下载完整pdf文档
M M August 11, 2021 09:06 PM GMT Global Semiconductor Memory – Winter Is Coming It is always difficult to predict cyclical turns, and investors' natural temptation is to follow the playbook and lock in gains, rather than wait for the upgrade cycle to fully unfurl. However, DRAM conditions are losing steam, and our inflection signposts suggest caution from here. Morgan Stanley does and seeks to do business with companies covered in Morgan Stanley Research. As a result, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of Morgan Stanley Research. Investors should consider Morgan Stanley Research as only a single factor in making their investment decision. For analyst certification and other important disclosures, refer to the Disclosure Section, located at the end of this report. + = Analysts employed by non-U.S. affiliates are not registered with FINRA, may not be associated persons of the member and may not be subject to NASD/NYSE restrictions on communications with a subject company, public appearances and trading securities held by a research analyst account. M M Contributors Morgan Stanley Asia Limited+ Shawn Kim Equity Analyst +852 3963-1005 Shawn.Kim@morganstanley.com Morgan Stanley Taiwan Limited+ Daniel Yen, CFA Equity Analyst +886 2 2730-2863 Daniel.Yen@morganstanley.com Morgan Stanley & Co. International plc+ Dominik Olszewski, CFA Equity Analyst +44 20 7425-7238 Dominik.Olszewski@morganstanley.com Morgan Stanley & Co. International plc, Seoul Branch+ Ryan Kim Equity Analyst +82 2 399-4939 Ryan.G.Kim@morganstanley.com Morgan Stanley & Co. LLC Joseph Moore Equity Analyst +1 212 761-7516 Joseph.Moore@morganstanley.com Morgan Stanley Taiwan Limited+ Ray Wu, CFA Equity Analyst +886 2 2730-2871 Ray.Wu@morganstanley.com Morgan Stanley Asia Limited+ Gilbert Wong, CFA Quantitative Strategist +852 2848-7102 Gilbert.Wong@morganstanley.com Morgan Stanley & Co. LLC Ethan Puritz Research Associate +1 212 761-1744 Ethan.Puritz@morganstanley.com Morgan Stanley Taiwan Limited+ Charlie Chan Equity Analyst +886 2 2730-1725 Charlie.Chan@morganstanley.com Morgan Stanley MUFG Securities Co., Ltd.+ Kazuo Yoshikawa, CFA Equity Analyst +81 3 6836-8408 Kazuo.Yoshikawa@morganstanleymufg.com Morgan Stanley Asia Limited+ Daisy Dai, CFA Equity Analyst +852 2848-7310 Daisy.Dai@morganstanley.com Morgan Stanley Asia Limited+ Duan Liu Research Associate +852 2239-7357 Duan.Liu@morganstanley.com 2 M M Memory – Winter Is Coming More risk than reward entering late cycle. While pricing is still moving higher, the rate of change is approaching peak as supply is catching up to demand. Our cycle indicator has shifted out of 'midcycle' to 'late-cycle' for the first time since 2019 and this phasechange has historically meant a challenging backdrop for forward returns. We expect earnings growth expectations to reverse, near 30% PB valuation contraction and higher chances of positioning reset. Inflection signposts – sell signals accumulate. We were prepared to become more constructive under the right conditions, but ultimately it is about inflections in the cycle and trajectory of earnings estimate revision breadth – the former approaching an earlier peak YoY pricing and latter approaching negative earnings risk that follows. Our belief is that: (i) the next cyclical downturn begins from 1Q22 and DRAM will stay fundamentally oversupplied in 2022, exacerbated by inventory builds; (ii) recent growth indicators have downticked, while demand held back by component supply; (iii) valuation is no longer compelling on the way down; and tactically buying on the dip no longer makes sense from a quant perspective – we expect better entry points in future. Higher sustained DRAM returns through the cycle. The industry is facing higher capital intensity and a flattening cost curve, which has created a fundamental change in chipmakers’ capital allocation decisions and return requirements. This is evident with DRAM industry operating margins troughing at a low-30% level in 2019, vs. losses in previous downturns and capital expenditure discipline via capex cuts, underpinning a steady rise in return on capital. What’s Changed SK Hynix Rating Price Target Samsung Electronics Price Target Samsung Electronics Price Target Micron Technology Inc. Price Target Rating S. Korea Technology Industry View From Overweight W156,000 To Underweight W80,000 W84,000 W77,000 W98,000 W89,000 $105.00 Overweight $75.00 Equal-weight In-Line Cautious We prefer NAND/NOR over DRAM and we have downgraded Hynix to UW. We also downgrade Micron, Nanya Tech and Winbond to EW. History says that stocks will tell us when it's time to buy, once they stop declining in reaction to bad news – but the bad news has to come first, and our downgrades partly reflect this process. We remain OW on Samsung, WDC and GigaDevice as we expect the former to outperform on the way down as it has in the past, and despite some concerns on NAND, we see value in WDC. Our EWs are NAND controller IC vendors Phison and SIMO (upgrading from UW), and we cut Macronix to EW on higher capex burden. Within our coverage, we are now an average 10% below consensus on 2022e EPS and see stocks moving back below mid-cycle multiples which form the basis of our PTs. Morgan Stanley Research 3 M M Contents 5 Memory – Late Cycle Brings More Risk Than Reward 17 1) Cycle Watch – Entering a Late Stage 22 2) Growth Indicators – Not Supported by Data 26 3) Earnings Estimate Revisions – Decelerating 28 4) Valuation – Less Compelling 30 QuanTECH – Tactical View on Memory 33 Implications for US Memory 38 Implications for Korea Memory 45 Implications for Greater China Memory 48 Implications for Japanese Semi Cap Equipment Stocks (Yoshikawa-san) 56 Financial Summary: Samsung Electronics 60 Financial Summary: SK Hynix 4 M M Memory – Late Cycle Brings More Risk than Reward What's changed? Cyclical conditions for DRAM have started to roll over. Initially, we believed that as long as demand remained strong into 4Q, prices could continue to go up since bit supply has outpaced production and help tighten inventory. However, while demand has stayed strong on a relative basis, it has worsened in recent weeks which has led to a downtick in pricing expectations. Initial indications are pointing to a more challenging environment for pricing into 4Q following prolonged 3Q negotiations that resulted in continued price hikes, and a reversal in trend into 2022. What this means in practice is that the investment case for memory will migrate from "what is the earnings potential?" toward "why should I own DRAM at this late stage, and what price is fair?". We are trimming estimates to reflect worsening DRAM conditions assuming an earlier peak (3Q21) and high single digit quarterly declines in 1H22. The story is changing and there is now better risk-reward elsewhere. Within memory, we relatively more positive on NAND is still relatively healthy according to our checks. However, with similar end-market exposures and an overlapping customer base, cyclicality vs. DRAM tends to be somewhat synchronized, which has us somewhat cautious on the direction of prices going forward. But there appears to be much less NAND inventory at customers vs. DRAM, given the fact that supply/demand tightened only recently and investor sentiment is much more cautious on NAND. We see a better near-term outlook for NOR on supply reduction in China and stay OW on semi equipment as DRAM is more about the inventory cycle and overall spending levels relatively benign. Stock implications – Downgrading DRAM... We are downgrading Hynix to UW as well as downgrading Micron, Nanya Tech and Winbond from OW to EW. This is a cyclical rotation and valuation call for DRAM and more structural for Macronix. Pricing is likely to track historical mean reversion on a price per gigabit basis from here, and this should create an incremental headwind to forecasts in 2022 and 2023. We see 0-6% downside to our base case price targets, and our bear cases revolve around a shorter down cycle, where an alignment of technicals and fundamentals drive multiples closer to 2016 levels or 29-40% lower from here. Key event catalysts for share prices near term include: (i) pricing negotiations which have become ad hoc (no longer quarterly) and rate of change in inventory – both leading indicators of the cycle; (ii) expected pre-announcements and management guidance turning to caution; and (iii) potential order cuts from the incremental demand weakness in consumer parts of tech such as PCs, Chinese smartphones and TVs. We downgrade Hynix to UW: Fundamentally, we believe Hynix's NAND exposure could have higher upside from here with Intel SSD acquisition pending approval but the stock is still largely driven by DRAM markets that could be over-earning. While guidance is bullish into 2H21 (SK Hynix: Off the Call - Bullish Guide Supports Consensus), investor expectations are clearly matching and demand conditions peaking. Our lower price target is driven by a reduction in multiple closer to historical trough valuation of 1.2x NTM P/B as we now see risk to future earnings revisions that would justify a discount. We see a 2:1 bear/bull skew and using our residual income valuation on our base case earnings, the shares are trading above fair value in what should be viewed as an optimistic earnings scenario. We downgrade Micron to EW: While we have been impressed by the structural improvements at the company, and we see limited downside, we see the stock as effectively rangebound in an environment where DRAM prices start to decline. August quarter results should still have upside, and that could persist into November. But the 35% increase in DRAM prices in the last two quarters, with further modest increase expected through year end, creates risks of a steeper decline next year, given the elevated state of customer inventories in some markets. We are trimming our estimates, assuming high single digit blended DRAM price declines in the February and May quarters and we lower our PT from US$105 to US$75, driven by 10x through-cycle earnings of US$7.50, down from 14x – as we have noted in the past, the multiple on through cycle earnings tends to compress to about 10 during periods of falling prices. Micron has highlighted potential customer shift from "just-in-time" to "just-in-case" inventory levels in some cases; we think that is a good articulation of what is happening – namely that customers are holding high inventory levels, but that they are willing to move those inventories even higher as overall supply conditions remain tight. That is a general theme across semiconductors at the moment, given rolling outages. The problem we see for DRAM is that modest price declines could snowball quickly, as customers are likely to react to such shifts by slowing purchasing to draw down inventory. We see Morgan Stanley Research 5 M M that as an issue in PCs and in servers, and while we didn't see that as likely in mobile, the company's recent commentary that mobile DRAM is somewhat elevated as well creates some concerns there. Having said that, we do highlight some offsets: Micron's balance sheet inventory is very lean, implying minimal effective supply growth in 2H; this kept us OW until now, as even with some issues in the end markets, the company will effectively see supply contract in 2h (as the inventory drawdown from their own balance sheet in 1H will not recur in 2H); this is why there are still selective price increases in 3Q, and why we thought there could be further increases in 4Q, but recent checks show that isn't the case. We also see continued strength in embedded DRAM, graphics DRAM, and NAND persisting through 4Q, as customer inventories in those areas seem much less elevated. That is roughly half of the company's revenue that is still showing improvement, should serve to protect high earnings levels for at least a couple more quarters. Changes to Micron estimates: We are making changes to our model to reflect DRAM headwinds, particularly in the PC and server markets. While 2021 estimates remain unchanged, our 2022e revenue, gross margin and EPS on a non-GAAP basis come down meaningfully from US$36.4bn, 43.8% and US$9.98 to US$33.2bn, 38.3% and US$7.32, respectively. On a ModelWare basis, which includes stock compensation expense, our 2022e EPS comes down from US$9.87 to US$7.18. We downgrade specialty DRAM player Nanya Tech to EW: As we believe the cycle for specialty DRAM may peak soon in 2H21, and we believe 2H21 should be the peak for NTC's business – after its sequential growth of specialty DRAM price in 3Q21. We therefore think Nanya Tech's DRAM ASP may turn to flat Q/Q in 4Q21. We also see specialty DRAM spot price weakness in near term, and we see more risk on specialty DRAM pricing starting from 4Q21. The market has already fully anticipated that the specialty DRAM contract price will catch up to the spot price in 3Q21. While the specialty DRAM spot premium is now only at 4% (vs. 70-90% in 1H21 vs. 25% 1.5 month ago), this means that it is increasingly difficult to have a clear view about the contract price trend into 4Q21. Moreover, specialty DRAM spot price has further declined 8% in the past month, which means that the higher spot price is not be sustainable to its end-clients. We are also UW on Powerchip Semiconductor (PSMC), as we believe its Pricing Power Won't Last. We believe the market expectation for the stock might be too high as bull camps are bullish on its logic semi business pricing uptrend (55% revenue contribution), while we see emerging risk on its DRAM foundry business (45% revenue exposure) during a DRAM pricing downturn. As PSMC is only working on A DRAM foundry business, its ASP volatility should be only 60-70% of DRAM IDM players. Bear in mind that its DRAM business ASP fell 50% in 2019 during the downcycle. … Relative Preference for Samsung, NAND, NOR, Semi Cap Our price target moves from US$105 to US$75. We still view through-cycle earnings potential at about US$7.50, but we have commented in the past that multiples on through cycle tend to peak when conditions are improving, and come under pressure during periods of downward revision; our prior target was based on 14x through cycle, but we are lowing that to about 10x, based on historical patterns. While 10x should limit downside, we do think that a weaker multiple is appropriate given that free cash flow per share should run about 40% below reported EPS, on average, given the large disparity between depreciation and capital spending, and given that earnings deceleration is seldom graceful – even in recent years where we have been impressed by the trough earnings power. Exhibit 1: Changes to Micron estimates for 2021 and 2022 We favor Samsung – quality with a more defensive tilt. Quality can mean a lot of things and for many investors it simply means large cap growth wins in a decelerating environment. While Samsung fits the high-quality bill, we think the quality rotation will now begin to favor Samsung's more defensive properties like earnings stability rather than growth. In previous corrections, Samsung has consistently outperformed the memory sector and often the MSCI Asia benchmark, and we believe the business attributes are fundamentally more appealing in the current environment given cost/product leadership as well as optionality on seizing new addressable markets. Cash conversion, growth and ROCE characteristics are close to best-in-class among quality tech – we think the valuation along with high and secure dividend yield/cash returns are supportive in a down cycle. Revenue (US$mn) Gross Margin EPS (US$) New $30,306 42.3% $7.41 CY2021e Old $30,306 42.3% $7.41 % Change 0.0% 0.0% 0.0% New $33,171 38.3% $7.32 CY2022e Old % Change $36,386 -8.8% 43.8% -5.5% $9.98 -26.7% Source: Morgan Stanley Research estimates 6 M M We are relatively more positive in NAND and stay OW WDC. Even though DRAM conditions have started to erode, NAND is still relatively healthy according to our checks. However, with similar endmarket exposures and an overlapping customer base, cyclicality between the two commodities tends to be somewhat synchronized. Typically, cycles occur within about three months of one another, which has us somewhat cautious on the direction and choppiness of NAND prices. But there appears to be much less NAND inventory at customers vs. DRAM, given the fact that supply/demand tightened only recently. Also important is that investor sentiment is much more cautious on NAND given recent countercyclical capital spending, and Chinese entry - we note that the implied valuation of WDC's NAND business, is less than 1x sales if we value their hard disk drive business in line with comparables, vs. 3.5x for MU - and any incremental positives, such as sector consolidation, capital spending contraction, or setbacks with Chinese competition would be an incremental positive for the stock. The NOR upcyle has more legs, but sentiment could be affected (refer to NOR Flash - cycle not peaking yet where in June we believed that that 3Q NOR pricing could be up by 10-20% Q/Q, with some specs up to 25%, and 4Q pricing could rise further, driven by widened supply/demand gap). 3Q NOR pricing uptrend is consistent with our view and we now expect the NOR pricing could go up by mid-singledigits Q/Q in 4Q, led by high-density NOR. The supply reduction in China and increasing structural demand from PC continues to drive up the pricing. We believe the outlook is better than DRAM and thus suggest stocks with NOR exposure could outperform global DRAM peers. We thus have raised our PT for GigaDevice given the better NOR Flash pricing into 4Q. We downgrade Winbond PT to EW as better NOR pricing and MCU could be offset by weaker DRAM. We also think Macronix's earnings could decline in 2022e given the higher depreciation despite better NOR, and thus we have decided to downgrade to EW. For Japan semi equipment (covered by Kazuo Yoshikawa). If memory enters a cyclical deceleration phase from 1H22, the risk of memory capex cut/push-out is likely to grow, as in past deceleration phases. Risk/reward profiles for semi cap equipment stocks are likely to deteriorate. However, we think Tokyo Electron (TEL) will outperform among semi cap equipment stocks, and we maintain our OW rating on TEL. • First, given rising capital intensity of advanced devices and trends of localizing chip production, we think the potential downside for WFE markets is less than during previous cycles. • Second, field solutions account for 20-25% of the company's semi cap equipment sales, and as the installation base expands we anticipate relatively stable growth in this area. In fact, during the previous memory downturn in F3/20, new equipment sales fell 14% YoY (including DRAM equipment -44%, NVM equipment -55%), but field solutions sales actually increased 6%. • Third, we estimate TEL's end-F3/22 net cash position at about ¥370bn (~¥2,400/share). The company's target financial model assumes RoE of 30%, and we think it is likely to pursue share buybacks in 2022 in order to achieve that goal, providing downside support for the share price. For US semi equipment (covered by Joe Moore): We budget for some deceleration in NAND spending next year, and could see some choppier spending in DRAM in 2022. But, we note that absolute levels of DRAM spending have remained reasonably constrained – for us, this cycle is more about the inventory cycle around covid than about excess spending – and while there is the expectation of a surge in 2h21, overall spending levels are relatively benign. It is harder to make that case in NAND, but we would still note that spending outside of Samsung and mainland China has actually been quite low, and we would expect continued recovery in spending for other participants. The bigger factor will be spending in foundry and logic, which looks to remain reasonably robust next year. Our sense is that cutting-edge foundry spending could decelerate at the margin, but spending on older nodes will likely remain strong. Further, sovereign drivers of foundry/logic spending are likely to remain quite strong, with particular impact in China and at Intel, which looks to be spending at an elevated level next year. For EU semi equipment (covered by Dom Olszewski), weakness in the DRAM cycle could affect orders from mid-2022 (particularly around immersion and DUV tools). However, we retain our Overweight rating on ASML and would expect the stock to outperform during a memory inventory correction as envisioned. • Context on recent DRAM capex cycles: Capital expenditure for DRAM was cut by 23% in 2019 and stayed flat in 2020 while the 2015 downturn only saw 9% cuts. However, capital intensity is higher this cycle – at ~30% of sales in 2019 vs, an average of ~23% since 2015. This could keep absolute capex levels elevated in upcoming downturns. ASML is most affected by developments in DRAM (as opposed to NAND where exposure is more limited). • How could ASML be affected by a DRAM downturn? 1) The most obvious immediate impact is on cancelled immersion Morgan Stanley Research 7 M M tools orders (demanded in large volume by DRAM customers). We model immersion shipments remaining at an elevated level over 2021-23 before declining back towards 68 per annum in 2025. 2) Loss of DRAM orders would also bring the margin mix lower (given the strong margins for immersion). 3) Significant weakness in memory demand could delay/disrupt plans for DUV capacity expansion (or result in overcapacity). Such an expansion in DUV was not expected earlier this year, but came into view as demand for older DUV tools continued to outstrip supply (thanks to unexpectedly strong demand at legacy nodes). • However, on balance we continue to view ASML as an outperformer during a DRAM inventory correction (as opposed to a deeper DRAM drawdown). ASML is protected by: 1) strategic importance of retaining high capex intensity – this is a defensive measure for incumbent memory chipmakers, to retain market share and ensure they are best positioned into the next upturn; 2) long lead times for equipment – which is in particular shortage today. While immersion tools (often bought by DRAM customers) have shorter lead times than EUV (12-18 months, mostly purchased by foundry customers today) – total litho equipment is still among the leadtime purchases; 3) ASML typically sees some offset from the counter-cyclical upgrade business (which typically underperforms in up cycles with tools fully utilised, and outperforms during downturns as tools have lower utilisation). Downgrade our Korea Technology Industry View from In-Line to Cautious. The risk/reward has turned less favorable from here, with many of the stocks having already priced in a recovery; however, downside to earnings across most stocks in our coverage and a mean reversion of valuation multiples is a norm post cycle inflection. Therefore, we prefer quality/large cap with Samsung remaining as our only OW-rated stock in our coverage, although this is now a relative rather than an absolute call. Consequently, we believe an In-Line industry view is no longer warranted. Exhibit 2: Breakdown of cost for a typical DRAM fab – immersion (ArFi) tools accounting for largest part of litho spending. Split of lithography spend by node (for 100kwspm) 100% 90% 80% 70% EUV 60% 50% ArFi 40% ArF 30% KrF 20% i-line 10% 0% 38nm 2 x H 2 x M 2 x L 1 x H 1 x M (2 1 x L (4 EUV EUV layers) layers) Source: Company data, Morgan Stanley Research estimates Exhibit 3: Split of ASML's lithography sales by technology (MSe) – immersion tools estimated to peak in 2022 before declines. Unit litho sales by technology: EUV ArF immersion ArF dry KrF i-line High-NA 2018 18 86 16 78 26 0 2019 26 82 22 65 34 0 2020 31 68 22 103 34 0 2021 41 94 28 128 35 0 2022 55 95 40 140 60 0 2023 63 88 20 120 30 2 2024 65 68 20 86 30 5 2025 68 68 20 86 30 8 Source: Company data, Morgan Stanley Research estimates 8 M M We are entering late cycle faster than normal What prompted us to downgrade? Cyclical peak – the rate of change in prices has been a consistent indicator of inflections in the memory cycle and stock prices. DRAM stocks have become the victims of their own success approaching cyclical peaks in terms of YoY rate of change in pricing and revenues – traditionally a point where market participants reduce their exposure to the group. When perceptions build for future quarterly prices to plateau and/or fall, customer focus turns to reduce inventory to minimize the negative impact of depreciating inventory. This has become an increasingly likely scenario for DRAM by 4Q21. Earlier pricing inflection. Looking into 4Q, our sense is that pricing will decline as inventory is still elevated at customers and with greater confidence of price declines on the horizon, the need to hold inventory also declines which can act as a feedback loop, driving prices lower. DRAM PC spot prices have begun to decelerate sharply in the last week by 7%, now only 3.8% above contract and a good leading indicator for contract prices by about two to three months. Our channel checks have ticked down in recent weeks. What was weaker channel pricing relative to contract for graphics and mobile DRAM has now moved to PCs. Even though DRAM conditions have started to roll over, NAND is still relatively healthy according to our checks. However, with similar end market exposures and an overlapping customer base, cyclicality between the two commodities tends to be somewhat synchronized. Customers are still swimming in chips. Our recent channel checks in memory suggest that customers have built sufficient inventory ahead of aggressive pricing from producers, despite the decline in producers' inventory exiting 2Q21. As a result, several customers are passing on supply, first in the cloud service provider space and more recently from PC/server OEMs, in hopes that they can purchase at lower prices down the road. Earnings growth expectations are decelerating sharply. The revision breadth is likely to get significantly worse before bottoming. History tells us to be patient in these instances as it will be challenging for stocks to outperform from here and earnings estimate revisions momentum moving toward negative territory can be dangerous. Memory stocks are not pricing in challenging conditions. It is no longer about whether it is cheap in absolute terms, it's about whether one is over- or underpaying for stability. Valuations (12-month forward P/E) have quickly snapped back to 7.2-16.4x P/E for memory stocks, 1.5-3.2x EV/Sales and 1.6-2.1x P/B – but still at a large premium and well above mid-cycle multiples. We are seeing incremental weakness in DRAM vs. improved NAND fundamentals per our discussions with industry contacts and also based on our supply/demand analysis – i.e. the rate of change on growth should peak by 4Q21 for DRAM, excess inventory at customers spilling over to producers soon, and valuations are falling. We are likely to see a softening of average selling prices in the DRAM market – they will likely start plateauing through the end of this year and then start to decline as capacity begins to come online and ramps up, leading to bigger tumbles in market prices by 2H22. The component supply constraints have us entering a late-cycle environment for DRAM quicker, and market internals are reflecting that – recent corrections in DRAM stocks are sending a powerful message about the durability of current cycle gains and the next stage for the cycle. Taking a long-term view. The market may overlook the extent to which industry consolidation, supply-side discipline and constraints, and secular demand growth can soften down cycles and drive longterm upside to return on capital. The current negative sentiment and relatively low valuations in DRAM suggest that the market may have overlooked the extent to which the increasing capital intensity of capacity additions, and, broadening demand should help to soften the inevitable down cycles and drive long-term earnings growth. It may seem like over-investment is the root cause of the increasing mismatch between demand and supply with DRAM capex moving higher – up 30-40% in 2021 – however, much of this reflects rising capital intensity rather than an acceleration in wafer capacity. Morgan Stanley Research 9 M M Exhibit 4: DRAM supply growth peaking in 2Q21 is a tailwind Exhibit 5: Long-term DRAM ROIC requires higher through-cycle against slowing demand in 2022 margins to compensate for rising capital intensity Source: Morgan Stanley estimates, Gartner, DRAMeXchange Source: Bloomberg, Morgan Stanley Research; Notes: Average ROIC of Samsung Electronics, SKY Hynix, Micron, WDC, and Nanya Tech Winter is coming sooner A cyclical downturn is likely from early 2022... Playing the second derivative inflection in DRAM (measured in terms of the monthly rate of change in contract pricing illustrated in Exhibit 6 ) could see contract pricing fall a long way – typically -50% YoY for both blended and spot prices – and we forecast material downside for memory prices and earnings. Memory valuations have come down since their 1Q peak but earnings expectations have not. Both supply and inventory present higher downside risk as we move to 1Q22, despite structural growth in cloud that continues. The current memory cycle is currently in a delicate phase with its own version of inflation (price) vs. growth scare (component constraints) explained in Memory - The Ambiguous Cycle. As we move closer to 4Q21, the main investor debate at present will be around the durability of cyclical positioning against a backdrop of peaking prices. … and inventory driven, given limited expansion in new capacity and underlying supply growth. The duration of the down cycles has more or less remained the same, but the magnitude of the sales decline may not be as large as before. In our view, this is due to the structural changes that have taken place in the DRAM industry – consolidation, a traditional barrier to entry – has increased and demand drivers are broadening and we expect the trend of softer down-cycles (higher lows) to continue to be the norm going forward. What were changes in PC demand as the dominant end-use market in the past, has now shifted to the number of end-use products that continues to grow, such as AI-enabled chips (in mobile, server and automotive indus- Exhibit 6: DRAM contract price YoY or the cycle likely peak in 4Q21 4-years 4-years 4-years 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 4-years 140% 120% 100% 80% 60% 40% 20% 0% -20% -40% -60% -80% 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Source: DRAMExchange, Morgan Stanley Research 10 Samsung EV/Sales DRAM Contract Price YoY (RHS) M M Exhibit 7: In a downturn, DRAM tends to underperform for ~4 quarters before finding a bottom Rebased to M 1.2 0.8 0.4 M M+6 '96 Cycle* '00 Cycle * Excluded SK Hynix as the Company was private until Dec-96 Source: Thomson Reuters, Morgan Stanley Research '15 Cycle 50% to Peak '18 Cycle M+12 21 Cycle tries) in need of larger die sizes (DDR5), and applications requiring high memory band width. This permeation into virtually every sector of the broader economy has made the industry more dynamic by diversifying its long-term growth prospects and likely resulting in milder down-cycles. Downcycle transitions commonly see P/B valuations contract by 50-64%. DRAM stocks can pull back 32-63% from peak to trough at any given downturn – we are now 12% lower from the April 2021 high – and the average stock price was down 43% in the last two down- turns. The risk of disappointment is high in the late stages of the cycle as price hikes have historically been a 'sell the news' events. While the risk premium compresses following excitement around recovery, eventually positioning resets, and a slowing rate of change in growth offset this and sent multiples lower. Structurally higher earnings and margins through the cycle mitigates downside, but the pricing and inventory correction that follows limits the upside. The net effect is a tug-of-war between earnings/valuation, tepid returns over the next 12 months, and a likely 30-40% correction in P/B valuations between here and there. Morgan Stanley Research 11 M M Exhibit 8: DRAM P/B cycles see multiples correct 50-64%... 2.3 -14% 1.9 Exhibit 9: ...with average market cap falling 34% to 41% 250,000 200,000 150,000 -15% 1.5 -50% 1.1 -64% -59% 100,000 -34% 50,000 -41% -40% 0.7 Dec-10 Dec-11 Dec-12 Dec-13 Dec-14 Dec-15 Source: Bloomberg, Morgan Stanley Research Dec-16 Dec-17 Dec-18 Dec-19 Dec-20 0 Dec-10 Dec-11 Dec-12 Dec-13 Dec-14 Dec-15 Source: Company data, Morgan Stanley Research Dec-16 Dec-17 Dec-18 Dec-19 Dec-20 What works in the late-to-down cycle transition? As a rule of thumb, the magnitude to the share price downside is proportional to the excess created in the current upturn. Broadly speaking, memory is not part of this 'excess' group but moving up the quality curve is a good idea as the quality factor has consistently outperformed during downturns. One example is Samsung stock as best positioned for a memory cyclical slowdown as its strong balance sheet, capital returns, optionality on growth, and cost advantage led to its shares always outperforming MSCI IT in a declining price environment. For pure plays such as Hynix and Nanya Tech, cyclical concerns could lead investors to shy away from these names given their relatively higher financial and operating leverage. Exhibit 10: Share price performance has bounced back since March lows last year 280 260 Rebased to 100 240 March 31, 2020 220 200 180 160 140 120 100 80 Jan-20 Apr-20 Source: Bloomberg, Morgan Stanley Research Jul-20 NOR Oct-20 DRAM NAND Jan-21 Semi Equipment Apr-21 Jul-21 12 M M Our key signposts for a peak in memory We acknowledge that a lot of good news was priced in, in 1Q21. However, hopes of a super cycle have faded since with higher production run rates in DRAM and memory valuation being less compelling in a downturn. While we would not exclude a positioning rally nearterm as an upside risk given the YTD under-performance and consensus 'funding short' positions, we think fundamentals eventually matter more than technicals, and see clear incremental risk that an earnings recession is ahead for the sector in 2022. We expect consensus earnings estimate revisions to turn incrementally more negative, thus, investors need to pay attention to these changes since earnings achievability will likely determine whether stocks can rally, and which ones remain vulnerable. We have analyzed our five signposts for DRAM stabilization across five factors, which we believe are necessary to call a fundamental trough for stocks – all pointing to caution on valuation, with some exception: 1. Cycle watch – approaching a second derivative inflection. We have compared the current cyclical downturn against prior cycles across five peak-to-trough absolute and relative factors. Capital spending has not been as high compared to the previous peak – with back-to-back years of increases of just +17% and +10% in 2021-22e DRAM capital spending, (compared to +23% in 2017 and another +72% in 2018). This is likely to result in a mild oversupply that we think may be shorter than previous normal cycles to reach equilibrium. Aside from cloud and enterprise, core end-markets in consumer continue to deteriorate and inventory is likely to extend, suggesting a path for price erosion into 2022. 2. Growth indicators – hurt by component shortages. Memory cycles are supply driven and the memory revenue cycle is often independent of the economic cycle, although a severe economic downturn can exacerbate a memory downturn. As such, we believe the magnitude of an economic recovery would be not relevant for memory in 2022, and a fully supportive Fed policy or stimulus would not change the cycle. Nevertheless, component shortages have been a hindrance on demand – and these have resulted in DRAM customers' inventory swelling since 2Q21 – and we forecast endmarkets to decelerate in 2022. 3. Earnings – optimistic but consensus. We struggle to find any fundamental EPS improvements in 2022 as we move toward peak cycle in 4Q21 and into a deceleration phase just at the beginning stages. The downside risks to DRAM pricing to mean revert back to previous trough levels (-40% from here) may be mitigated by supply growth challenges, but directionally, the revision breadth is likely to worsen before finding a bottom. We cut 2022e EPS forecasts by 0-25% are now 0-38% below consensus numbers. Memory stocks' relative earnings estimate revisions breadth has began to decelerate since 2Q this year, and such shift in momentum readings are typically not a good signal for forward performance, and relative underperformance moves well ahead of the turn in relative negative revisions. 4. Valuation – not compelling in a downturn. Stocks are at near +1SD above mid-cycle valuations and 80% and 60% from recent trough P/B and EV/Sales multiples, respectively. P/E is a contrarian indicator and currently stands slightly below +1SD mid-cycle multiples, but this should be looked at relative to trough EPS, which is likely to decline. If further evidence emerges of a deeper downturn in 2022, or no signs of recovery in 2H demand, then we expect stocks to move closer to trough multiples, as in past downturns. The real problem on the upside is that we see fundamentally no driver for continued sector demand and/or EPS growth. This will limit any upside to valuations, in our view. 5. Tactical – isthe stock (not fundamental) risk/reward better now? From a tactical perspective, we think stocks have near-term potential to rally before correcting sharply from here to reflect underlying fundamentals. From a top down perspective, our EM strategist, Jonathan Garner, sees better value in other cyclical industries APxJ/EM Sector Allocation). Our quant analyst, Gilbert Wong, believes his QuanTECH framework (QuanTECH – Tactical View on Memory) suggests to turn cautious on Memory players as we are now standing at the late stage of market cycle because of: (i) memory players used to lead DRAM price to rotate into down-cycles; (ii) the peaked Asia PMI cycle is elevating the downward earnings revisions risk among Growth stocks, especially memory stocks; and (iii) Asia/EM market is undergoing risk-off trajectory to push valuation lower. Morgan Stanley Research 13 M M Exhibit 11: Samsung Electronics Support Levels… 100000 90000 80000 Last Price SMAVG (50) SMAVG (100) 70000 SMAVG (200) Shares Moving Towards 60000 50000 Exhibit 12: … While SK Hynix Shares Breaking Below 160000 140000 Last Price SMAVG (50) 120000 SMAVG (100) SMAVG (200) 100000 80000 60000 40000 40000 30000 20000 2016 2017 2018 Source: Bloomberg, Morgan Stanley Research 2019 2020 2021 20000 2016 2017 2018 Source: Bloomberg, Morgan Stanley Research 2019 2020 2021 Memory stock valuations: In addition to the top picks from our analysts below, we highlight stocks that we think are favored and disfavored in the current rally using our quantamental methodology in Exhibit 5 . Exhibit 13: Valuation for memory stocks – what's in the price? Ticker Name 005930.KS Samsung Electronics 2344.TW Winbond Electronics Corp SIMO.O Silicon Motion 8299.TWO Phison Electronics Corp 000660.KS SK Hynix WDC.O Western Digital MU.O Micron Technology Inc. 2337.TW Macronix International Co Ltd 603986.SS GigaDevice Semiconductor Beijing Inc 2408.TW Nanya Technology Corp. Sub-sector DRAM Specialty Memory Specialty Memory Specialty Memory DRAM NAND DRAM Specialty Memory Specialty Memory Specialty Memory Market Cap (US$ mn) 410,586 4,545 2,118 3,357 71,901 19,901 86,801 2,650 20,473 7,940 Source: Bloomberg, Morgan Stanley Research Current Trailing PB 1.82 1.99 3.53 2.63 1.56 1.94 1.98 2.14 10.29 1.48 5Y Peak 2.03 2.25 4.94 3.69 2.21 2.80 3.09 4.99 26.19 9.70 % to Peak 12% 13% 40% 41% 41% 45% 57% 134% 154% 557% Blended Abs. Valn. Blended Rel. Valn. 5Y Z-Score 5Y Z-Score 1.97 0.42 2.02 0.14 0.59 0.55 1.01 0.19 0.58 0.11 0.42 1.10 0.37 -0.21 0.15 0.20 -0.79 0.17 -0.16 0.57 What happens in a downturn? The rate of change in DRAM contract prices has been a consistent indicator of inflections in the memory cycle and stock prices: Investors have not been rewarded in past cycles for buying stocks that are cheap on peak earnings, once meaningful second derivative weakness has begun. Second derivative weakness at cycle peaks is such a consistent leading indicator because there are positive feedback loops with customer inventory behavior that are far more durable than week-to-week demand variances. When perceptions build for future quarterly prices to plateau and/or fall, customer focus turns to reduce inventory to minimize the negative impact of depreciating inventory. This has become an increasingly likely scenario for DRAM by 4Q21. 14 Share prices normally lead DRAM and concurrently with the NAND cycle at trough: In most cases the trough of stock prices was concurrent or lagged by one month before the inflection point (second derivative), measured by the rate of revenue growth for DRAM. In the 2018 cycle, DRAM YoY contract pricing peaked in September 2018 and troughed by April 2019 while stock prices bottomed one quarter ahead in December 2018. Conversely for NAND, contract pricing bottomed in February 2016, while share prices began to outperform concurrently around the same time. The main reason was lack of visibility relative to DRAM due to a larger number of suppliers and demand elasticity on pricing dynamics. Duration tends to be four to eight quarters in a downturn and four to nine quarters in an upturn: The slowdown tends to be shorter M M than the upturn in the memory cycle because it is easy to cancel orders and rationalize inventory but a lot harder to bring back capacity and ramp up production. The 2018 downturn lasted only for four quarters for share prices from September 2017 to April 2019. Given our forecast of the peak in DRAM contract pricing by 4Q21, we are currently moving toward late stages of the cycle. DRAM share price performance is driven by valuation in the early stage of a down cycle: In the 2018 cycle, EPS estimate revisions lagged the peak of share prices by three months, while de-ratings drove the initial downward leg of share price performance, anticipating the inflection of the DRAM cycle. Multiples tend to bottom around the inflection in the second derivative (either pricing or monthly revenues), which is often the catalyst for share prices to reverse performance. In terms of EPS estimate revisions breadth, peak to trough magnitude is almost 2x (+102% to -88%) on average and duration of EPS estimate revisions is 12 months. Margins in previous cycles have compressed by 45ppt for Hynix (2018 cycle from 65% in 3Q18 to 20% in 4Q19) and are currently (3Q21e) at +43%. When to buy memory? We expect poor sentiment and forced selling to weigh on share prices until four conditions are met (none likely before year-end): • DRAM prices find a bottom. We expect 3Q21 contract pricing outcomes closer to +10% for DRAM (NAND also +10%) to be better than market expectations and DRAM peak pricing by 4Q21. However, demand for chips is moderating as customers have been constrained in certain types of components and overstocked on others. This makes suppliers vulnerable if component constraints extend or sell-through falters, which could eventually result in an air pocket for memory demand. Separately, smartphone data points have remained relatively weak in China, along with parts of consumer (mid/small TV, Chrome PC) and graphics DRAM deteriorating sharply in recent weeks. Cloud demand looks more durable and will grow strongly throughout the year as cloud providers start to spend more countercyclically and consume a lot of DRAM content. In the previous three cycles, DRAM prices have consistently adjusted lower by about 50% YoY before finding a bottom. • Inventory adjustment cycle. The recent DRAM excess builds at PC, cloud and mobile customers – owing to constraints in the supply chain – are driving end customers to hold more inventory. In a downturn, we note that falling prices lead to accelerated customer inventory depletion and should spill over to producers before returning to normal levels. During an upturn, lead times for the delivery of chips to manufacturers of PCs, servers and smartphones tend to grow longer. This prompts those customers to order more than they need, and start hoarding that inventory in case those lead times grow into a real shortage. It is the opposite in a downturn and a potential bottoming should shape up in late 2022/early 2023, we estimate, after all those excess chip inventories are used up. • Capex budgets pared down while capital intensity moves higher. There is always either too much or too little supply in memory. The industry is competitive but also rational, and should therefore be responsive to periods of price disequilibrium. In response to the increase in those inventory and prospects of lower profitability, chip manufacturers reduce their capacity growth to manufacture chips to meet that demand. Capital expenditure for DRAM was cut by 23% in 2019 and stayed flat in 2020 while the 2015 downturn only saw 9% cuts. However, capital intensity is higher this time around as it jumped to ~30% of sales in 2019 vs, an average of ~23% since 2015. This is likely to keep absolute capex levels elevated in upcoming downturns and also prompts companies to require higher returns to justify those investments. • An upcoming wave of negative earnings revisions passes. When the cycle turns, we expect revenues to decline (52% on average for DRAM), and push down margins on the fixed cost infrastructure, which in turn will cut earnings. While there could be further margin expansion near-term as companies look to rebuild inventory, expansion from here is likely to prove more moderate. The market often runs ahead of the actual recovery in earnings and we would look to the earnings revision breadth to bottom in terms of rate of change. A turnaround in consensus estimates from a downgrade to an upgrade cycle is needed to give the equity market confidence that earnings power is not cyclically impaired – however, this is potentially some way out. This view is not without risks Way too early to fade this trade. We are not at peak earnings (we forecast in 4Q21) and could still have an extended period of sustained earnings. Supply tightness across a wide range of components and the perceived pent-up demand that was unfulfilled for the past two quarters could support a cycle that is "stronger for longer" compared to past upturns. The PC and server industries were simply caught flat-footed after a lack of available capacity from foundries, while at the same time the world went virtual and accelerated everything digital. Morgan Stanley Research 15 M M Exhibit 14: Memory stocks preference table Source: Morgan Stanley Research estimates, Refinitiv. Pricing data as of August 10, 2021. 16 M M 1) Cycle Watch – Entering a Late Stage Finding the sweet spot – share prices follow the cycle. The July 2021 performance in memory stocks is inconsistent with our inflection playbook. Exhibit 10 shows the consistent relationship – consistent on timing, consistent on magnitude – between DRAM prices' rate of change and the multiple of stocks. Supply adjustments coming in steps drive inflections, and inventory levels push the magnitude of the inflection points. Demand also catches up, but only gradually, and is less of factor to reaching equilibrium vs. cycles that are largely supply driven. Looking specifically at stocks, we find that shares tend to rally into a YoY trough and performance flattens 6-12 months after but still yields positive returns. However, stocks start to struggle 12 months after cycle troughs. We think that the recent rally in memory stocks is large relative to what usually happens after trough inflections, but we have yet to reach a inflection point on pricing. Bear market rallies are common in the memory sector, and we advocate a disciplined valuation and trading approach to participate in such rallies, while acknowledging and positioning for the greater downtrend. However, we think memory stocks have generally overshot to the upside at present and are due another deep correction, especially given what we think will be sobering earnings in the coming quarters, with even worse results likely to be seen in 2H22. A historical case study of five downturns over the past 25 years points to further absolute downside to a distant trough from here. We do think the fundamentals are deteriorating more than the market generally expects, and see scope for disappointment in both the depth and duration of the decline. Fundamentally not different to previous "boom and bust" cycles. The price increases that started at the end of 2016 until 3Q18, or two years of chip shortages, is about normal with the only difference being shortages were extreme due to the upturn in cloud spending from 2016. This triggered a simultaneous increase in production with the magnitude of investment in capacity the highest in history given most of companies had excess capital to expand capacity with profits from the peak years in 2017-18, as evidenced by Exhibit 4 . We think there has been too much capital spending in the last two years – backto-back years of big increases in DRAM capital spending up 23% in 2017 and another 72% in 2018 – which will cause an oversupply in the coming two years. The price per gigabyte is the primary product differentiator for DRAM, and market concentration has less influence due to the strong commoditization effect in DRAM industry and innovation (tech migration) driving cost reduction, which eventually increases intensity of competition. At the same time, large productivity gains associated with technology migration in the past have reduced sharply in recent cycles due to the unit density of DRAM plateauing, which may have a higher impact on margins. Exhibit 15: YoY PC DRAM spot price cycle – closer to peak now 3.2 2.7 2.2 1.7 1.2 0.7 Jul-13 Jul-14 Jul-15 Jul-16 Jul-17 Jul-18 Jul-19 Jul-20 DRAM EV/Sales Source: DRAMeXchange, Morgan Stanley Research DRAM Spot Price YoY (RHS) 120% 100% 80% 60% 40% 20% 0% -20% -40% -60% -80% Jul-21 Morgan Stanley Research 17 M M Exhibit 16: Long-term DRAM capital spend – sizeable investments in 2018 drove supply excess 35 30 25 22.0 30.9 30.9 28.1 23.8 24.1 20 17.0 15 13.3 10.6 9.3 10 6.5 6.6 4.8 5 11.3 12.3 17.9 16.0 14.6 13.4 7.8 7.1 5.9 4.6 200% 150% 100% 50% 0% -50% 0 -100% 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21e 22e Source: SIA, SEMI, Morgan Stanley Research (e) estimates How far ahead of the next upcycle will the market discount improvement? The negative earnings estimate revision cycle and the lack of any sign of recovery suggest to us that memory stocks should continue to underperform. Short-term sentiment rallies are common during memory downcycles, but the window to trade them is often short. Our expectation of a 2020 recovery implies that the market should begin to discount the next upcycle in late 2019. In addition, there are better "early cycle" sectors to play that provide better risk/return than DRAM at this point in the cycle, such as cloud. In fact, downcycles usually begin with a powerful inventory correction that runs for 9-12 months. Order rates stabilize much later, but that is essentially when the price/margin begins to tick up. So, essentially, by focusing on the top-line 'second derivative', investors could be missing the much more important deteriorating rate of change in EPS growth. The last downcycle taught us this valuable lesson; this cycle could be far worse in this regard. Share price performance lagged EPS revisions by one to two months in previous downturns: Despite, some short-lived ups and downs, in 2016 memory stocks only started to rebound meaningfully one to two months later than the recovery in EPS revision breadth. For Samsung, earnings estimate revision breadth turned negative from December 2015, deteriorating sharply to -57% by January 2016, before recovering and turning positive again in April 2016. Meanwhile, its share price performance started to rebound in February 2016. Similarly, for SK Hynix, earnings estimate revision breadth turned negative from July 2015, deteriorating to -85% in March 2016, while its share price performance started to react to upward revisions from May 2016, two months into a bottom in earnings estimate revisions. Exhibit 17: Earnings revisions matter for Samsung – EPS revision breadth vs. share price performance … 100% 120% 50% 0% -50% 80% 40% 0% -40% -100% -80% Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan-17 Jan-18 Jan-19 Jan-20 Jan-21 Earnings Revision Breadth Source: Refinitiv, Morgan Stanley Research Share Price Perf. YoY (RHS) Exhibit 18: ...as well as for SK Hynix – EPS revision breadth vs. share price performance Source: Refinitiv, Morgan Stanley Research 18 M M Anatomy of Downturns – 2018 vs. 2016 There are evident parallels with the downturn of 2016 and not 2018: Memory makers cannot control DRAM demand, which is difficult to predict – but they can control supply, which is influenced by the perception of where future demand growth is headed rather than supply discipline. Memory manufacturers have to guess years in advance what the market size will be, without knowing how demand will grow, to ensure they get a share of the end-market. This is similar to the classic example of the 'cobweb' model, which describes production delays in the agriculture industry arising from the long period between seeding and harvest. When it comes to DRAM and NAND, it takes a long time (two years) to complete manufacturing facilities, and the delay between investment decisions and actual production causes cycles to repeat. Downturns tend to be far more painful when led by a perceived "Super Cycle" (such as data centers) – the risk of missing out on key markets like AI or 5G is far worse for companies than having too much product at too low a price when making investments. And the bigger the perceived demand – such as in the 1993-96 PC boom and similarly, the 2017-18 data center server boom – the greater the excess supply and the harder the fall. In comparison, the 2016 downturn lacked any new demand drivers and had the same demand dynamics (PC demand weakness) and reasons for limited supply growth, resulting in a downturn that was mild. Exhibit 19: QUANTECH: DRAM price delta from equilibrium (grey dots) vs. quant model predicted line: Our model predicts DRAM price to rise until end of 2021 Delta from Equilibrium Supercycle 1 (PC Cycle) Supercycle 2 (Smartphone Cycle) 2.5 Log Supercycle 3 (Cloud Cycle) 2.0 Our model predicts DRAM 1.5 price to rise until end of 2021 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 Source: Morgan Stanley Research. Data as of January-end 2019. Morgan Stanley Research 19 M M The 2015 DRAM downturn lasted five quarters: Revenue declined 34% and ASP fell 41% throughout the downturn (revenues fell 48% in most recent downturn). 2014 (similar to 2021) was a year of high DRAM prices for the desktop PC market as a combination of low production and high demand caused prices to surge. Production issues were caused by an earthquake that struck Taiwan in April 2014, damaging wafer yields for many vendors, as well as a fire at Hynix's Wuxi fab in late 2013 that continued to slow its production. On the demand side, production was focused towards mobile devices rather than the declining PC desktop market putting further pressure on desktop DRAM supply. By 2015, however, that trend reversed with the DRAM industry moving back to oversupply. In particular, supplies of DRAM was excessive on weak PC sales and decelerating smartphone demand, and many manufacturers were forced to adjust their production and greatly reduce average selling prices in order to move parts. How long does it last? Length, depth and rate of change of downturn: We expect this downcycle to be similar to the 2015 downturn – shorter and shallower than the market expects, with an accelerating rate of change through 2022. Looking at the two most important "verticals" – data centers and smartphones – data center server demand growth has accelerated but should still stay relatively robust, and smartphones are in a downturn post the surge in 2021 after maturity and a transition year to 5G. We believe the lion’s share of pain in the consumer tech markets post WFH benefits will be felt throughout 2022, with a full recovery not likely until 2023. Given our view that an average downcycle has lasted seven quarters on average, if we are indeed close to peak into the current DRAM up cycle, any up cycle enthusiasm at present is most likely misplaced. Exhibit 20: Samsung and SK Hynix blended DRAM ASP decline by half during downturns 1.9 6 quarters of downcycle 1.7 7 quarters of downcycle 5 quarters of downcycle 1.5 1.3 1.1 -49% 0.9 0.7 0.5 -46% -61% 0.3 Mar-11 Mar-12 Mar-13 Mar-14 Source: Company Reports, Morgan Stanley Research estimates Mar-15 Mar-16 Mar-17 Samsung Blended DRAM ASP (US$) Mar-18 Mar-19 Mar-20 SK Hynix Blended DRAM ASP (US$) Mar-21 Mar-22 1Q90 3Q91 1Q93 3Q94 1Q96 3Q97 1Q99 3Q00 1Q02 3Q03 1Q05 3Q06 1Q08 3Q09 1Q11 3Q12 1Q14 3Q15 1Q17 3Q18 1Q20 Exhibit 21: DRAM cycles perspective in the last 30 years 30 25 DRAM Rev (US$ bn) 20 15 2008 GFC 10 5 0 Source: Morgan Stanley Research, WSTS Note: the current DRAM down cycle started in 3Q18; data as of Dec. 2018 Downturn Duration (Q) 1996-98 10 2000-03 11 2008-09 2 2010-12 8 2014-16 5 2018-19 12 Rev -77% -61% -51% -41% -34% -48% ASP -95% -87% -46% -67% -41% -59% 20 M M DRAM supply-driven cycles: While the DRAM market grows larger every year in terms of bits, when measured in dollars, the market remains a boom-bust cycle – hence the volatility is not in the demand, but in the price, largely dictated by the supply side. Exhibit 22 shows that DRAM demand grows at a pretty steady pace that does not vary significantly, when compared to price per gigabyte history of the same DRAM gigabytes. There is very little demand response to these dramatic price swings – when DRAM prices plummet, demand does not spike in response. Likewise, when prices are stable, demand continues to grow despite the cost to OEMs. The transition to DDR5 should constrain supply growth as a typical die will be 10-15% larger depending on the production node. The bigger die is due to the addition of an error correction code (ECC) imbedded in the chip, which is a buffer function added in case of read/ write defects. This ECC function was previously included at the module levels for DDR4 but will be part of the DRAM die from DDR5. The adoption of DDR5, however, will be limited to 5-10% of overall mix in 2022 given that its application to Intel's Eagle Stream platform (Sapphire Rapids CPU) is delayed from 1Q22 to 3Q22. Exhibit 22: DRAM demand grows every year in terms of bit shipments (1Gb eq)... 180,000 160,000 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0 200320042005200620072008200920102011201220132014201520162017201820192020 DRAM Shipment (mn 1Gb eq) Source: WSTS, Morgan Stanley Research 11/1/2004 3/1/2006 7/1/2007 11/1/2008 3/1/2010 7/1/2011 11/1/2012 3/1/2014 7/1/2015 11/1/2016 3/1/2018 7/1/2019 Exhibit 23: ...but DRAM prices are dictated by supply and hence the DRAM cycle 100% 80% 60% 40% 20% 0% DRAM -B2i0t%ASP YoY (%) -40% -60% -80% -100% Source: WSTS, Morgan Stanley Research DRAM Bit ASP YoY (%) Exhibit 24: DRAM in the context of the semiconductor cycle – revenue YoY is highly correlated for memory with DRAM at a more advanced stage relative to NAND 200% 80% 150% 60% 100% 40% 50% 20% 0% 0% -50% -20% -100% Oct-02 Oct-03 Oct-04 Oct-05 Oct-06 Oct-07 Oct-08 Oct-09 Oct-10 Oct-11 Oct-12 Oct-13 Oct-14 Oct-15 Oct-16 Oct-17 Oct-18 Oct-19 Oct-20 Source: SIA, Morgan Stanley Research DRAM NAND Global Semiconductor(RHS) -40% Morgan Stanley Research 21 M M 2) Growth Indicators – Not Supported by Data We look for proof of 2022 demand acceleration beyond seasonal strength: We believe the foundry sector will be important to watch, since its lead time on product cycles is the longest. We emphasize that memory cycles have often been independent of tech seasonality and global downturns (e.g., 2004, 2015). What drives most memory cycles is supply and the memory sales cycle is often independent of economic cycles, but an economic downturn can exacerbate a downturn. Previous downturns were driven less by changes in demand but rather by problems bringing in supply as quickly as people were expecting. Nevertheless, global growth is an important driver for markets and our economics team forecasts that it will remain strong into 2022. In the near term, however, PMIs have been peaking, trade data have been moderating, and earnings estimate revision numbers have been decelerating across every major vertical for semis. As part of our PMI playbook, we have shown in previous PMI downcycles since 1999, the derating has been more acute at the Cyclicals with multiples compressing in the nine months following a PMI peak. A 2H recovery happens every single year...: There is no historical correlation between equity returns on memory stocks and seasonal strength, even over long time periods. The debate now moves to what extent companies can beat a raised bar in 2H21 and in terms of consensus earnings forecasts. If we listen to the companies themselves, most have not said this is the peak and that things are plateauing. Rather, most are saying that things should get better. ...and the Fed cannot change earnings or the DRAM cycle: Global growth is likely to rise above pre-COVID path in 3Q21 driven by a much sharper recovery in global capex. Our economics team expects that this economic cycle will be normal, strong and short and forecasts 6.5% global GDP growth for 2021 and 4.9% in 2022. Policy support, elevated excess savings (especially in the US) is an important cushion amid inventories at historical lows. Exhibit 25: The DRAM cycle has no correlation with seasonality – price, cycle and stock price 80% 60% 40% 20% 0% -20% -40% -60% -80% Jan-14 Jul-14 Jan-15 Jul-15 Jan-16 Jul-16 Jan-17 Jul-17 Jan-18 Jul-18 Jan-19 Jul-19 Jan-20 Jul-20 Jan-21 Shipment Growth YoY Source: DRAMeXchange, WSTS, Thomson Reuters, Morgan Stanley Research Bit ASP Growth YoY Avg DRAM Stock Perf. YoY (RHS) 120% 100% 80% 60% 40% 20% 0% -20% -40% -60% 22 M M DRAM market outlook worsened recently Our recent channel checks in memory suggest that customers have built sufficient inventory ahead of aggressive pricing from producers, despite the decline in producers' inventory exiting 2Q21. Meanwhile, the magnitude of price hike is higher than previous market expectations following a very strong 2Q (high teens blended DRAM ASP), and inventory stocking is likely to shift to producers soon, raising risk of an inventory correction into 2022. The rate of increase in DRAM spot prices (DDR4 8Gb) decelerated in the last month (US$4.86) after increasing 30% in 1Q21, and currently at a 3.8% premium to contract prices. Contract prices for server DRAM increased by 8% in July alone and likely a 10% increase for 3Q, and most likely smaller increase in 4Q, although visibility is limited by the lack of transactions in the market. PC DRAM also increased 7% in July, and we expect Q3 PC DRAM to be up another 8-9% and mobile DRAM up near ~20% given the price per gigabit gap vs. server DRAM that has widened to more than 20%. Demand for chips is moderating with lack of component availability: Customers have been constrained in certain types of components and overstocked on others. Constraints for PC (audio codec, ethernet) and server (lockdowns in Malaysia affecting the back end) are still widespread, causing demand to slow in the near term and customers to build sufficient DRAM inventory. This makes suppliers vulnerable if sell-through falters, which could eventually result in an air pocket for memory demand. Separately, smartphone data points have remained relatively weak in China, along with parts of consumer (TV) and graphics DRAM. Cloud demand looks like it will grow strongly throughout the year and be more durable as cloud providers start to spend more countercyclically and consume a lot of DRAM inventory in the quarter. Exhibit 26: DDR4 8Gb - 1Gx8 2666Mbps contract prices now 3.6% below spot price Source: DRAMeXchange, Bloomberg, Morgan Stanley Research Exhibit 27: NAND 64Gb MLC spot price declined 12% since May and now at 27% discount to contract price $6.5 $5.5 $4.5 $3.5 $2.5 $1.5 2013 2014 2015 2016 2017 2018 2019 Spot price Contract price Source: DRAMeXchange, Bloomberg, Morgan Stanley Research 2020 2021 Morgan Stanley Research 23 M M End-Market Dynamics We expect weaker end smartphones and PCs, while server markets remain strong given the following: No sign of recovery in smartphone demand and rising pricing pressure: We expect 2022 to be a challenging year for high-end smartphones given the strong transition to 5G behind and strength in 2021. The high price points in a mature smartphone industry did not bode well for unit shipments, especially in China. Meanwhile, China's June smartphone shipments were down 9% YoY, to 25mn units, and despite high-end smartphone price cuts driving some volume pickup, we have seen increased pricing pressure for components. Servers – cloud strength in an accelerating cloud capex environment: Cloud capex looks like it will grow strongly through the year following a brief period of digestion and could to be smoother going forward as cloud providers start to spend more countercyclically. Demand conditions have remained strong with restocking held back by shortages of some key components such as PMIC. The sequential pickup in Aspeed's June sales suggests upside to an already strong cloud demand. The 2H demand is stronger across the board, and recently some US customers even started to place orders for 1H22. Exhibit 28: Smartphone unit growth is expected to accelerate to 7.7% in 2021 following the 6.7% decline in 2020 Exhibit 29: Total smartphone shipments in China reached 25.2M units in June 2021, down 9% Y/Y. Source: IDC, Morgan Stanley Research Source: CAICT, Morgan Stanley Research 1Q17 2Q17 3Q17 4Q17 1Q18 2Q18 3Q18 4Q18 1Q19 2Q19 3Q19 4Q19 1Q20 2Q20 3Q20 4Q20 1Q21 2Q21e 3Q21e 4Q21e 1Q14 2Q14 3Q14 4Q14 1Q15 2Q15 3Q15 4Q15 1Q16 2Q16 3Q16 4Q16 1Q17 2Q17 3Q17 4Q17 1Q18 2Q18 3Q18 4Q18 1Q19 2Q19 3Q19 4Q19 1Q20 2Q20 3Q20 4Q20 1Q21 2Q21e 3Q21e 4Q21e 1Q22e 2Q22e 3Q22e 4Q22e Exhibit 30: Excluding Amazon, Cloud Capex growth is expected to grow in the mid-20% range throughout 2021 Cloud Capex Growth (Y/Y) - Weighted Avg. Ex. Amazon 100% 80% 83% 72% 60% 40% 20% 0% 28% 20% 3% 5% -20% 43% 35% 26% 26% 26% 17% 3% 9% 5% 6% 1% -10% -9% -10% Source: Company data, Thomson Reuters, Morgan Stanley Research, irs.gov. Note: Cloud capex includes capex from Alphabet, Microsoft, Facebook, Tencent, Baidu, Apple, IBM, and Oracle. Forward estimates include MSe for Tencent and Baidu, consensus est. for the others. Exhibit 31: While cloud capex spending patterns should continue to ebb and flow, the general trajectory is up and to the right as hyperscalers invest to support growing cloud demand Top 10 Cloud Providers: Total Quarterly Capex ($M) and Q/Q Growth 45,000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 - 16% 11%e 12%e 17% -9%-1%e4%e -9%e-2%e1%e 14% 11% 10% -3% 12% 3% -4% 9% 12% 10% -25% 23% 17% 23% 5% -15% 4% 10%12% -9% 7% 14%10%-17%12% -15% Source: Company data, Thomson Reuters, Morgan Stanley Research, irs.gov. Note: Cloud capex includes capex from Alphabet, Amazon, Microsoft, Facebook, Tencent, Baidu, Apple, IBM, and Oracle. Forward estimates include MSe for Tencent and Baidu, consensus est. for the others. Note: 24 M M US PC channel inventories continues to rise, driven by HPQ. Our proprietary US PC channel inventory tracker shows that as of the week ending July 18th, total PC units ex-Samsung are now 57% above pre-COVID levels and at a more than two year high with total notebook units (ex-Samsung) 76% above pre-COVID levels. A deeper look at PC market channel inventory increases by SKU suggests ~95% of the total increase over the past three weeks is due to Chromebooks, signalling slower demand in the consumer and education markets while commercial inventory levels still appear low. Overall PC lead times have come down over the past three weeks to 26 days but remain elevated from March quarter levels. Exhibit 32: PC units are expected to grow strongly in 2021 but decelerate in 2022 Exhibit 33: Notebook PC levels excluding Samsung are now 57% above pre-COVID levels Source: Morgan Stanley Research Source: Morgan Stanley Research Morgan Stanley Research 25 M M 3) Earnings Estimate Revisions – Decelerating Taking into account the risk to mean reversion on DRAM pricing forecasts and the recent deceleration in consensus earnings growth expectations, the revision breadth is likely to get significantly worse before bottoming. History tells us to be patient in these instances as jumping the gun too early when earnings estimate revisions momentum is moving toward negative territory can be dangerous. And with stocks up 70-115% from the lows, we prefer to wait for significantly depressed earnings revisions before adding new money. Pricing and margin mean reversion: We continue to believe that margins will decline at a faster rate of change through 2022 as excess capacity and inventory impacts lead to pricing pressure. Pricing is approaching peak levels in server DRAM, which should be close to the 3Q18 peak in terms of weighted average density. The risk in the near term is that DRAM prices mean-revert back to the 2019 lows (current ~US$160 in 3Q21 could potentially fall back near US$100 as prices tend to gravitate toward costs over time) with both customers and DRAM makers looking to convert inventory into cash, since prices falling sharply mean the value of DRAM declines. Exhibit 34: FY2 earnings estimate revisions vs. share price performance – decelerating trends since Feb 150% 200% 100% 50% 0% -50% 150% 100% 50% 0% -50% -100% -100% FY2 Earnings Revision Source: Thomson Reuters, Morgan Stanley Research Share Price Perf. YoY (RHS) Exhibit 35: DRAM prices tend to follow cost declines – 1Gb equiv. blended ASP are above trend line (US$) 100.0 10.0 DRAM 1Gb equiv. Blended ASP Blended DRAM ASP PC DRAM contract price 1.0 Exhibit 36: Samsung DRAM margins mean reversion scenario as prices move closer to trend line 90% 70% 50% 30% Average @30.6% 10% -10% -30% 1Q02 1Q03 1Q04 1Q05 1Q06 1Q07 1Q08 1Q09 1Q10 1Q11 1Q12 1Q13 1Q14 1Q15 1Q16 1Q17 1Q18 1Q19 1Q20 1Q21 0.1 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 (CY) Source: WSTS, Morgan Stanley Research Samsung DRAM OPM Source: WSTS, Company data, Morgan Stanley Research estimate 26 M M What we learned in 2Q – every memory company beat and raised, but shares sold-off: Growth has gone from +7% to +13% in one quarter. Companies' forward guidance is up sharply with hope for a 2H21 demand recovery. The greater risk to numbers lies with 2022 estimates, in our view – they have not adjusted because of the perceived near-term recovery. 2022e earnings have growth of +20% for companies that have reported (appear high), with the top line accelerating to +13% growth. Consensus earnings estimates for Asian memory continued to drop in 2022: The breadth of Asian semi earnings revisions has decreased to 27% since its peak of 59% in April 2021. The benefits from the work-from-home situation due to the pandemic has boosted companies share performance in 1Q21 but concerns over excess supply have turned analysts to become incrementally bearish in their earnings forecasts. On an absolute basis, revisions for West Digital suffered the most in the past 30 days due to weakness in both hard disk drives and NAND. Relative revisions breadth for memory is all negative, meaning revisions for these segments fared even worse than those for the market as a whole. Reducing numbers and price targets. We reduce estimates for Samsung Electronics, SK Hynix and Nanya Tech in our coverage, mostly reflecting lower pricing expectations and shipments (owing to significantly higher inventory build) entering late 2021, leading to a reduction in price targets. We lower our 2022 revenue by 4% and EPS estimates by 12.5%, on average. Our 2022 EPS forecasts are 6% below consensus numbers for Samsung and 16% below for SK Hynix, on average. Exhibit 37: DRAM NTM EPS growth vs. earnings estimate revision breadth 300% 250% 200% 150% 100% 50% 0% -50% -100% Nov-10 Nov-11 Nov-12 Nov-13 Nov-14 Nov-15 Nov-16 Nov-17 Nov-18 Nov-19 Nov-20 100% 80% 60% 40% 20% 0% -20% -40% -60% -80% -100% DRAM Stocks Avg. NTM EPS YoY Growth DRAM Industry Earnings Revision Breadth (RHS) Source: Thomson Reuters, Morgan Stanley Research Exhibit 38: SK Hynix consensus EPS vs. MS forecast 25,000 20,000 15,000 10,000 MSe 2022 EPS MSe 2021 EPS 5,000 0 Jan-19 Jul-19 Jan-20 Jul-20 2020 2021 Source: IBES, Morgan Stanley Research 2022 Jan-21 Mse 2021 Jul-21 Mse 2022 Morgan Stanley Research 27 M M 4) Valuation – Less Compelling Memory stocks are essentially pricing-in a peak: It is no longer about whether it is cheap in absolute terms, it's about whether one is over- or underpaying for stability. Valuations (12-month forward P/E) have quickly snapped back to 7.2-16.4x P/E for memory stocks, 1.5-3.2x EV/Sales and 1.6-2.1x P/B. We favor high P/E multiples and low/trough P/B and EV/sales multiples. Being OW on these stocks now is like saying rates will fall further – and that may be a big risk to take. Capital returns are at risk from declining FCF. Over the long term, we want to buy stocks offering dividends that provide yield or are below the respective country 10-year yield but are growing (not declining). We would avoid yields that are too high (above 6%) because that is the point at which risk of a cut starts to grow. P/E valuation multiples for memory stocks are contrarian, i.e., it is better to buy them at high earnings multiples, which occur closer to cyclical troughs, in the anticipation that the multiple will compress as the earnings rise with the cycle. As such, investors need to consider the high starting point for margins and brace for significant earnings declines and much higher P/E multiples from here, in our view. Exhibit 39: Memory sector P/B valuation 2.5 2.5x 2.0 1.5 1.0 0.5 Dec-09 Dec-11 Dec-13 Memory Sector Trailing PB Source: Bloomberg, Morgan Stanley Research Dec-15 Avg 1.0x Dec-17 +1 s.d. Dec-19 -1 s.d. Exhibit 40: Average memory EV/Sales reversed from its peak 2.7 2.5 2.6x 2.3 2.1 1.9 1.7 1.5 1.4x 1.3 1.1 0.9 0.7 Dec-10 Dec-11 Dec-12 Dec-13 Dec-14 Dec-15 Samsung SK Hynix Avg EV/12M Fwd Sales Dec-16 Avg Dec-17 Dec-18 +1 s.d. Dec-19 -1 s.d. 0.8x Dec-20 Dec-21 Source: Bloomberg, Morgan Stanley Research Exhibit 41: Memory stocks P/E mean reversion... 21.0 18.0 15.0 12.0 9.0 6.0 3.0 Exhibit 42: ...as well as relative P/E 1.4 1.2 1.0 0.8 0.6 0.4 0.2 Jan-13 Jun-13 Nov-13 Apr-14 Sep-14 Feb-15 Jul-15 Dec-15 May-16 Oct-16 Mar-17 Aug-17 Jan-18 Jun-18 Nov-18 Apr-19 Sep-19 Feb-20 Jul-20 Dec-20 May-21 Jan-13 Jun-13 Nov-13 Apr-14 Sep-14 Feb-15 Jul-15 Dec-15 May-16 Oct-16 Mar-17 Aug-17 Jan-18 Jun-18 Nov-18 Apr-19 Sep-19 Feb-20 Jul-20 Dec-20 May-21 Memory PE Avg Source: Bloomberg, Morgan Stanley Research +1 s.d. -1 s.d. Memory Rel. PE Avg +1 s.d. Source: Bloomberg, Morgan Stanley Research. Note: Relative PE vs. MSCI AC Asia -1 s.d. 28 M M Stocks and market sentiment no longer discount a positive scenario: Book value multiples have retreated but are still above historical mid-cycle levels. However, this in itself is insufficient without a second-derivative cyclical catalyst – an earnings downtick – for the market to call a fundamental downturn. The sector is not at distressed levels – one year forward EV/Sales is currently at 2.3x, well above the historical average of 1.4x and the historical trough of 0.6x – following the large earnings and FCF upcycle seen YTD. An alternate path to our bull case would be for DRAM and NAND producers to adjust production more aggressively, thereby maintaining supply-demand equilibrium into 1H22. The bear case – long down cycle similar to 1996, politics and trade: The bear case of our economists’ narrative is a weaker global growth amid a prolonged COVID-19 outbreak and inability to keep up with the pace of variants. And, more important, key drivers of this scenario are US corporate sector risks and a deterioration in COVID related outbreaks. Where We Could Be Wrong – Bull and Bear Scenarios The signposts and recommendations above outline our thinking around the base case (60% subjective probability). But, as always, there are risks around this narrative – bullish and bearish. The bull case – a very short cycle: Cyclical inflections in the midst of high inventory are rare outside a supply adjustment such as fire or price levels reaching cash costs. In our bull case, cloud demand comes roaring back in 1H22 against the backdrop of a stronger-than-expected rebound in global growth. Key signposts to watch in this context would be cloud capex growth improvements that could improve the long-term outlook and/or earnings trends holding up better than our expectations. As always, capacity expansion and production (supply) growth are another path to our bear case. If trade dominates, the pain will be routed through the corporate sector, while sentiment would synchronize with current weak fundamentals. In both cases, the alignment of technicals and fundamentals is likely to deteriorate significantly for memory. We model multiples to test the bottom end of the GFC in this scenario. We do not see China's entry in memory as having a tremendous impact on the DRAM market as the companies have not been executing well. At the very least, not the leading-edge DRAM market because their technology will be anywhere from one to two generations behind what leading memory companies are producing today. Morgan Stanley Research 29 M M QuanTECH – Tactical View on Memory Gilbert Wong QuanTECH framework also suggests to turn cautious on Memory players: We share a similar view to our sector analysts that we are now standing at the late stage of the market cycle to invest in memory stocks because of: (i) memory players used to lead DRAM prices to rotate into down-cycles; (ii) the peaked Asia PMI cycle is elevating the downward earnings revisions risk among Growth stocks, especially memory stocks; and (iii) Asia/EM market is undergoing risk-off trajectory to push valuations lower. We elaborate on these arguments in this section. 1) DRAM price cycle vs. memory players performance: To re-cap what we discussed in May – DRAM - Navigating Complexity – we highlighted that our DRAM price forecasting model expected DRAM price to peak in Dec 2021, while market used to start declining 3-4 months ahead of the cycle peak of DRAM price. Therefore, we suggested investors to take profits from Memory players by end of September the latest. Since that point onward, the development of DRAM price has been largely in line with our model forecasts ( Exhibit 43 ); given we are soon stepping into August, it would be prudent to refresh our recommendation here that it is the time to turn cautious on Memory stocks, and take profits from DRAM players, while staying with NAND exposure for relative resilience in the late-cycle. Exhibit 43: DRAM price delta from equilibrium (grey dots) versus quant model predicted line: We are in the second minicycle of the third supercycle driven by cloud/datacenter demand Delta from Equilibrium Supercycle 1 (PC Cycle) Supercycle 2 (Smartphone Cycle) 2.5 Log Supercycle 3 (Cloud Cycle) 2.0 Our model predicts DRAM 1.5 price to rise until end of 2021 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0 Exhibit 44: DRAM ASP per 1Gb and our model forecasts plotted in log terms Log 4096 1024 256 DRAM ASP (1Gb Eq) Forecasted DRAM Price 64 16 Our model predicts DRAM price to rise until end of 2021 4 1 1/4 Jan-91 Jan-93 Jan-95 Jan-97 Jan-99 Jan-01 Jan-03 Jan-05 Jan-07 Jan-09 Jan-11 Jan-13 Jan-15 Jan-17 Jan-19 Jan-21 Source: Worldwide Semiconductor Trading Statistics, Morgan Stanley Research Source: Morgan Stanley Research 30 M M Exhibit 45: DRAM supplier stocks' absolute performance in the current DRAM cycle versus the average relative performance in the prior six cycles (assuming DRAM cycle peaks in December 2021) Average Absolute Performance of Key Memory Players in Prior 6 Cycles Current DRAM Cycle (Assuming DRAM peaks in Dec-2021) 120 110 tactical absolute return upside: +15% 100 ...but stock price tended to decline 3 months earlier than the peak of DRAM price 90 t=12 t=11 t=10 t=9 t=8 t=7 t=6 t=5 t=4 t=3 t=2 t=1 t=0 t=-1 t=-2 t=-3 t=-4 t=-5 t=-6 t=-7 t=-8 t=-9 t=-10 t=-11 t=-12 80 Before DRAM price peaked 70 After DRAM price peaked Source: MSCI, FactSet, Morgan Stanley Research. Data as of July 30, 2021 2) The peaked Asia PMI is imposing higher risk for NTM EPS cut among memory stocks: The outperformance of memory players in the last one year was partially driven by strong economic momentum post the pandemic lockdown. However, we also observe there is early signal for the Asia PMI to lose momentum since June, which might further drag the market expectation of memory players' earnings, imposing additional downside risk on their stock prices. Exhibit 46 shows the NTM EPS of major DRAM players (namely Micron Tech, Hynix, Samsung, Nanya Tech) had been historically highly sensitive to the Asia PMI cycle. The latest earnings momentum has again been slowed down together with PMI deceleration, reinforcing our latest cautious view on the memory players with high DRAM exposure. Exhibit 46: DRAM players aggregated NTM EPS vs. Asia Pacific ex Japan PMI composite 200% 150% 100% 50% 0% -50% -100% 2011 2012 2013 2014 2015 Memory DRAM - NTM EPS Source: FactSet, IHS Markit, Morgan Stanley Research; data coverage from Jan 2011 to Jul 2021 2016 2017 2018 2019 APxJ PMI Morgan Stanley Research 2020 58 56 54 52 50 48 46 2021 31 M M 3) Risk sentiment is turning to risk-off with elevated valuation pressure amid rising UST 10Y yield: The low rate and global easing environment have significantly lifted the valuation of both market and tech stocks. We flagged Exhibit 47 in May to remind investors that majority return of semiconductor stocks was expectation driven, and contributed by multiplier expansion. Although the improved industry fundamentals met market expectations in 1H2021, the valuation premium that the market had previously priced in should be normalized over time either by further pick-up in EPS or an absolute market correction. We see the current market dynamics is developing toward the latter scenario that Fed's tapering is coming soon, while US Treasury 10Y yield tended to rise than decline in 2H2021 amid high inflation. On the other hand, investors are taking risk off from their portfolio YTD (ex). These effects in combined would have net negative implications on the valuation on semiconductor industry. Given memory players are subject to higher risk of earnings momentum slowdown, we think the embedded downside risk is emerging as we move into late-cycle of DRAM price. Exhibit 47: Long-term relationship between PHLX Semiconductor Index return and the changes of its underlying next-twelve-month EPS 800 The wide valuation gap here is the results of low rate and global easings, which would be gradually narrowed as time goes by 400 200 100 SOX NTM EPS 50 2010 2011 2012 2013 2014 2015 2016 2017 Source: FactSet, IBES, Morgan Stanley Research; data as of July 30, 2021; notes: both index return and EPS are indexed to 100 on January 1, 2013. 2018 2019 2020 2021 Exhibit 48: MS Asia Risk Sentiment Indicator vs. MSCI APxJ - As of July 30, 2021 550 500 Risk Sentiment: Greedy 100% 90% 450 80% MSARSI: 58th 400 70% 350 60% 300 50% 250 40% 200 30% 150 20% 100 Risk Sentiment: Fearful 10% 50 0% 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 MSCI APxJ - LHS Sell Signal Buy Signal Asia Risk Sentiment Indicator - RHS Source: FactSet, Bloomberg, EPFR, MSCI, Morgan Stanley Research. Note: Buy/Sell signals are generated if the average weekly readings of risk sentiment indicator in last month were below 15th percentile or above 85th percentile. The above red/green lines are tracking the forward 6M returns since signal appeared. MSARSI methodology and its other details can be referred to insight #1 in QuantASIA launching report. 32 M M Implications for US Memory Joe Moore We downgrade Micron to EW: While we have been impressed by the structural improvements at the company, and we see limited downside, we see the stock as effectively rangebound in an environment where DRAM prices start to decline. August quarter results should still have upside, and that could persist into November. But the 35% increase in DRAM prices in the last two quarters, with further modest increase expected through year end, creates risks of a steeper decline next year, given the elevated state of customer inventories in some markets. We are trimming our estimates, assuming high single digit blended DRAM price declines in the February and May quarters, and we lower our PT from US$105 to US$75, driven by 10x through-cycle earnings of US$7.50, down from 14x – as we have noted in the past, the multiple on through cycle earnings tends to compress to about 10 during periods of falling prices. Micron has highlighted potential customer shift from "just-in-time" to "just-in-case" inventory levels in some cases; while we think that is a good articulation of what is happening - namely that customers are holding high inventory levels, but that they are willing to move those inventories even higher as overall supply conditions remain tight. That is a general theme across semiconductors at the moment, given rolling outages. The problem we see for DRAM is that modest price declines could snowball quickly, as customers are likely to react to such shifts by slowing purchasing to draw down inventory. We see that as an issue in PCs and in servers, and while we didn't see that as likely in mobile, the company's recent commentary that mobile DRAM is somewhat elevated as well creates some concerns there. Having said that, we do highlight some offsets: Micron's balance sheet inventory is very lean, implying minimal effective supply growth in 2H; this kept us OW until now, as even with some issues in the end markets, the company will effectively see supply contract in 2H (as the inventory drawdown from their own balance sheet in 1H will not recur in 2H); this is why there are still selective price increases in 3Q, and why we thought there could be further increases in 4Q, but recent checks show that isn't the case. We also see continued strength in embedded DRAM, graphics DRAM, and NAND persisting through 4q, as customer inventories in those areas seem much less elevated. That's roughly half of the company's revenue that is still showing improvement, should serve to protect high earnings levels for at least a couple more quarters. Our price target moves from US$105 to US$75. We still view through-cycle earnings potential at about $7.50, but we have commented in the past that multiples on through cycle tend to peak when conditions are improving, and come under pressure during periods of downward revision; our prior target was based on 14x through cycle, but we are lowing that to about 10x, based on historic patterns. While 10x should limit downside, we do think that a weaker multiple is appropriate given that free cash flow per share should run about 40% below reported EPS, on average, given the large disparity between depreciation and capital spending, and given that earnings deceleration is seldom graceful - even in recent years where we have been impressed by the trough earnings power. Changes to Micron estimates – We are making significant changes to our model to reflect worsening DRAM conditions. While 2021 estimates remain unchanged, our 2022e revenue, gross margin and EPS on a non-GAAP basis come down meaningfully from US$36.4bn, 43.8% and US$9.98 to US$33.2bn, 38.3% and US$7.32, respectively. On a ModelWare basis, which includes stock compensation expense, our 2022e EPS comes down from US$9.87 to US$7.18. Exhibit 49: Changes to Micron estimates for 2021 and 2022 Revenue (US$mn) Gross Margin EPS (US$) New $30,306 42.3% $7.41 CY2021e Old $30,306 42.3% $7.41 Source: Morgan Stanley Research estimates % Change 0.0% 0.0% 0.0% New $33,171 38.3% $7.32 CY2022e Old % Change $36,386 -8.8% 43.8% -5.5% $9.98 -26.7% Morgan Stanley Research 33 M M We are relatively more positive in NAND and stay OW WDC. Even though DRAM conditions have started to roll over, NAND is still relatively healthy according to our checks. However, with similar endmarket exposures and an overlapping customer base, cyclicality between the two commodities tends to be somewhat synchronized. Typically, cycles occur within about three months of one another, which has us somewhat cautious on the direction and choppiness of NAND prices. But, there appears to be much less NAND inventory at customers vs. DRAM, given the fact that supply/demand tightened only recently. Also important is that investor sentiment is much more cautious on NAND given recent countercyclical capital spending, and Chinese entry – we note that the implied valuation of WDC's NAND business, is less than 1x sales if we value their hard disk drive business in line with comparables vs. 3.5x for MU – and any incremental positives, such as sector consolidation, capital spending contraction or setbacks with Chinese competition would be an incremental positive for the stock. Exhibit 50: MU: Projected income statement Micron Income Statement ($ in millions; fiscal year ends in Aug) Revenues COGS COGS - MW Gross Profit Non-GAAP Non-GAAP Gross Margin R&D % of Revs. SG&A % of Revs. C4Q16A Nov-16 F1Q17A 3,970 2,959 2,938.0 C1Q17A Feb-17 F2Q17A 4,648 2,944 2,859.0 C2Q17A May-17 F3Q17A 5,566 2,957 2,895.0 C3Q17A Aug-17 F4Q17A 6,138 3,026 2,991.0 C4Q17A Nov-17 F1Q18A 6,803 3,056 3,034.0 C1Q18A Feb-18 F2Q18A 7,351 3,081 3,055.0 C2Q18A May-18 F3Q18A 7,797 3,074 3,047.0 C3Q18A Aug-18 F4Q18A 8,440 3,289 3,261.0 C4Q18A Nov-18 F1Q19A 7,913 3,298 3,243.0 C1Q19A Feb-19 F2Q19A 5,835 2,971 2,907.0 C2Q19A May-19 F3Q19A 4,788 2,960 2,904.0 C3Q19A Aug-19 F4Q19A 4,870 3,475 3,379.0 C4Q19A Nov-19 F1Q20A 5,144 3,778 3,739.0 C1Q20A Feb-20 F2Q20A 4,797 3,442 3,399.0 C2Q20A May-20 F3Q20A 5,438 3,675 3,634.0 C3Q20A Aug-20 F4Q20A 6,056 3,988 3,945.0 C4Q20A Nov-20 F1Q21A 5,773 4,037 3,989.0 C1Q21A Feb-21 F2Q21A 6,236 4,587 4,182.0 C2Q21A May-21 F3Q21A 7,422 4,296 4,237.0 C3Q21E Aug-21 F4Q21E 8,196 4,481 4,401.4 C4Q21E Nov-21 F1Q22E 8,452 4,734 4,654.0 C1Q22E Feb-22 F2Q22E 8,273 4,939 4,858.9 C2Q22E May-22 F3Q22E 8,155 5,151 5,071.5 C3Q22E Aug-22 F4Q22E 8,474 5,304 5,223.9 1,032 26.0% 470.0 11.8% 159.0 4.0% 1,789 38.5% 473.0 10.2% 187.0 4.0% 2,671 48.0% 434.0 7.8% 204.0 3.7% 3,147 51.3% 447.0 7.3% 193.0 3.1% 3,769 55.4% 448.0 6.6% 191.0 2.8% 4,296 58.4% 523.0 7.1% 196.0 2.7% 4,750 60.9% 603.0 7.7% 211.0 2.7% 5,179 61.4% 567.0 6.7% 215.0 2.5% 4,670 59.0% 611.0 7.7% 209.0 2.6% 2,928 50.2% 601.0 10.3% 209.0 3.6% 1,884 39.3% 606.0 12.7% 206.0 4.3% 1,491 30.6% 623.0 12.8% 212.0 4.4% 1,405 27.3% 640.0 12.4% 211.0 4.1% 1,398 29.1% 681.0 14.2% 223.0 4.6% 1,804 33.2% 649.0 11.9% 216.0 4.0% 2,111 34.9% 630.0 10.4% 231.0 3.8% 1,784 30.9% 647.0 11.2% 214.0 3.7% 2,054 32.9% 641.0 10.3% 214.0 3.4% 3,185 42.9% 670.0 9.0% 230.0 3.1% 3,795 46.3% 700.0 8.5% 255.0 3.1% 3,798 44.9% 707.5 8.4% 260.0 3.1% 3,414 41.3% 715.0 8.6% 265.0 3.2% 3,084 37.8% 722.5 8.9% 270.0 3.3% 3,250 38.4% 730.0 8.6% 275.0 3.2% Other OpEx Other OpEx - MW Operating Income - Non-GAAP Operating Margin % 23.0 40.0 438 11.0% 51.0 1,177 25.3% 8.0 53.0 2,071 37.2% (30.0) 54.0 2,546 41.5% 11.0 56.0 3,157 46.4% (16.0) 29.0 3,630 49.4% (44.0) (4.0) 4,017 51.5% (8.0) 32.0 4,439 52.6% 36.0 67.0 3,887 49.1% 97.0 113.0 2,100 36.0% 6.0 58.0 1,110 23.2% (90.0) 40.0 694 14.3% (3.0) 42.0 594 11.5% 11.0 47.0 542 11.3% 10.0 56.0 981 18.0% 50.0 48.0 1,302 21.5% 9.0 58.0 973 16.9% 131.0 52.0 1,257 20.2% 427.0 29.0 2,364 31.9% 53.0 2,895 35.3% 53.0 2,886 34.1% 51.0 2,489 30.1% 53.0 2,146 26.3% 53.0 2,300 27.1% Total Interest Income (net) and Other Income Total Interest Income (net) and Other Income - MW PBT - GAAP PBT - Non-GAAP PBT - MW Income Tax Expense (Benefit) GAAP Income Tax Expense (Benefit) - Non-GAAP Income Tax Expense (Benefit) - MW PAT - GAAP PAT - Non-GAAP PAT - MW Equity Method Income Equity Method Income - Non GAAP adjustment Net Income Extraordinary Net Income-GAAP Net Income- Non-GAAP Net Income - MW Net income attributable to non-controlling interest Net Income Attributable to Micron (GAAP) Net Income Attributable to Micron (Non GAAP) Net Income Attributable to Micron (MW) (146.0) (100.0) 213 338 292 31.0 18.0 18.0 182 320 274 (2.0) 17.0 180 92 180 335 272 180 335 272 (119.0) (122.0) 925 1,055 1,000 38.0 31.0 31.0 887 1,024 969 7.0 894 82 894 1,031 976 894 1,031 976 (226.0) (113.0) 1,737 1,958 1,901 92.0 64.0 64.0 1,645 1,894 1,837 2.0 1,647 192 1,647 1,896 1,839 1,647 1,896 1,839 (181.0) (100.0) 2,321 2,446 2,389 (47.0) 60.0 60.0 2,368 2,386 2,329 1.0 2,369 (39.0) 2,369 2,387 2,330 (1.0) 2,368 2,386 2,329 (305.0) (72.0) 2,792 3,085 3,034 114.0 91.0 91.0 2,678 2,994 2,943 2,678 265 2,678 2,994 2,943 2,678 2,994 2,943 (114.0) (35.0) 3,453 3,595 3,543 143.0 99.0 99.0 3,310 3,496 3,444 1.0 3,311 134 3,311 3,497 3,445 (2.0) 3,309 3,495 3,443 (237.0) (20.0) 3,716 3,997 3,949 (109.0) 97.0 97.0 3,825 3,900 3,852 (2.0) 3,823 27 3,823 3,898 3,850 3,823 3,898 3,850 (31.0) 6.0 4,346 4,445 4,398 20.0 131.0 131.0 4,326 4,314 4,267 - 4,326 (59.0) 4,326 4,314 4,267 (1.0) 4,325 4,313 4,266 14.0 24.0 3,773 3,911 3,850 477.0 400.0 400.0 3,296 3,511 3,450 (1.0) 3,295 154 3,295 3,510 3,449 (3.0) 3,292 3,507 3,446 (53.0) 30.0 1,904 2,130 2,073 280.0 165.0 165.0 1,624 1,965 1,908 1.0 1,625 284 1,625 1,966 1,909 (6.0) 1,619 1,960 1,903 (294.0) 23.0 716 1,133 1,075 (135.0) (75.0) (75.0) 851 1,208 1,150 - 851 299 851 1,208 1,150 (11.0) 840 1,197 1,139 5.0 18.0 655 712 672 71.0 52.0 52.0 584 660 620 2.0 586 36.0 586 662 622 (25.0) 561 637 597 43.0 11.0 561 605 564 55.0 42.0 42.0 506 563 522 2.0 508 16 508 565 524 (17.0) 491 548 507 (13.0) (7.0) 427 535 487 21.0 17.0 17.0 406 518 470 1.0 407 64 407 519 471 (2.0) 405 517 469 (18.0) (12.0) 870 969 921 68.0 29.0 29.0 802 940 892 3.0 805 90 805 943 895 (2.0) 803 941 893 (32.0) (26.0) 1,125 1,276 1,224 136.0 46.0 46.0 989 1,230 1,178 1.0 990 189.0 990 1,231 1,179 (2.0) 988 1,229 1,177 (25.0) (18.0) 841 955 904 51.0 71.0 71.0 790 884 833 13.0 803 43 803 897 846 803 897 846 (28.0) (20.0) 635 1,237 1,182 48.0 125.0 125.0 587 1,112 1,057 16.0 603 470 603 1,128 1,073 603 1,128 1,073 7.0 15.0 1,806 2,379 2,326 65.0 200.0 200.0 1,741 2,179 2,126 (6.0) 1,735 385 1,735 2,173 2,120 1,735 2,173 2,120 (31.0) (31.0) 2,729 2,864 2,811 252.9 252.9 252.9 2,476 2,611 2,558 - 2,476 82 2,476 2,611 2,558 2,476 2,611 2,558 (31.0) (31.0) 2,720 2,855 2,802 252.1 252.1 252.1 2,468 2,603 2,550 - 2,468 82 2,468 2,603 2,550 2,468 2,603 2,550 (31.0) (31.0) 2,323 2,458 2,407 216.7 216.7 216.7 2,107 2,242 2,191 - 2,107 84 2,107 2,242 2,191 2,107 2,242 2,191 (31.0) (31.0) 1,980 2,115 2,062 185.6 185.6 185.6 1,795 1,930 1,877 - 1,795 82 1,795 1,930 1,877 1,795 1,930 1,877 (31.0) (31.0) 2,134 2,269 2,216 199.5 199.5 199.5 1,935 2,070 2,017 - 1,935 82 1,935 2,070 2,017 1,935 2,070 2,017 GAAP EPS EPS Non-GAAP EPS MW EPS $ 0.16 $ 0.77 $ 1.40 $ 1.99 $ 2.19 $ 2.67 $ 3.10 $ 3.56 $ 2.81 $ 1.42 $ 0.74 $ 0.50 $ 0.43 $ 0.36 $ 0.71 $ 0.87 $ 0.71 $ 0.53 $ 1.52 $ 2.17 $ 2.19 $ 1.90 $ 1.64 $ 1.79 $ 0.16 $ 0.77 $ 1.40 $ 2.00 $ 2.19 $ 2.67 $ 3.10 $ 3.56 $ 2.81 $ 1.42 $ 0.75 $ 0.52 $ 0.45 $ 0.36 $ 0.71 $ 0.88 $ 0.71 $ 0.53 $ 1.52 $ 2.17 $ 2.19 $ 1.90 $ 1.64 $ 1.79 $ 0.32 $ 0.90 $ 1.62 $ 2.02 $ 2.45 $ 2.82 $ 3.15 $ 3.53 $ 2.97 $ 1.71 $ 1.05 $ 0.56 $ 0.48 $ 0.45 $ 0.82 $ 1.08 $ 0.78 $ 0.98 $ 1.88 $ 2.27 $ 2.29 $ 2.00 $ 1.75 $ 1.90 $ 0.25 $ 0.84 $ 1.56 $ 1.96 $ 2.40 $ 2.78 $ 3.12 $ 3.51 $ 2.94 $ 1.67 $ 1.02 $ 0.55 $ 0.46 $ 0.42 $ 0.79 $ 1.04 $ 0.75 $ 0.94 $ 1.85 $ 2.24 $ 2.26 $ 1.97 $ 1.71 $ 1.86 Source: Company Data, Morgan Stanley Research FY2018A 30,391 12,500 12,397.0 FY2019A 23,406 12,704 12,433.0 FY2020A 21,435 14,883 14,717.0 FY2021E 27,627 17,401 16,809.4 FY2022E 33,355 20,128 19,808.2 FY2023E 34,054 21,856 21,536.2 17,994 59.2% 2,141.0 7.0% 813.0 2.7% 10,973 46.9% 2,441.0 10.4% 836.0 3.6% 6,718 31.3% 2,600.0 12.1% 881.0 4.1% 10,818 39.2% 2,658.0 9.6% 913.0 3.3% 13,546 40.6% 2,875.0 8.6% 1,070.0 3.2% 12,518 36.8% 2,995.0 8.8% 1,150.0 3.4% (57.0) 113.0 15,243 50.2% 49.0 278.0 7,791 33.3% 68.0 193.0 3,419 16.0% 567.0 192.0 7,489 27.1% 210.0 9,821 29.4% 210.0 8,593 25.2% (687.0) (121.0) 14,307 15,122 14,924 168.0 418.0 418.0 14,139 14,704 14,506 (1.0) - 14,138 367.0 14,138 14,703 14,505 (3.0) 14,135 14,700 14,502 (328.0) 95.0 7,048 7,886 7,670 693.0 542.0 542.0 6,355 7,344 7,128 2.0 - 6,357 773.0 6,357 7,346 7,130 (45.0) 6,312 7,301 7,085 (20.0) (34.0) 2,983 3,385 3,196 280.0 134.0 134.0 2,703 3,251 3,062 7.0 2,710 359.0 2,710 3,258 3,069 (23.0) 2,687 3,235 3,046 (77.0) (76.0) 6,011 7,413 7,201 416.9 648.9 648.9 5,594 6,764 6,552 23.0 - 5,617 980.0 5,617 6,787 6,575 5,617 6,787 6,575 (124.0) (124.0) 9,157 9,697 9,487 853.9 853.9 853.9 8,304 8,844 8,634 8,304 330.0 8,304 8,844 8,634 8,304 8,844 8,634 (124.0) (124.0) 7,929 8,469 8,259 743.3 743.3 743.3 7,185 7,725 7,515 7,185 330.0 7,185 7,725 7,515 7,185 7,725 7,515 $ 11.51 $ 5.52 $ 2.38 $ 4.92 $ 7.52 $ 6.88 $ 11.51 $ 5.56 $ 2.40 $ 4.92 $ 7.52 $ 6.88 $ 11.96 $ 6.35 $ 2.83 $ 5.90 $ 7.95 $ 7.33 $ 11.81 $ 6.24 $ 2.71 $ 5.76 $ 7.82 $ 7.19 Exhibit 51: MU: Projected balance sheet Micron Technology, Inc Balance Sheet ($ in millions; fiscal year ends in Aug) Assets Cash and cash equivalents Short-term investments Restricted cash Accounts receivables, net Inventory Prepaid and deferred income taxes Prepaid expenses and other current assets Total Current Assets C4Q16A Nov-16 F1Q17A C1Q17A Feb-17 F2Q17A C2Q17A May-17 F3Q17A C3Q17A Aug-17 F4Q17A C4Q17A Nov-17 F1Q18A C1Q18A Feb-18 F2Q18A C2Q18A May-18 F3Q18A C3Q18A Aug-18 F4Q18A C4Q18A Nov-18 F1Q19A C1Q19A Feb-19 F2Q19A C2Q19A C3Q19A C4Q19A May-19 Aug-19 Nov-19 F3Q19A F4Q19A F1Q20A C1Q20A Feb-20 F2Q20A C2Q20A C3Q20A C4Q20A May-20 Aug-20 Nov-20 F3Q20A F4Q20A F1Q21A C1Q21A Feb-21 F2Q21A C2Q21A May-21 F3Q21A C3Q21E Aug-21 F4Q21E C4Q21E Nov-21 F1Q22E C1Q22E Feb-22 F2Q22E C2Q22E May-22 F3Q22E C3Q22E Aug-22 F4Q22E $ 4,139.0 30.0 - 2,453.0 2,750.0 132.0 9,504.0 $ 3,633.0 265.0 - 2,891.0 3,000.0 156.0 9,945.0 $ 4,065.0 265.0 - 3,497.0 3,064.0 132.0 11,023.0 $ 5,109.0 319.0 - 3,759.0 3,123.0 147.0 12,457.0 $ 6,008.0 166.0 - 3,876.0 3,160.0 148.0 13,358.0 $ 7,828.0 214.0 - 4,437.0 3,184.0 173.0 15,836.0 $ 6,808.0 263.0 - 4,912.0 3,369.0 147.0 15,499.0 $ 6,506.0 296.0 - 5,478.0 3,595.0 164.0 16,039.0 $ 4,447.0 1,116.0 5,418.0 3,876.0 182.0 15,039.0 $ 6,353.0 1,180.0 4,416.0 4,390.0 211.0 16,550.0 $ 5,157.0 1,532.0 3,257.0 4,905.0 215.0 15,066.0 $ 7,152.0 803.0 - 3,195.0 5,118.0 235.0 16,503.0 $ 6,969.0 619.0 - 3,419.0 4,943.0 217.0 16,167.0 $ 7,118.0 363.0 - 3,049.0 5,208.0 238.0 15,976.0 $ 8,267.0 391.0 - 3,603.0 5,405.0 233.0 17,899.0 $ 7,624.0 518.0 - 3,912.0 5,607.0 304.0 17,965.0 $ 5,985.0 1,047.0 3,691.0 5,521.0 285.0 16,529.0 $ 6,507.0 677.0 - 3,353.0 4,743.0 1,999.0 17,279.0 $ 7,759.0 590.0 - 4,231.0 4,537.0 1,444.0 18,561.0 $ 6,716.2 590.0 - 5,240.1 4,979.3 1,444.0 18,969.7 $ 7,498.2 590.0 - 5,403.9 5,260.0 1,444.0 20,196.1 $ 8,257.0 590.0 - 5,289.5 5,487.6 1,444.0 21,068.1 $ 8,866.1 590.0 - 5,214.0 5,723.8 1,444.0 21,838.0 $ 9,422.0 590.0 - 5,417.9 5,893.2 1,444.0 22,767.2 Property and equipment, net Equity method investments Deferred tax assets, non current Goodwill Intangible assets, net Long-term marketable securities Other non-current assets Total Assets 15,321.0 1,401.0 599.0 445.0 155.0 411.0 27,836.0 19,098.0 38.0 679.0 1,190.0 425.0 589.0 391.0 32,355.0 19,014.0 38.0 679.0 1,190.0 425.0 471.0 427.0 33,267.0 19,431.0 16.0 766.0 1,228.0 387.0 617.0 434.0 35,336.0 20,723.0 - 731.0 1,228.0 368.0 314.0 469.0 37,191.0 21,864.0 - 1,026.0 1,228.0 348.0 520.0 441.0 41,263.0 22,705.0 - 989.0 1,228.0 334.0 487.0 603.0 41,845.0 23,672.0 - 1,022.0 1,228.0 331.0 473.0 611.0 43,376.0 24,807.0 - 842.0 1,228.0 356.0 1,565.0 758.0 44,595.0 26,204.0 - 762.0 1,228.0 350.0 1,614.0 779.0 47,487.0 27,138.0 - 817.0 1,228.0 339.0 1,167.0 533.0 46,288.0 28,240.0 - 837.0 1,228.0 339.0 1,164.0 576.0 48,887.0 29,352.0 - 783.0 1,228.0 333.0 599.0 1,187.0 49,649.0 29,647.0 - 764.0 1,228.0 332.0 586.0 1,115.0 49,648.0 30,081.0 - 775.0 1,228.0 332.0 577.0 1,113.0 52,005.0 31,031.0 - 707.0 1,228.0 334.0 1,048.0 1,365.0 53,678.0 32,229.0 - 726.0 1,228.0 336.0 1,264.0 1,379.0 53,691.0 31,848.0 - 726.0 1,228.0 342.0 1,316.0 1,396.0 54,135.0 32,209.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 55,943.0 32,242.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 56,384.7 33,105.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 58,474.1 33,918.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 60,159.1 34,681.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 61,692.0 35,394.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 63,334.2 Liabilities & Shareholders' Equity Accounts payable Accrued expenses and other current liabilities Deferred income Income taxes payable Equipment purchase contracts Current debt Total Current Liabilities 4,155.0 - 236.0 - 1,155.0 5,546.0 3,801.0 - 289.0 - 1,117.0 5,207.0 3,656.0 - 326.0 - 1,161.0 5,143.0 3,664.0 - 408.0 - 1,262.0 5,334.0 3,766.0 - 416.0 - 1,401.0 5,583.0 4,194.0 - 427.0 - 1,514.0 6,135.0 3,998.0 - 431.0 - 1,454.0 5,883.0 4,611.0 - 284.0 - 859.0 5,754.0 4,200.0 - 591.0 - 398.0 5,189.0 4,062.0 - 665.0 - 2,634.0 7,361.0 3,494.0 - 557.0 - 1,346.0 5,397.0 4,626.0 - 454.0 - 1,310.0 6,390.0 5,408.0 - 447.0 - 462.0 6,317.0 5,077.0 - 508.0 - 237.0 5,822.0 5,364.0 - 491.0 - 330.0 6,185.0 5,817.0 - 548.0 - 270.0 6,635.0 4,856.0 - 559.0 - 273.0 5,688.0 4,550.0 - 560.0 - 323.0 5,433.0 4,427.0 - 738.0 - 297.0 5,462.0 6,224.2 - 738.0 - 297.0 7,259.2 6,575.0 - 738.0 - 297.0 7,610.0 6,859.5 - 738.0 - 297.0 7,894.5 7,154.8 - 738.0 - 297.0 8,189.8 7,366.5 - 738.0 - 297.0 8,401.5 Long-term debt (includes ST debt) Deferred income taxes Long-term deferred income taxes payable Other long-term liabilities Minority interest Shareholders' Equity 8,490.0 - 632.0 848.0 11,308.0 - 705.0 848.0 10,485.0 - 620.0 848.0 9,872.0 - 661.0 848.0 7,644.0 - 571.0 867.0 7,802.0 - 760.0 869.0 12,320.0 14,287.0 16,171.0 18,621.0 22,526.0 25,697.0 5,890.0 - 554.0 869.0 3,777.0 - 681.0 870.0 3,734.0 - 933.0 870.0 3,604.0 - 1,092.0 863.0 3,563.0 - 1,138.0 867.0 4,541.0 - 1,088.0 987.0 5,188.0 - 1,546.0 98.0 5,188.0 - 1,517.0 98.0 6,356.0 - 1,546.0 98.0 6,373.0 - 1,674.0 - 6,356.0 - 1,740.0 - 6,298.0 - 1,741.0 - 6,418.0 - 1,804.0 - 3,933.0 - 1,804.0 - 3,911.6 - 1,804.0 - 3,889.9 - 1,804.0 - 28,649.0 32,294.0 33,869.0 34,567.0 35,323.0 35,881.0 36,500.0 37,023.0 37,820.0 38,996.0 39,907.0 40,663.0 42,259.0 43,388.4 45,148.5 46,570.7 3,868.0 - 1,804.0 - 47,830.2 3,845.8 - 1,804.0 - 49,282.9 Total Liabilities & Equity 27,836.0 32,355.0 33,267.0 35,336.0 37,191.0 41,263.0 41,845.0 43,376.0 44,595.0 47,487.0 46,288.0 48,887.0 49,649.0 49,648.0 52,005.0 53,678.0 53,691.0 54,135.0 55,943.0 56,384.7 58,474.1 60,159.1 61,692.0 63,334.2 FY2017A FY2018A FY2019A FY2020A FY2021E FY2022E FY2023E $ 5,109.0 319.0 - 3,759.0 3,123.0 147.0 12,457.0 $ 6,506.0 296.0 - 5,478.0 3,595.0 164.0 16,039.0 $ 7,152.0 803.0 - 3,195.0 5,118.0 235.0 16,503.0 $ 7,624.0 518.0 - 3,912.0 5,607.0 304.0 17,965.0 $ 6,716.2 590.0 - 5,240.1 4,979.3 1,444.0 18,969.7 $ 9,422.0 590.0 - 5,417.9 5,893.2 1,444.0 22,767.2 $ 11,879.6 590.0 - 5,603.4 6,154.5 1,444.0 25,671.4 19,431.0 16.0 766.0 1,228.0 387.0 617.0 434.0 35,336.0 23,672.0 - 1,022.0 1,228.0 331.0 473.0 611.0 43,376.0 28,240.0 - 837.0 1,228.0 339.0 1,164.0 576.0 48,887.0 31,031.0 - 707.0 1,228.0 334.0 1,048.0 1,365.0 53,678.0 32,242.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 56,384.7 35,394.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 63,334.2 37,746.0 - 822.0 1,228.0 350.0 1,399.0 1,374.0 68,590.4 3,664.0 - 408.0 - 1,262.0 5,334.0 9,872.0 - 661.0 848.0 18,621.0 35,336.0 4,611.0 - 284.0 - 859.0 5,754.0 3,777.0 - 681.0 870.0 32,294.0 43,376.0 4,626.0 - 454.0 - 1,310.0 6,390.0 4,541.0 - 1,088.0 987.0 35,881.0 48,887.0 5,817.0 - 548.0 - 270.0 6,635.0 6,373.0 - 1,674.0 - 38,996.0 53,678.0 6,224.2 - 738.0 - 297.0 7,259.2 3,933.0 - 1,804.0 - 43,388.4 56,384.7 7,366.5 - 738.0 - 297.0 8,401.5 3,845.8 - 1,804.0 - 49,282.9 63,334.2 7,693.1 - 738.0 - 297.0 8,728.1 3,754.7 - 1,804.0 - 54,303.6 68,590.4 Source: Company Data, Morgan Stanley Research 34 M M Exhibit 52: MU: Projected cash flow statement MU, Inc. Statement of Cash Flows ($ in millions; fiscal year ends in Aug) Cash flows provided by (used for) operating activities: Net income Adjustments to reconcile net income to net cash provided Depreciation and amortization Stock-based compensation Tax effect from stock-based compensation Loss (gain) on equity investments, net Write-down of inventories Gain/Loss on disposal of property, plant and equipment Deferred income taxes Non-cash interest expense and other non-cash items Changes in current assets and liabilities: Accounts receivable Inventory Prepaid and deferred income taxes Prepaid expenses and other assets Accounts payable Accrued expenses and other current liabilities Income taxes payable Deferred revenues Other changes in assets and liabilities Net cash provided by operating activities C4Q16A Nov-16 F1Q17A C1Q17A Feb-17 F2Q17A C2Q17A May-17 F3Q17A C3Q17A Aug-17 F4Q17A C4Q17A Nov-17 F1Q18A C1Q18A Feb-18 F2Q18A C2Q18A May-18 F3Q18A C3Q18A Aug-18 F4Q18A C4Q18A Nov-18 F1Q19A C1Q19A Feb-19 F2Q19A C2Q19A May-19 F3Q19A $ 180.0 $ 894.0 $ 1,647.0 $ 2,369.0 $ 2,678.0 $ 3,311.0 $ 3,823.0 $ 4,326.0 $ 3,295.0 $ 1,625.0 $ 851.0 $ 803.0 46.0 2.0 - 1,034.0 55.0 (7.0) (71.0) 1,051.0 - 1,098.0 114.0 (3.0) 99.0 1,119.0 51.0 - 195.0 1,177.0 52.0 23.0 1,178.0 48.0 - 246.0 1,231.4 47.0 (79.0) 1,353.0 61.0 (14.0) 1,295.0 57.0 - 112.0 1,399.0 58.0 - 288.0 (401.0) 139.0 299.0 70.0 1,138.0 (372.0) 35.0 - 100.0 - (263.0) 1,405.0 (291.0) 2,407.0 (878.0) (124.0) 165.0 363.0 3,203.0 (121.0) (37.0) - (261.0) - 12.0 3,636.0 (509.0) (25.0) 354.0 - (262.0) - 227.0 4,348.0 (547.0) (184.0) (55.0) 262.0 (216.0) (294.0) 4,261.0 (557.0) (226.0) 511.0 216.0 (469.0) 5,000.4 189.0 (286.0) (46.0) 257.0 4,809.0 1,013.0 (514.0) (280.0) 320.0 (192.0) 3,436.0 1,171.0 (515.0) (377.0) (125.0) (39.0) 2,711.0 C3Q19A Aug-19 F4Q19A C4Q19A Nov-19 F1Q20A C1Q20A Feb-20 F2Q20A C2Q20A May-20 F3Q20A C3Q20A Aug-20 F4Q20A C4Q20A Nov-20 F1Q21A C1Q21A Feb-21 F2Q21A C2Q21A May-21 F3Q21A C3Q21E Aug-21 F4Q21E 586.0 $ 508.0 $ 407.0 $ 805.0 $ 990.0 $ 803.0 $ 603.0 $ 1,735.0 $ 2,475.6 1,377.0 67.0 60.0 1,296.0 72.0 (32.0) 1,365.0 85.0 6.0 1,422.0 82.0 6.0 1,567.0 89.0 6.0 1,487.0 92.0 7.0 1,549.0 97.0 8.0 1,557.0 96.0 - 454.0 1,587.0 96.0 - 58.0 (213.0) 529.0 (45.0) (186.0) 2,233.0 (208.0) 175.0 178.0 19.0 3.0 2,011.0 312.0 (265.0) 79.0 19.0 (7.0) 2,001.0 (565.0) (196.0) 443.0 (12.0) 38.0 2,023.0 (262.0) (203.0) 25.0 53.0 6.0 2,271.0 251.0 86.0 - (753.0) - (24.0) - 18.0 1,967.0 282.0 543.0 (24.0) 13.0 (14.0) 3,057.0 (873.0) 185.0 468.0 (83.0) 21.0 3,560.0 (1,009.1) (442.3) 1,797.2 - 4,504.3 Cash flows from investing activities: Capital expenditures, net Additions to equity method investments Return of equity method investment Proceeds from sale of PP&E Restricted cash Sale/(Purchases) of securities Acquisitions/dispositions Other investing activities Net cash used for investing activities (1,264.0) - 483.0 - (109.0) (890.0) (1,164.0) - (666.0) (2,634.0) (12.0) (4,476.0) (1,302.0) - 429.0 (873.0) (1,004.0) - (86.0) - (208.0) (1,298.0) (1,956.0) - 453.0 - 69.0 (1,434.0) (2,261.0) - (255.0) - 107.0 (2,409.0) (2,411.0) - (16.0) - 183.0 (2,244.0) (2,251.0) - (18.0) - 140.0 (2,129.0) (2,700.0) - (1,910.0) - 183.0 (4,427.0) (2,649.0) - (105.0) - 262.0 (2,492.0) (2,403.0) - 98.0 - 239.0 (2,066.0) (2,028.0) - 744.0 - 184.0 (1,100.0) (1,943.0) - 744.0 - 10.0 (1,189.0) (2,056.0) - 272.0 - 74.0 (1,710.0) (1,944.0) - (16.0) - 8.0 (1,952.0) (2,280.0) - (585.0) - 127.0 (2,738.0) (2,738.0) - (741.0) - 61.0 (3,418.0) (3,018.0) - 316.0 - 146.0 (2,556.0) (2,259.0) - 3.0 - 175.0 (2,081.0) (1,620.0) - (1,620.0) Cash flows from financing activities: Proceeds from equipment-sale leaseback trans. Cash received from noncontrolling interest Acquisition of noncontrolling interest Distributions to noncontrolling interest Payments on equipment purchase contracts Change in LT Debt Dividends Share Repurchases Other financing activities Net cash used for financing activities (24.0) (172.0) (13.0) (3.0) (212.0) 2,961.0 - 24.0 (384.0) 13.0 (61.0) 2,553.0 175.0 - (1,218.0) - (91.0) (1,134.0) (3,136.0) - 2,527.0 - (35.0) (214.0) (858.0) (133.0) (2,594.0) 1,445.0 (1,282.0) 133.0 (785.0) 514.0 (138.0) (170.0) (3,388.0) 535.0 (3,023.0) (36.0) (3,427.0) 130.0 (3,333.0) (20.0) (577.0) (1,838.0) (2,435.0) (17.0) (128.0) (2,568.0) 3,665.0 952.0 (17.0) (1,671.0) (159.0) (1,847.0) (21.0) (964.0) (2.0) 1,879.0 892.0 (11.0) (1,415.0) 434.0 (992.0) (18.0) (261.0) (159.0) 223.0 (215.0) (20.0) (2,610.0) (44.0) 3,746.0 1,072.0 (14.0) (80.0) (48.0) (40.0) (182.0) (97.0) (84.0) (33.0) (214.0) (26.0) (19.0) 50.0 5.0 (16.0) (1,241.0) 1,029.0 (228.0) (2,485.0) (1,442.2) (3,927.1) Effect of exchange rate changes on cash and cash equivalent (37.0) 12.0 32.0 (19.0) (6.0) 10.0 (8.0) (33.0) (10.0) 9.0 7.0 20.0 (14.0) - 6.0 19.0 27.0 16.0 1.0 - Cash and cash equivalents, beginning of period Change in cash flow due to restatement Net increase (decrease) in cash and cash equivalents Cash and cash equivalents, end of period 4,140.0 (1.0) 4,139.0 4,139.0 (506.0) 3,633.0 3,633.0 432.0 4,065.0 4,065.0 16.0 1,028.0 5,109.0 5,109.0 (15.0) 914.0 6,008.0 6,008.0 9.0 1,811.0 7,828.0 7,828.0 (6.0) (1,014.0) 6,808.0 6,808.0 192.6 (494.6) 6,506.0 6,506.0 4.0 (2,063.0) 4,447.0 4,447.0 1.0 1,905.0 6,353.0 6,353.0 (1.0) (1,195.0) 5,157.0 5,157.0 (50.0) 2,045.0 7,152.0 7,152.0 1.0 (184.0) 6,969.0 6,969.0 73.0 76.0 7,118.0 7,118.0 1,149.0 8,267.0 8,267.0 (13.0) (630.0) 7,624.0 7,624.0 (1.0) (1,638.0) 5,985.0 5,985.0 522.0 6,507.0 6,507.0 1,252.0 7,759.0 7,759.0 (1,042.8) 6,716.2 Source: Company Data, Morgan Stanley Research FY2017A FY2018A FY2019A FY2020A FY2021E FY2022E FY2023E $ 5,090 $ 14,138 $ 6,357 $ 2,710 $ 5,617 $ 8,303.5 $ 7,185.4 3,986.0 215.0 (8.0) 28.0 4,705.4 198.0 385.0 5,424.0 243.0 446.0 (1,651.0) 50.0 - 564.0 - (121.0) 8,153.0 (1,734.0) (472.0) 549.0 (524.0) 17,245.4 2,431.0 (1,528.0) (174.0) 150.0 (160.0) 13,189.0 5,650.0 328.0 (14.0) (723.0) (489.0) 725.0 79.0 40.0 8,306.0 6,180.0 381.0 469.0 - (1,349.1) 371.7 - 1,488.2 - (94.0) - 25.0 13,088.3 6,848.0 384.0 (177.8) (913.9) - 1,142.3 - 15,586.2 7,648.0 384.0 (185.4) (261.3) 326.6 - 15,097.3 (4,734.0) - (269.0) (2,634.0) 100.0 (7,537.0) (8,879.0) - 164.0 - 499.0 (8,216.0) (9,780.0) - (1,173.0) - 868.0 (10,085.0) (8,223.0) - 415.0 - 219.0 (7,589.0) (9,635.0) - (422.0) - 382.0 (9,675.0) (10,000.0) - (10,000.0) (10,000.0) - (10,000.0) 753.0 (35.0) (369.0) 349.0 (206.0) (10,194.0) 2,624.0 (7,776.0) (12.0) (37.0) 4,140.0 16.0 953.0 5,109.0 5,109.0 180.6 1,216.4 6,506.0 (75.0) (3,340.0) (2,729.0) 3,706.0 (2,438.0) 26.0 6,506.0 (46.0) 692.0 7,152.0 (63.0) (4,366.0) (251.0) 4,363.0 (317.0) 11.0 7,152.0 61.0 411.0 7,624.0 (139.0) (3,829.0) (1,442.2) 1,046.0 (4,364.1) 44.0 7,624.0 (1.0) (906.8) 6,716.2 (87.3) (2,793.1) (2,880.3) - 6,716.2 - 2,705.8 9,422.0 (91.1) (2,548.6) (2,639.7) - 9,422.0 - 2,457.5 11,879.6 Morgan Stanley Research 35 M M Risk Reward – Micron Technology Inc. (MU.O) We are Equal-Weight around worsening memory fundamentals PRICE TARGET $75.00 10x through cycle earnings of US$7.50 Consensus Price Target Distribution Source: Thomson Reuters, Morgan Stanley Research $75.00 MS PT $117.62 $172.00 Mean Morgan Stanley Estimates RISK REWARD CHART AND OPTIONS IMPLIED PROBABILITIES (12M) USD 160 $120.00(++5588.0.088%%) ) Prob (>120.00)~5.4% 120 80 $75.91 $75.00(--11..2200%%) ) Prob (> 75.00) ~43.6% 40 $40.00(--4477.3.31%1%) ) Prob (<40.00)~8.7% 0 AUG '20 FEB '21 AUG '21 AUG '22 Key: Historical Stock Performance Current Stock Price Price Target Source: Thomson Reuters, Morgan Stanley Research, Morgan Stanley Institutional Equities Division. The probabilities of our Bull, Base, and Bear case scenarios playing out were estimated with implied volatility data from the options market as of 10 Aug, 2021. All figures are approximate risk-neutral probabilities of the stock reaching beyond the scenario price in either three-months’ or one-years’ time. View explanation of Options Probabilities methodology here EQUAL-WEIGHT THESIS ▪ While the underlying demand trends are strong and producer inventory levels are low heading into a period of seasonal strength, there are some signs of inventory adjustments short term after customers built inventory. ▪ We see demand growth on the back of seasonality, memory elasticity/higher content per unit, and low customer inventories, and very slow supply growth in DRAM given declines in capex. ▪ We continue to believe that memory stocks have a relatively well defined earnings cycle, though highs and lows are likely to be better than they have been historically. Consensus Rating Distribution MS Rating 85% Overweight 15% Equal-weight 0% Underweight Source: Thomson Reuters, Morgan Stanley Research Risk Reward Themes Secular Growth: Market Share: Positive Negative View descriptions of Risk Rewards Themes here BULL CASE $120.00 12x US$10.00 Bull case EPS Strong memory cycle continues through 2021 and into 2022 in both segments. DRAM remains in allocation mode around stronger content in mobile and elevated PC demand. NAND also remains in shortage on stronger enterprise demand. Gross margin improvement continues driven by scale and utilization rates, and cost improvements. Company behavioral changes make investors believe sustainable profits are to be had here; fab utilization remains low and capacity shrinks further. BASE CASE $75.00 10x through cycle earnings of US$7.50 We previously utilized a book value methodology given the fading earnings power during the downturn. However, now that the industry has emerged from the bottom and the company has greater earnings power, we now use an earningsbased methodology with an assumption that the stock can trade to 10x through-cycle earnings of US$7.50. BEAR CASE $40.00 ~1x tangible book value Memory downturn continues but is less severe than before. The precariously high levels of producer inventory throughout the industry mean that any faltering of end demand would put us back into declining trajectory, where cash flow would be zero or worse 36 M M Risk Reward – Micron Technology Inc. (MU.O) KEY EARNINGS INPUTS Drivers 2020 GAAP Revenue ($, mm) 21,435 Non Gaap Gross Margin (%) 31.3 Non-GAAP EPS ($) 2.83 Inventory ($, mm) 5,607 DOI 135.6 2021e 27,627 39.2 5.90 4,979 103.0 2022e 33,355 40.6 7.95 5,893 105.4 2023e 34,054 36.8 7.33 6,155 101.4 CATALYST CALENDAR Date 14 Jan 2022 18 Jan 2022 Event Source: Thomson Reuters, Morgan Stanley Micron Technology Inc Annual Shareholders Meeting INVESTMENT DRIVERS Improved pricing and demand strength drive earnings growth GLOBAL REVENUE EXPOSURE 0-10% Europe ex UK 0-10% Japan 10-20% APAC, ex Japan, China and India Mainland 10-20% North America 50-60% Mainland China Source: Morgan Stanley Research Estimate View explanation of regional hierarchies here MS ALPHA MODELS 4/5 24 Month 4/5 3 Month BEST Horizon MOST Horizon Source: Thomson Reuters, FactSet, Morgan Stanley Research; 1 is the highest favored Quintile and 5 is the least favored Quintile RISKS TO PT/RATING RISKS TO UPSIDE Pricing continues to increase and demand continues to improve Customers continue to demonstrate an appetite to take on inventory around macroeconomic uncertainty RISKS TO DOWNSIDE Pricing can turn quickly; a reversal in the price uptrend would hamper Micron's earnings power and weigh on the stock The current downturns in NAND and DRAM last longer than expected OWNERSHIP POSITIONING MS ESTIMATES VS. CONSENSUS FY Aug 2021e Sales / Revenue ($, mm) 27,480 27,627 27,650 27,798 EBITDA ($, mm) 11,932 Net income ($, mm) 6,508 13,457 13,329 14,006 6,787 6,847 7,068 Inst. Owners, % Active HF Sector Long/Short Ratio HF Sector Net Exposure 63% 2.9x 33.7% 5.90 EPS ($) 5.76 6.25 5.98 Thomson Reuters; MSPB Content. Includes certain hedge fund exposures held with MSPB. Information may be inconsistent with or may not reflect broader market trends. Long/Short Ratio = Long Exposure / Short exposure. Sector % of Total Net Exposure = (For a particular sector: Long Exposure - Short Exposure) / (Across all sectors: Long Exposure – Short Exposure). Mean Morgan Stanley Estimates Source: Thomson Reuters, Morgan Stanley Research Morgan Stanley Research 37 M M Implications for Korea Memory Shawn Kim Samsung Electronics Samsung should outperform in a memory downturn and while also acknowledging that memory is approaching peak cycle, we see optionality on growth and capital returns, which keeps us OW. We can see room for the share to get some credit and appetite in the market to go back into this high-quality play. At 1.6x P/B, 12.4x P/E and 3.8% FCF yield for 2021e, we believe the stock is particularly attractive in an industry context and stay Overweight. Fundamentals are moving well past uncertainty in 3Q; strong top line and margin expansion are clear positives. Uncertainty is more on 4Q and 2022, and reason for lack of price action. What are upside risks to its revenue trajectory, as we move into 2022? 2021 has been a choppier year in DRAM than we expected coming into the back half of the year, but we have been impressed by Samsung’s relative earnings performance, continued willingness to return cash, and the capabilities the company has built in the 5G sector. The current margins are moving toward peak cycle and now almost higher than previous normal cycle peaks, which speaks to Samsung's superior sustainable returns due to technology/cost structure. While we fully expect cyclicality, we also do expect underlying secular growth, and in the context of the other positives we have raised, we would argue that our current 2022 estimates are not inflated peak estimates. We cut our 2022 and 2023 EPS estimates by 10% and 9%, respectively, to reflect an earlier peak in the DRAM cycle by one quarter and same rate of subsequent quarterly price declines. Our recent channel checks with buyers and sellers of memory point to a shift in customer behavior and difficulty moving bits, causing customer inventory to likely stay elevated through the end of the year. The earlier setback in customers' ability to procure components is a case of "Demand Interrupted" as opposed to "Demand Destroyed" and could lead to an extended cycle in 1H22 where we would look for confirmation. l We revise our DRAM ASP assumptions for 2022 to -9% YoY (vs. -2% previously) and unchanged for 2023 to -18% to reflect a deterioration in demand conditions for 4Q21 and pricing moving in line with costs in 2022. We estimate 4Q21 DRAM bit shipments to be 1% Exhibit 53: Samsung tends to outperform in downcycles 5% 2% 0% -5% -6% -10% -8% -15% -20% -25% -30% -30% -35% Rest of DRAM MSCI Tech Source: Bloomberg, Morgan Stanley Research MSCI Asia Samsung QoQ, due to concerns on demand growth and some inventory build in the quarter on the company's efforts to minimize the impact on ASP. Our bit shipment forecast stays at 26% YoY in 2021 reflecting inventory depletion and 19% in 2022 reflecting the impact of likely supply growth adjustments. l Unchanged NAND forecasts. We model a better NAND pricing environment as demand has improved and pricing negotiations have improved post Q2 results. We estimate NAND prices to fall -10% YoY in 2022 and -12% in 2023. The acceleration in price decline for Samsung will be based on a combination of higher market share competition in the enterprise SSD market and demand elasticity that pushes back further due to excess supply and producer inventory in 2022. l We estimate an improvement in smartphone sales moving into 2022, to 296mn units, or 5% higher than 281mn estimates in 2021. Despite the production hiccup in Vietnam and regional lockdown in India due to the pandemic, we think conditions will normalize in 2022 and return to 290-300mn units sales per year. For 2H21, Samsung will introduce new Galaxy Fold models, and we estimate shipments to reach ~6mn units. We estimate IM margin to stay in low-teen percentages throughout 2021-23, backed by scale and better production efficiency. 38 M M Earnings estimate revisions breadth has decelerated and likely move into negative territory, and the EPS revision ratio (3M average) was at 24% in 3Q21 (vs. 11% in 2Q21). We see downside risks to our EPS revisions, given the potential DRAM ASP peak out in 4Q21. With the above changes, we lower our EPS estimates by -3% for 2021, -10% for 2022, which has a knock-on impact of -9% for 2023. Our long-term revenue estimates are largely unchanged, and margin reflect our concerns on near-term earnings due to an earlier cyclical downturn in memory. Price target changes: By reflecting aforementioned estimates changes, we reduce our price target for Samsung common (005930.KS) and preferred (005935.KS) to W89,000 and W77,000, respectively. While the common share price target is derived by residual-income model, the preferred share price target is a result of a 14% discount to the common share, which is in line with historical average discount level. Exhibit 54: Samsung's earnings estimate revisions breadth is moving towards a trough level 100% 120% 50% 0% -50% 80% 40% 0% -40% -100% -80% Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan-17 Jan-18 Jan-19 Jan-20 Jan-21 Earnings Revision Breadth Source: Quantiwise, Morgan Stanley Research Share Price Perf. YoY (RHS) Exhibit 55: Samsung consensus EPS vs. MS forecasts 8,000 7,500 MSe 2022 EPS 7,000 6,500 MSe 2021 EPS 6,000 5,500 5,000 4,500 4,000 3,500 3,000 Jan-19 Apr-19 Jul-19 Oct-19 Jan-20 Apr-20 Jul-20 Oct-20 2020 2021e Source: Thomson Reuters, Morgan Stanley Research 2022e Jan-21 Apr-21 Jul-21 Exhibit 56: Earnings forecast revisions (earnings drivers) (Unit: KRW bn) Sales Semiconductor Display IT & Mobile Consumer Electronics Operating Profit Semiconductor Display IT & Mobile Consumer Electronics OP Margin (%) Semiconductor Display IT & Mobile Consumer Electronics MW EPS (W) Previous 270,099 95,658 32,610 107,394 51,643 55,530 32,787 4,411 14,122 3,808 20.6% 34.3% 13.5% 13.1% 7.4% 6,867 2021E Revised 268,465 93,821 32,610 107,394 51,643 53,694 30,951 4,411 14,122 3,808 20.0% 33.0% 13.5% 13.1% 7.4% 6,636 Change -0.6% -1.9% 0.0% 0.0% 0.0% -3.3% -5.6% 0.0% 0.0% 0.0% -0.6pp -1.3pp 0.0pp 0.0pp 0.0pp -3.4% Source: Company data, Morgan Stanley Research (E) estimates Previous 290,403 114,601 35,164 113,110 52,609 61,455 41,821 3,693 12,692 2,646 21.2% 36.5% 10.5% 11.2% 5.0% 7,396 2022E Revised 285,301 108,869 35,164 113,110 52,609 55,723 36,088 3,693 12,692 2,646 19.5% 33.1% 10.5% 11.2% 5.0% 6,676 Change -1.8% -5.0% 0.0% 0.0% 0.0% -9.3% -13.7% 0.0% 0.0% 0.0% -1.6pp -3.3pp 0.0pp 0.0pp 0.0pp -9.7% Previous 305,488 123,080 36,804 117,969 54,040 63,814 41,665 3,994 14,588 2,432 20.9% 33.9% 10.9% 12.4% 4.5% 7,550 2023E Revised 300,468 117,439 36,804 117,969 54,040 58,173 36,024 3,994 14,588 2,432 19.4% 30.7% 10.9% 12.4% 4.5% 6,856 Change -1.6% -4.6% 0.0% 0.0% 0.0% -8.8% -13.5% 0.0% 0.0% 0.0% -1.5pp -3.2pp 0.0pp 0.0pp 0.0pp -9.2% Morgan Stanley Research 39 M M Exhibit 57: Long-term estimate changes (Unit: KRW bn) Total Semiconductor Display IM (IT&Mobile) CE (Consumer Elec) Revenue CAGR (2019-29E) Prior Revised 8.0% 7.8% 15.2% 14.7% 5.4% 5.4% 4.0% 4.0% 3.5% 3.5% Source: Company data, Morgan Stanley Research (E) estimates Change -0.2pp -0.5pp 0.0pp 0.0pp 0.0pp Average OPM (2019-29E) Previous Revised 21.1% 19.8% 33.7% 31.4% 10.1% 10.1% 11.3% 11.3% 4.9% 4.9% Change -1.3pp -2.3pp 0.0pp 0.0pp 0.0pp Exhibit 58: Samsung consolidated P&L Revenue by Division (KRW tr) Semiconductor Memory DRAM NAND Display IM (IT & Mobile) Mobile CE (Consumer Electronics) VD Appliances Total Revenue 1Q20 17.6 13.1 7.6 5.5 6.6 26.0 25.0 10.3 5.7 4.7 55.3 2Q20 18.2 14.6 8.6 5.8 6.7 20.8 19.8 10.2 5.3 4.9 53.0 3Q20 18.8 14.3 8.2 6.1 7.3 30.5 29.8 14.1 8.2 5.9 67.0 Revenue Breakdown (%) Semiconductor Display IM (IT & Mobile) CE (Consumer Electronics) 32 34 28 12 13 11 47 39 46 19 19 21 Operating Profit by Division (KRW tr) Semiconductor Memory DRAM NAND Display IM (IT & Mobile) CE (Consumer Electronics) Total OP 4.0 5.4 5.5 3.5 5.0 4.9 2.9 3.5 3.4 0.6 1.5 1.4 -0.3 0.3 0.5 2.7 2.0 4.5 0.5 0.7 1.6 6.4 8.1 12.4 Source: Company data, Morgan Stanley Research (E) estimates 4Q20 18.2 13.5 7.8 5.7 10.0 22.3 21.5 13.6 8.5 5.1 61.6 30 16 36 22 3.9 3.6 2.7 0.9 1.8 2.4 0.8 9.0 1Q21 19.3 14.6 8.4 6.2 7.0 29.6 28.6 13.0 7.2 5.8 65.4 29 11 45 20 3.4 3.4 2.5 0.9 0.4 4.4 1.1 9.4 2Q21 3Q21E 4Q21E 1Q22E 2Q22E 3Q22E 4Q22E 22.7 25.9 25.9 28.1 29.0 27.3 24.5 17.9 21.1 21.1 22.8 23.5 21.6 18.8 10.9 12.5 12.7 13.3 13.4 12.0 9.9 7.0 8.5 8.4 9.5 10.1 9.7 8.8 6.9 9.7 9.0 8.1 7.5 10.1 9.5 22.7 27.9 27.2 30.6 24.6 29.1 28.8 21.4 26.6 25.9 29.3 23.3 27.9 27.5 13.4 11.4 13.8 13.2 13.7 11.6 14.1 7.2 6.2 8.6 7.3 7.2 6.2 8.7 6.2 5.2 5.3 5.9 6.4 5.4 5.4 63.7 69.1 70.3 73.5 68.7 71.9 71.2 36 38 37 38 42 38 34 11 14 13 11 11 14 13 36 40 39 42 36 40 40 21 17 20 18 20 16 20 6.9 10.3 10.3 11.1 10.9 8.6 5.5 6.7 9.8 9.6 10.4 10.1 7.7 4.5 5.0 7.0 6.8 7.0 6.6 4.9 2.9 1.7 2.8 2.9 3.4 3.6 2.8 1.6 1.3 1.4 1.4 0.5 0.4 1.4 1.3 3.2 3.4 3.0 3.8 2.7 3.2 3.0 1.1 0.7 0.9 0.5 0.9 0.6 0.6 12.6 16.0 15.8 16.0 15.1 14.0 10.7 2018 86.3 72.4 47.7 24.7 32.5 100.7 96.5 42.1 25.3 16.8 243.8 2019 65.8 50.2 31.6 18.3 31.0 107.3 102.3 44.9 26.2 18.7 230.4 2020 72.9 55.5 32.2 23.1 30.6 99.6 96.0 48.2 27.7 20.5 236.8 2021E 93.8 74.6 44.5 30.1 32.6 107.4 102.6 51.6 29.2 22.5 268.5 2022E 108.9 86.6 48.5 38.1 35.2 113.1 108.0 52.6 29.5 23.2 285.3 2023E 117.4 93.0 47.8 45.2 36.8 118.0 112.6 54.0 30.2 23.8 300.5 35 29 31 35 38 39 13 13 13 12 12 12 41 47 42 40 40 39 17 19 20 19 18 18 44.6 14.1 18.8 31.0 36.1 36.0 43.5 12.5 17.0 29.5 32.7 32.4 33.3 12.9 12.6 21.3 21.4 17.8 10.2 -0.4 4.4 8.2 11.4 14.6 2.6 1.6 2.2 4.4 3.7 4.0 10.2 9.3 11.5 14.1 12.7 14.6 2.0 2.7 3.6 3.8 2.6 2.4 58.9 27.8 36.0 53.7 55.7 58.2 Exhibit 59: Samsung LTM P/B at 1.8x or 12% lower than historical high 2.0x 2.0 1.8 1.8X 1.6 1.4 1.2 1.0 0.8 0.8x 0.6 Dec-09 Dec-10 Dec-11 Dec-12 Dec-13 Dec-14 Dec-15 Dec-16 Dec-17 Dec-18 Dec-19 Dec-20 Samsung Trailing PB Avg +1 s.d. -1 s.d. Source: Bloomberg, Morgan Stanley Research Exhibit 60: Net buying activities of Samsung (W bn) (%) 48,000 60 40,000 58 32,000 24,000 56 16,000 54 8,000 0 52 -8,000 50 -16,000 -24,000 48 Jan-18 Jun-18 Nov-18 Apr-19 Sep-19 Feb-20 Jul-20 Dec-20 May-21 Foreign shares ratio (RHS) Foreign Domestic Retail Source: Bloomberg, Morgan Stanley Research 40 M M SK Hynix The obvious issue now – is the recovery in memory sustainable beyond 4Q? The DRAM cycle has further to run – YoY pricing is accelerating in 3Q, most likely up near 10%, and management has raised its 2021 demand growth forecast from 20% to 25%. Ultimately, the market debate is really around the price of perceived growth and safety. We continue to see a structural growth story in enterprise NAND via Intel's NAND acquisition as particularly attractive, but we fully expect cyclicality in 2022, which is likely to influence current market sentiment. W80,000 (from W156,000), mainly due to lowered assumptions on blended DRAM ASP for 2022-23 as well as de-rating back to trough cycle multiples. We are 14% and 28% below consensus on EBIT for 2022e and 2023e, respectively, while relatively the same for 2021e. We believe the wide variance between our forecasts and consensus arises from the market's: (i) underestimating the impact of excess inventory on demand and hence pricing despite relatively high expectations for cloud and enterprise server DRAM demand driving growth into 2022-23e; and (ii) expectation of seasonal strength in mobile, which is less likely for similar reasons (elevated customer inventory), in our view. We lower our EPS forecasts by 7-27% for 2021-23e: We trim our top-line forecasts to reflect a more negative outlook for DRAM pricing. Comments that demand will be better in 2H are just statements of likelihood, but they do not by themselves change the overall supply demand situation into 2022. Our reduced top-line growth and earnings reflects: (i) falling DRAM blended ASP one quarter earlier; due to (ii) likely falling end-demand from recent changes in customer behavior and rising channel inventories nearterm; and (iii) lowered margin profiles. Our PT is reduced to l We revise our DRAM ASP assumptions lower for 4Q21 to 0% (vs. 7% previously), 2021 at 17.4% (vs. 19.7%), 2022 at -9.8% (vs. -3.6%), and 2023 stays unchanged at -18%. Due to piling inventories at customers and fading WFH tailwinds, we think DRAM prices will peak in 4Q21 and model quarterly declines throughout 2022. We did not change DRAM bit growth estimates, considering continued tech migrations and capacity ramp-ups in its manufacturing facility site in Korea, M16. Exhibit 61: SK Hynix's earnings estimate revisions breadth recently turned negative... Source: Refinitiv, Morgan Stanley Research Exhibit 62: … which point to downside risks to consensus ahead 30,000 (KRW) 25,000 20,000 15,000 10,000 MSe 2022 EPS MSe 2021 EPS 5,000 0 Feb-17 Feb-19 2020 2021 Source:Thomson Reuters, Morgan Stanley Research 2022 Feb-21 Morgan Stanley Research 41 M M Earnings estimate revisions breadth has turned bearish, and the EPS revision ratio (3M average) was at -21% in 3Q21, followed by -5% in 2Q21. From a trading perspective, retail investors have been a key driver of the YTD performance. Hynix's foreign ownership now stands at 48%, which is around 60bps lower YTD. Our 3Q21 and 4Q21 operating profit estimates are in line with the latest Bloomberg con- sensus, while 2022 and 2023 estimates are lower at -14% and -28%, respectively. According to Bloomberg, out of 39 brokers, only 3% (or one broker) rated the stock 'Sell'. Average target price of 39 brokers is W163,000 or 55% upside. Exhibit 63: Key earnings forecast revisions (W bn) Sales DRAM NAND Operating Profit DRAM NAND OP Margin DRAM NAND Net Profit EPS (W) Previous 44,703 32,429 12,274 13,389 13,006 382 29.9% 40.1% 3.1% 10,022 13,767 2021E Revised 43,805 31,794 12,011 12,490 12,371 119 28.5% 38.9% 1.0% 9,349 12,842 Change -2.0% -2.0% -2.1% -6.7% -4.9% -68.9% -1.4pp -1.2pp -2.1pp -6.7% -6.7% Previous 58,099 40,308 17,791 19,126 17,697 1,429 32.9% 43.9% 8.0% 14,389 19,765 2022E Revised 53,950 36,982 16,968 14,977 14,371 606 27.8% 38.9% 3.6% 11,278 15,492 Change -7.1% -8.3% -4.6% -21.7% -18.8% -57.6% -5.2pp -5.0pp -4.5pp -21.6% -21.6% Previous 60,078 39,663 20,415 15,848 14,429 1,419 26.4% 36.4% 6.9% 11,931 16,389 2023E Revised 55,861 36,390 19,471 11,631 11,156 475 20.8% 30.7% 2.4% 8,769 12,045 Change -7.0% -8.3% -4.6% -26.6% -22.7% -66.5% -5.6pp -5.7pp -4.5pp -26.5% -26.5% Assumption Changes DRAM bit growth DRAM ASP growth NAND bit growth NAND ASP growth USD/KRW Rate 23.4% 19.7% 52.4% -2.3% 1,110 23.4% 17.4% 52.4% -4.2% 1,110 Source: Company data, Morgan Stanley Research (E) estimates 0.0pp -2.3pp 0.0pp -1.9pp 0.0pp 21.3% -3.6% 34.5% -7.7% 1,180 21.3% -9.8% 34.5% -10.2% 1,180 0.0pp -6.2pp 0.0pp -2.5pp 0.0pp 20.0% -18.0% 35.0% -15.0% 1,180 20.0% -18.0% 35.0% -15.0% 1,180 0.0pp 0.0pp 0.0pp 0.0pp 0.0pp Exhibit 64: Long-term estimate changes (W bn) Total DRAM NAND Revenue CAGR (2019-29E) Prior Revised Change 17.1% 16.5% -0.6pp 13.6% 12.7% -0.9pp 25.6% 25.5% -0.1pp Source: Company data, Morgan Stanley Research (E) estimates Average OPM (2019-29E) Previous Revised Change 24.0% 21.2% -2.8pp 35.7% 31.1% -4.6pp -1.6% -1.0% 0.7pp 42 M M Our long-term estimates are mostly unchanged, despite our concerns on near-term earnings risks in 2021-22, due to the near-term memory market's cyclicality. While long-term revenue estimates are largely unchanged, we lower our margin estimate by 280bps, mainly from the DRAM business. Exhibit 65: SK Hynix Consolidated P&L (KRW bn) Sales QoQ/YoY 1Q20 7,199 4% 2Q20 8,607 20% 3Q20 8,129 -6% 4Q20 7,966 -2% 1Q21 8,494 7% 2Q21 10,322 22% 3Q21E 12,201 18% 4Q21E 12,788 5% 1Q22E 14,317 12% 2Q22E 14,468 1% 3Q22E 13,439 -7% 4Q22E 11,727 -13% DRAM Sales QoQ/YoY Bit Shipments (1Gb mn) QoQ/YoY Blended ASP (USD) QoQ/YoY 5,051 -5% 11,288 -4% 0.39 3% 6,107 21% 11,513 2% 0.45 15% 5,956 -2% 11,974 4% 0.42 -7% 5,906 -1% 13,291 11% 0.39 -7% 6,214 5% 13,847 4% 0.40 4% 7,699 24% 14,608 5% 0.47 17% 8,808 14% 15,193 4% 0.52 10% 9,072 3% 15,648 3% 0.52 0% 9,924 9% 16,431 5% 0.51 -2% 10,003 1% 17,433 6% 0.49 -5% 9,205 -8% 18,653 7% 0.42 -14% 7,850 -15% 19,400 4% 0.34 -18% NAND Sales QoQ/YoY Bit Shipments (8Gb mn) QoQ/YoY Blended ASP (USD) QoQ/YoY 2,148 33% 14,144 12% 0.13 7% 2,500 16% 14,792 5% 0.14 8% 2,173 -13% 16,123 9% 0.13 -10% 2,060 -5% 17,413 8% 0.12 -8% 2,280 11% 21,069 21% 0.11 -7% 2,623 15% 21,702 3% 0.12 10% 3,393 29% 25,391 17% 0.13 8% 3,715 10% 27,067 7% 0.13 2% 4,392 18% 28,420 5% 0.13 -2% 4,465 2% 30,409 7% 0.12 -5% 4,234 -5% 33,146 9% 0.11 -13% 3,877 -8% 36,129 9% 0.09 -16% Operating Profit QoQ/YoY DRAM OP NAND OP 800 239% 1,376 -576 1,947 143% 2,110 -163 1,300 -33% 1,790 -490 966 -26% 1,447 -481 1,324 37% 1,882 -558 2,695 104% 2,849 -154 4,071 51% 3,764 307 4,401 8% 3,877 524 5,103 16% 4,415 688 5,000 -2% 4,479 521 3,492 -30% 3,471 21 1,383 -60% 2,006 -624 OP Margin DRAM OP Margin NAND OP Margin 11.1% 27.2% -26.8% 22.6% 34.5% -6.5% 16.0% 30.1% -22.6% 12.1% 24.5% -23.4% 15.6% 30.3% -24.5% 26.1% 37.0% -5.9% 33.4% 42.7% 9.0% 34.4% 42.7% 14.1% 35.6% 44.5% 15.7% 34.6% 44.8% 11.7% 26.0% 37.7% 0.5% 11.8% 25.6% -16.1% Net Profit QoQ/YoY 648 -616% 1,263 95% 1,077 -15% 1,767 64% Source: Company data, Morgan Stanley Research (E) estimates 990 -44% 1,988 101% 3,060 54% 3,308 8% 3,838 16% 3,761 -2% 2,631 -30% 1,048 -60% 2018 40,445 34% 2019 26,991 -33% 2020E 31,901 18% 2021E 43,805 37% 2022E 53,950 23% 2023E 55,861 4% 32,220 41% 32,629 20% 0.94 22% 21,427 -33% 39,412 21% 0.46 -51% 23,020 7% 48,066 22% 0.41 -11% 31,794 38% 59,296 23% 0.48 17% 36,982 16% 71,917 21% 0.44 -10% 36,390 -2% 86,300 20% 0.36 -18% 8,225 15% 28,982 36% 0.29 -16% 5,564 -32% 44,063 52% 0.13 -56% 8,881 60% 62,472 42% 0.13 0% 12,011 35% 95,228 52% 0.12 -4% 16,968 41% 128,104 35% 0.11 -10% 19,471 15% 172,941 35% 0.10 -15% 20,843 52% 19,517 1,326 2,713 -87% 5,319 -2,606 5,013 85% 6,723 -1,710 12,490 149% 12,371 119 14,977 20% 14,371 606 11,631 -22% 11,156 475 51.5% 60.6% 16.1% 10.1% 24.8% -46.8% 15.7% 29.2% -19.3% 28.5% 38.9% 1.0% 27.8% 38.9% 3.6% 20.8% 30.7% 2.4% 15,540 46% 2,006 -87% 4,755 137% 9,349 97% 11,278 21% 8,769 -22% Exhibit 66: Hynix's Trailing P/B – less compelling in a cyclical downturn 3.0 2.7x 2.5 2.0 1.6x 1.5 1.0 0.8x 0.5 Dec-10 Dec-11 Dec-12 Dec-13 Dec-14 Dec-15 Dec-16 Dec-17 Dec-18 Dec-19 Dec-20 SK Hynix Trailing PB Avg Source: Bloomberg, Morgan Stanley Research +1 s.d. -1 s.d. Exhibit 67: Net buying activities of SK Hynix (W bn) (%) 4,000 53 3,000 52 2,000 51 1,000 50 0 49 -1,000 48 -2,000 47 -3,000 46 -4,000 45 -5,000 44 Jan-18 Jun-18 Nov-18 Apr-19 Sep-19 Feb-20 Jul-20 Dec-20 May-21 Foreign shares ratio (%) Foreign Domestic Retail Source: Bloomberg, Morgan Stanley Research Morgan Stanley Research 43 M M Our long-term estimates are mostly unchanged, despite our concerns on near-term earnings risks in 2021-22, due to the near-term memory market's cyclicality. While long-term revenue estimates are largely unchanged, we lower our margin estimate by 280bps, mainly from the DRAM business. Exhibit 65: SK Hynix Consolidated P&L (KRW bn) Sales QoQ/YoY 1Q20 7,199 4% 2Q20 8,607 20% 3Q20 8,129 -6% 4Q20 7,966 -2% 1Q21 8,494 7% 2Q21 10,322 22% 3Q21E 12,201 18% 4Q21E 12,788 5% 1Q22E 14,317 12% 2Q22E 14,468 1% 3Q22E 13,439 -7% 4Q22E 11,727 -13% DRAM Sales QoQ/YoY Bit Shipments (1Gb mn) QoQ/YoY Blended ASP (USD) QoQ/YoY 5,051 -5% 11,288 -4% 0.39 3% 6,107 21% 11,513 2% 0.45 15% 5,956 -2% 11,974 4% 0.42 -7% 5,906 -1% 13,291 11% 0.39 -7% 6,214 5% 13,847 4% 0.40 4% 7,699 24% 14,608 5% 0.47 17% 8,808 14% 15,193 4% 0.52 10% 9,072 3% 15,648 3% 0.52 0% 9,924 9% 16,431 5% 0.51 -2% 10,003 1% 17,433 6% 0.49 -5% 9,205 -8% 18,653 7% 0.42 -14% 7,850 -15% 19,400 4% 0.34 -18% NAND Sales QoQ/YoY Bit Shipments (8Gb mn) QoQ/YoY Blended ASP (USD) QoQ/YoY 2,148 33% 14,144 12% 0.13 7% 2,500 16% 14,792 5% 0.14 8% 2,173 -13% 16,123 9% 0.13 -10% 2,060 -5% 17,413 8% 0.12 -8% 2,280 11% 21,069 21% 0.11 -7% 2,623 15% 21,702 3% 0.12 10% 3,393 29% 25,391 17% 0.13 8% 3,715 10% 27,067 7% 0.13 2% 4,392 18% 28,420 5% 0.13 -2% 4,465 2% 30,409 7% 0.12 -5% 4,234 -5% 33,146 9% 0.11 -13% 3,877 -8% 36,129 9% 0.09 -16% Operating Profit QoQ/YoY DRAM OP NAND OP 800 239% 1,376 -576 1,947 143% 2,110 -163 1,300 -33% 1,790 -490 966 -26% 1,447 -481 1,324 37% 1,882 -558 2,695 104% 2,849 -154 4,071 51% 3,764 307 4,401 8% 3,877 524 5,103 16% 4,415 688 5,000 -2% 4,479 521 3,492 -30% 3,471 21 1,383 -60% 2,006 -624 OP Margin DRAM OP Margin NAND OP Margin 11.1% 27.2% -26.8% 22.6% 34.5% -6.5% 16.0% 30.1% -22.6% 12.1% 24.5% -23.4% 15.6% 30.3% -24.5% 26.1% 37.0% -5.9% 33.4% 42.7% 9.0% 34.4% 42.7% 14.1% 35.6% 44.5% 15.7% 34.6% 44.8% 11.7% 26.0% 37.7% 0.5% 11.8% 25.6% -16.1% Net Profit QoQ/YoY 648 -616% 1,263 95% 1,077 -15% 1,767 64% Source: Company data, Morgan Stanley Research (E) estimates 990 -44% 1,988 101% 3,060 54% 3,308 8% 3,838 16% 3,761 -2% 2,631 -30% 1,048 -60% 2018 40,445 34% 2019 26,991 -33% 2020E 31,901 18% 2021E 43,805 37% 2022E 53,950 23% 2023E 55,861 4% 32,220 41% 32,629 20% 0.94 22% 21,427 -33% 39,412 21% 0.46 -51% 23,020 7% 48,066 22% 0.41 -11% 31,794 38% 59,296 23% 0.48 17% 36,982 16% 71,917 21% 0.44 -10% 36,390 -2% 86,300 20% 0.36 -18% 8,225 15% 28,982 36% 0.29 -16% 5,564 -32% 44,063 52% 0.13 -56% 8,881 60% 62,472 42% 0.13 0% 12,011 35% 95,228 52% 0.12 -4% 16,968 41% 128,104 35% 0.11 -10% 19,471 15% 172,941 35% 0.10 -15% 20,843 52% 19,517 1,326 2,713 -87% 5,319 -2,606 5,013 85% 6,723 -1,710 12,490 149% 12,371 119 14,977 20% 14,371 606 11,631 -22% 11,156 475 51.5% 60.6% 16.1% 10.1% 24.8% -46.8% 15.7% 29.2% -19.3% 28.5% 38.9% 1.0% 27.8% 38.9% 3.6% 20.8% 30.7% 2.4% 15,540 46% 2,006 -87% 4,755 137% 9,349 97% 11,278 21% 8,769 -22% Exhibit 66: Hynix's Trailing P/B – less compelling in a cyclical downturn 3.0 2.7x 2.5 2.0 1.6x 1.5 1.0 0.8x 0.5 Dec-10 Dec-11 Dec-12 Dec-13 Dec-14 Dec-15 Dec-16 Dec-17 Dec-18 Dec-19 Dec-20 SK Hynix Trailing PB Avg Source: Bloomberg, Morgan Stanley Research +1 s.d. -1 s.d. Exhibit 67: Net buying activities of SK Hynix (W bn) (%) 4,000 53 3,000 52 2,000 51 1,000 50 0 49 -1,000 48 -2,000 47 -3,000 46 -4,000 45 -5,000 44 Jan-18 Jun-18 Nov-18 Apr-19 Sep-19 Feb-20 Jul-20 Dec-20 May-21 Foreign shares ratio (%) Foreign Domestic Retail Source: Bloomberg, Morgan Stanley Research 44 M M Implications for Greater China Memory Charlie Chan, Daniel Yen, Ray Wu Specialty DRAM (Downgrade NanyaTech to EW; reiterate UW on PSMC) DRAM pricing in 3Q21, we think Nanya Tech's DRAM ASP may turn flat Q/Q in 4Q21. We believe the cycle for specialty DRAM may peak soon in 2H21 we suggest to take profit, and we downgrade NanyaTech to EW (please see Quarterly Earnings to Peak in 2H21; Downgrade to EW). We see signs that specialty DRAM prices might become softer into year-end, and our recent channel checks also indicate some signs of DRAM demand weakness, and DRAM suppliers no longer enjoying the strong bargaining power they had in 1H21. Looking forward, the company’s bit shipments and ASP growth should become mild in the coming quarters, and we thus believe 2H21 should be the peak for NTC's business ( Exhibit 3 ). After its sequential growth of specialty We also see increasing concern from spot price weakness in the near term (declined 8% in the past month) – more risk on specialty DRAM pricing starting from 4Q21. The market has already fully anticipated that the specialty DRAM contract price will catch up to the spot price in 3Q21. While the specialty DRAM spot premium is now only at 4% (vs. 70-90% in 1H21 vs. 25% 1.5 months ago; Exhibit 68 ), this means that it is increasing difficult to have a clear view about the contract price trend into 4Q21. Moreover, the specialty DRAM spot price has further declined 8% in the past month, which means that the higher spot price is not sustainable to its end-clients. Exhibit 68: Consumer/specialty DRAM contract vs. spot price trend – spot premium now at 4% (vs. 25% 1.5 month ago) 4.0 3.5 3.0 (US$) 2.5 2.0 1.5 1.0 Oct-16 Feb-17 Jun-17 Oct-17 Feb-18 Jun-18 Oct-18 Feb-19 Jun-19 Oct-19 Feb-20 Jun-20 Oct-20 Feb-21 Jun-21 DDR3 4Gb 256M*16 Spot Price Source: DRAMeXchange, Morgan Stanley Research DDR3 4Gb 256M*16 Contract Price Morgan Stanley Research 45 M M We have an UW rating on PSMC, as we believe its Pricing Power Won't Last. We believe the market expectation for the stock might be too high as bull camps are bullish on its logic semi business pricing uptrend (55% revenue contribution), while we see emerging risk on its DRAM foundry business (45% revenue exposure) during a DRAM pricing downturn. As PSMC is only working on its DRAM foundry business, its ASP volatility should be only 60-70% of DRAM IDM players. Bear in mind that its DRAM business ASP fell 50% in 2019 during the downcycle. SSD controller IC and module (EW on Phison; Upgrade SIMO to EW) We are EW on Phison as its gross margin near peak: Phison has been aggressive on NAND Controller IC Price (Further Controller IC Price Hike), and this leads to beat on 2Q21 GM and the trend should continue for part of 3Q21. However, we think GM expansion may end in 3Q, and we see some areas of weakness for the NAND market emerging (please see Gross margin near peak – Stay EW). We have upgraded SIMO to EW due to the stronger outlook for 2022 (Share gain with better mix eases our concern on cycle; up to EW): Similar to Phison, SIMO has turned more aggressive on pricing strategy since 2Q21 (Additional foundry supply with improving pricing power). The company has noted its strong business outlook into 2022 from market share gains on the SSD controller business, which the company has received US$1.5bn order backlog for 2022, and our industry checks suggest that SIMO has secured more foundry resources at TSMC, where SIMO is likely to reach revenue of US$1.5bn by 2022. NOR Flash (Downgrade Winbond and Macronix to EW; covered by Daniel Yen) Please refer to our updated industry report also published today Dragged by DRAM. over DRAM suppliers since March 31 last year was mainly due to exposure to other structural opportunities (e.g., MCU). We believe the NOR upcyle will sustain until year-end: We said in June that 3Q NOR pricing could be up by 10-20% Q/Q, with some specs up to 25%, and 4Q pricing could rise further, driven by widened supply/demand gap. We maintain our view and believe NOR pricing could rise mid-single digits Q/Q in 4Q, led by high density NOR. Supply reduction in China and increasing structural demand from PC continues to drive up the pricing. Based on current conditions, we do not expect a sharp inventory correction into 1H22. Prefer GigaDevice (OW) over Macronix (EW) and Winbond (EW): We factor in the better NOR ASP in 2H into our models for all three NOR Flash players. We raise our PT for GigaDevice, and expect its share price to trade up to 68x our 2022e EPS. MCU provides the longer positive for the stock (refer to our global MCU report). Our EPS estimates are 11% and 21% above consensus for 2021 and 2022, respectively. We raise our Winbond earnings estimates but lower targeted multiple (rating down to EW) given NOR/MCU upside offset by the de-rating on DRAM weakness. We cut Macronix to EW from OW as capacity expansion for 3D NAND could lead to earnings declines in 2022 and 2023 with ROE contraction. China's memory localization progress YMTC 128L 3D NAND milestone in 2Q21: YMTC has said its 128L 3D NAND will reach mass production in 2Q21. Our industry checks suggest that it is now running at small volume testing production, and will have wafer-out mass production by May/June 2021. The yield rate is currently 30-40%, and is likely to improve. YMTC has enough funding for first-stage capacity expansion to reach 85kwpm by the end of 2021, while the second-stage funding is an unknown. As YMTC's NAND capacity ramps up, we believe Phison will increase its NAND supply from YMTC. We expect Phison will have 10-15% NAND supply from YMTC, up from ~5% in 2020. NOR vs. DRAM, fundamentals vs multiples: DRAM conditions have been losing steam recently, and our inflection signposts suggest caution from here (please refer to our global note Memory – Late Cycle Brings More Risk Than Reward ). NOR Flash is a sub-segment of memory, so sentiment could also be negatively impacted. That said, we believe pure NOR Flash players could outperform DRAM – 1) on supply, we expect capacity reduction in 2H from China; 2) on the demand front, we see structural demand increase for PC into 2H; 3) as to pricing, we expect NOR pricing Y/Y to further accelerate into 4Q; and 4) share prices usually move simultaneously with NOR pricing Y/Y. The share price outperformance of NOR Flash vendors 46 CXMT's round of Rmb15.6bn capital injection might lead to longer-term industry competition: According to a TechNews report (December 17, 2020), China's ChangXin Memory Technologies (CXMT) successfully conducted a round of capital injection of Rmb15.6bn, mainly from the China Integrated Circuit Industry Investment Fund. Although the 2021 DRAM industry should see easing competition from China in view of CXMT's quality and performance, the capital injection might lead to longer-term industry competition once CXMT's yield rate picks up. Near-term CXMT is facing the same issue – its DRAM modules are not stable (quality and performance). By applying a 60% yield to CXMT’s 19nm DRAM produc- M M tion of 60k-70kwpm, it would account for only 2% of global DRAM supply in 2021. Exhibit 69: China's self-designed DRAM and NAND capacity market share ACMR to show strong growth of +45-50% YoY in 2021; YMTC is the largest client: We think the progress of China's memory fab projects (YMTC, CXMT) and DRAM cycle (Hynix) should be the main growth driver in 2021. YMTC is aggressively increasing its NAND capacity – to reach 80-100k by 2021. This might become a swing factor for NAND prices in 2022, as YMTC will come out with more NAND supply. We remain OW on ACMR, as we believe its technology leadership among its China peers should allow it to gain share amid the growing China memory and logic foundry capex demand. We are UW on AMEC (Advanced Micro-Fabrication Equipment) given the high valuation. Source: Company data, Morgan Stanley Research (E) estimates Exhibit 70: China's self-designed DRAM market share Exhibit 71: China's self-designed NAND market share Source: Company data, Morgan Stanley Research (E) estimates Source: Company data, Morgan Stanley Research (E) estimates Morgan Stanley Research 47 M M Implications for Japanese Semi Cap Equipment Stocks (Yoshikawa-san) Kazuo Yoshikawa, CFA We estimate that in 1H21 the global wafer fab equipment (WFE) market rose +47% YoY, within which memory equipment increased by 50% YoY and non-memory equipment by 42% YoY. Both memory and foundry/logic investment were strong; DRAM capex—which was weak in 2020—recovered, and memory capex grew rapidly. We expect WFE to grow 35-40% in 2021. We currently expect the wafer fab equipment market to remain more or less flat (modest decline if EUV is excluded) in 2022, we look for a modest decline. This is below the sell side consensus (up 5-10% YoY). Exhibit 72: Global wafer fab equipment market (US$bn) 25.0 22.5 Memory Quarterly wafer fab equipment market 20.2 20.5 20.0 17.5 15.0 12.5 10.0 7.5 5.0 2.5 0.0 Non-memory 15.7 16.9 16.5 6.7 3.8 2.9 8.6 4.1 4.5 9.2 3.5 5.7 9.8 11.1 11.7 12.1 13.1 14.5 14.1 13.3 12.9 11.8 11.5 12.7 4.9 13.4 14.4 6.5 4.5 6.1 6.7 6.2 8.5 10.9 9.7 8.3 7.5 5.6 5.5 5.2 5.0 6.5 5.3 5.0 5.0 5.9 4.6 3.6 4.4 5.1 5.4 6.3 6.0 10.8 7.5 8.4 10.4 7.8 7.0 9.5 8.5 8.8 11.8 11.3 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 16 17 18 19 20 21 (CY) Source: Source: Morgan Stanley Research estimates 48 M M The sales weightings by client device of the semiconductor production equipment makers in our coverage are as follows. Wafer fab equipment makers: Seen in global terms, Tokyo Electron looks to have a balanced customer mix in line with the industry average, but it has a high memory weighting compared with SCREEN. (Note that about 25% of Tokyo Electron’s semi equipment sales are in the field solutions business.) SCREEN has only limited dealings with Korean major memory makers, giving memory a low sales weighting. Back-end equipment makers: Advantest and Tokyo Seimitsu both have low memory weightings. DISCO (6146.T) does not disclose memory and non-memory sales weightings, but we suspect that it is not that different from Advantest’s semiconductor test systems. Micronics Japan is highly competitive in memory probe cards, giving memory a high sales weighting. We maintain our OW rating on Tokyo Electron: TEL estimates memory will account for 49% of its new equipment sales in F3/22 (DRAM 22%, NVM 27%). If memory enters a cyclical deceleration phase from 1H22, the risk of memory capex cut/push-out is likely to grow, as in past deceleration phases. Risk/reward profiles for semi cap equipment stocks are likely to deteriorate. However, we think TEL will probably outperform among semi cap equipment stocks. First, given rising capital intensity of advanced devices and trends of localizing chip production, we think the potential downside for WFE markets is less than during previous cycles. Second, field solutions account for 20-25% of the company's semi cap equipment sales, and as the installation base expands we anticipate relatively stable growth in this area. In fact, during the previous memory downturn in F3/20, new equipment sales fell -14% YoY (including DRAM equipment -44%, NVM equipment -55%), but field solutions sales actually increased +6%. We think that even if the WFE market contracts slightly, field solutions will provide earnings support. Third, we estimate TEL's end-F3/22 net cash position at about ¥370bn (~¥2,400/share). The company's target financial model assumes RoE of 30%, and we think it is likely to pursue share buybacks in F3/23 in order to achieve that goal, providing further downside support for the share price. Exhibit 73: Semi cap equipment companies: Revenue breakdown 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 31% 60% 49% 61% 78% 72% 70% 78% 68% 70% 74% 74% 20% 28% 12% 23% 13% 25% 38% 31% 20% 20% 26% 25% 22% 17% 10% 12% 21% 7% 12% 18% 6% 16% 11% 21% 13% 17% 18% 8% 18% 8% 52% 65% 62% FY18 FY19 FY20 FY18 FY19 FY20 FY18 FY19 FY20 FY18 FY19 FY20 CY18 CY19 FY20 TEL SCREEN Advantest Tokyo Seimitsu MJC Source: Company data, Morgan Stanley Research Nonmemory NVM DRAM Morgan Stanley Research 49 M M Risk Reward – Samsung Electronics (005930.KS) We can see appetite in the market to go back into this high-quality play PRICE TARGET W89,000 Base case, residual income (RI) valuation model, cross-checked against P/BV analysis. At our price target, the 2021e P/B multiple would be 1.6x, which is largely in-line with one standard deviation higher than its mid-cycle NTM multiple. We assume an 11.5% cost of equity, based on a beta of 1.0, and our terminal growth rate assumption is 5%. W102,710 Consensus Price Target Distribution W73,200.00 W130,000.00 Source: Thomson Reuters, Morgan Stanley Research MS PT Mean Morgan Stanley Estimates RISK REWARD CHART AND OPTIONS IMPLIED PROBABILITIES (12M) KRW 160000 120000 80000 W80,200 W118,000(++4477.1.133%%) ) Prob (>118,000)~2.5% W89,000(++1100.9.977%%) ) Prob (> 89,000) ~25.0% 40000 W60,000(--2255.1.199%%) ) Prob (<60,000)~14.2% OVERWEIGHT THESIS ▪ Share gains in component businesses (DRAM, NAND, and OLED) continue to drive Samsung's earnings. ▪ Samsung should benefit from memory upcycle, as well as potential share gains and earnings upside from its foundry business. ▪ The smartphone market is mature and no longer growing, but Samsung maintains a 20%+ share with its mass model Galaxy A and Note/S series. Foldable smartphones should enable margin expansion in the IM (mobile) division in 2021. ▪ 3Q21's guidance was excellent for a stock that has now underperformed since 1Q – we can see appetite in the market to go back into this high-quality play, although clear earnings visibility into 2022 is still lacking. Consensus Rating Distribution MS Rating 93% Overweight 5% Equal-weight 2% Underweight Source: Thomson Reuters, Morgan Stanley Research 0 AUG '20 FEB '21 AUG '21 AUG '22 Key: Historical Stock Performance Current Stock Price Price Target Source: Thomson Reuters, Morgan Stanley Research, Morgan Stanley Institutional Equities Division. The probabilities of our Bull, Base, and Bear case scenarios playing out were estimated with implied volatility data from the options market as of 10 Aug, 2021. All figures are approximate risk-neutral probabilities of the stock reaching beyond the scenario price in either three-months’ or one-years’ time. View explanation of Options Probabilities methodology here Risk Reward Themes Secular Growth: Self-help: Positive Positive View descriptions of Risk Rewards Themes here BULL CASE W118,000 HPC drives sustained memory pricing High-performance computing drives sustained memory pricing strength and capital returns exceed expectations: We assume continued robust demand growth for memory and OLED with stabilization in the macro environment and China's failure to successfully enter the memory semiconductor industry. BASE CASE W89,000 Strong DRAM cycle in 2021 but peak out in 4Q21 We maintain an OW rating on Samsung with a constructive view on DRAM. Our QUANTECH model tactically favors DRAM players including Samsung and Hynix over NOR, based on optimal back-tested factors' weightings. From a fundamental perspective, we expect improving memory conditions to help drive relative outperformance. We estimate DRAM ASP to stay flat and peak out from 4Q21. BEAR CASE W60,000 Slower memory consumption, competition from China Prolonged consumption slowdown and Chinese competition: This scenario reflects a more severe Covid-19 impact and correction in US FANG (Facebook, Amazon, Netflix, and Google). Also, Chinese competition disrupts OLED and memory profitability. Weaker macro conditions stall near-term global consumption of IT products. We also assume demand for server content slumps amid a slowdown in data center growth. Faster-than-expected DRAM ASP declines in 2022-23. 50 M M Risk Reward – Samsung Electronics (005930.KS) KEY EARNINGS INPUTS Drivers 2020 Smartphone Shipment Growth (%) (12.1) DRAM Bit Growth (%) 22.7 DRAM ASP Growth (%) (18.1) NAND Bit Growth (%) 26.7 NAND ASP Growth (%) (1.5) 2021e 2022e 2023e 10.2 5.3 2.0 26.2 19.3 20.0 14.3 (8.6) (18.0) 43.4 40.4 35.0 (5.2) (9.9) (12.0) CATALYST CALENDAR Date 27 Oct 2021 01 Nov 2021 Event Source: Thomson Reuters, Morgan Stanley Q3 2021 Samsung Electronics Co Ltd Earnings Release INVESTMENT DRIVERS Supply-demand outlook for memory and OLED Smartphone market share and margins Capital returns of 50% FCF GLOBAL REVENUE EXPOSURE 0-10% India 0-10% Japan 0-10% Latin America 0-10% MEA 0-10% UK 10-20% APAC, ex Japan, China and India Mainland 10-20% Europe ex UK 20-30% Mainland China 20-30% North America Source: Morgan Stanley Research Estimate View explanation of regional hierarchies here MS ALPHA MODELS 2/5 MOST 3 Month Horizon Source: Thomson Reuters, FactSet, Morgan Stanley Research; 1 is the highest favored Quintile and 5 is the least favored Quintile RISKS TO PT/RATING RISKS TO UPSIDE New technological developments, especially in memory and foldable display Longer-lasting memory up-cycle from hyperscale data center growth RISKS TO DOWNSIDE Product and memory cycle, including Apple and new Chinese smartphone competition Earnings growth concentration in semiconductors MS ESTIMATES VS. CONSENSUS FY Dec 2021e Sales / Revenue (W, bn) 268,465 264,160 274,483 283,595 EBITDA (W, bn) 45,944 82,736 92,913 83,359 OWNERSHIP POSITIONING Inst. Owners, % Active 80.4% EBIT (W, bn) 43,807 53,694 55,530 52,787 Source: Thomson Reuters, Morgan Stanley Research EPS (W) 5,355 6,636 5,962 7,059 Mean Morgan Stanley Estimates Source: Thomson Reuters, Morgan Stanley Research SUSTAINABILITY & ESG +1.0 -1.0 Indicator of Change 0.27 Disclosure Rate 82% Morgan Stanley Research 51 M M Risk Reward – Samsung Electronics (005930.KS) SUSTAINABILITY AND ESG INDICATOR OF CHANGE OVER TIME 1.0 -0.07 -0.18 -0.45 SUMMARY All Env Gov In 2019, Samsung Electronics had an ESG indicator of Change of 0.27, which signals that the majority of its ESG data points in our database improved YoY. Key ESG metrics that improved in 2019 were CO2 emissions, total waste, and recycled water. 0.27 -0.33 -1.0 2015 2016 2017 2018 2019 Environment CO2 emissions (Scope 1 and 2) Energy consumption Total waste Waste recycled Water consumption Water recycled Unit kilo tons CO2 equivalent GWh tons Percentage kilo tons Percentage Dec '17 13,585 23,419 1,146,812 95% 120,618 47.00% Dec '18 15,173 26,028 1,210,521 96% 134,230 46.00% Dec '19 13,800 26,899 1,099,197 95% 134,479 51.00% Social Avg training hrs per employee Employee engagement Employee turnover Hours Percentage Percentage 73.50 82% - 62.20 87% - 67.20 89% 19.50% Governance Avg tenure of the board Female board members Females in management Independent board members Years Percentage Percentage Percentage 4.72 - 13.20% 33.33% 1.15 - 14.20% 44.44% 2.02 20.00% 14.70% 60.00% Indicator of change chart is plotted using the fiscal aligned methodology. View explanation of 'Indicator of Change' methodology here Source: Company Data, Morgan Stanley Research 52 M M Risk Reward – Samsung Electronics (005935.KS) Overweight with narrowed discount to common shares PRICE TARGET W77,000 Our preferred share price target is based on our price target of W90,000 for the common shares and our assumption that the preferred shares' discount to the common shares is 14% (average in the past two years). We assume an 11.5% cost of equity, based on a beta of 1.0, and our terminal growth rate assumption is 5%. W86,500 Consensus Price Target Distribution W63,000.00 W105,000.00 Source: Thomson Reuters, Morgan Stanley Research MS PT Mean Morgan Stanley Estimates RISK REWARD CHART KRW 100000 75000 50000 25000 W101,000(++3366.1.212%%) ) W74,200 W77,000(++33.7.777%%) ) W52,000(--2299.9.922%%) ) OVERWEIGHT THESIS ▪ Share gains in component businesses (DRAM, NAND, and OLED) continue to drive Samsung's earnings. ▪ Samsung should benefit from memory upcycle, and potential share gains and earnings upside from its foundry business. ▪ The smartphone market is mature and no longer growing, but Samsung maintains 20%+ share with its mass model Galaxy A and Note/S series. Foldable smartphones should enable margin expansion in the IM (mobile) division in 2021. ▪ 12-month forward implied P/B valuation of 1.8x is 29% higher than its mid-cycle valuation of 1.4x, but still far lower than historical peak multiple of 2.4x. Samsung is attractive compared to its peers including Micron, Intel, TSMC. Risk Reward Themes Secular Growth: Self-help: Positive Positive View descriptions of Risk Rewards Themes here 0 AUG '20 FEB '21 Key: Historical Stock Performance Current Stock Price Source: Thomson Reuters, Morgan Stanley Research AUG '21 Price Target AUG '22 BULL CASE W101,000 14% discount to bull case value for common shares High-performance computing drives sustained memory pricing strength and capital returns exceed expectations: We assume continued robust demand growth for memory and OLED with stabilization in the macro environment and failure of China's entry. BASE CASE W77,000 14% discount to base case value for common shares We maintain an OW rating on Samsung with a constructive view on DRAM. Our QUANTECH model tactically favors DRAM players including Samsung and Hynix over NOR, based on optimal back-tested factors' weightings. From a fundamental perspective, we expect improving memory conditions to help drive relative outperformance. We estimate DRAM ASP to stay flat and peak out from 4Q21. BEAR CASE W52,000 14% discount to bear case value for common shares Prolonged consumption slowdown, US FANG correction and Chinese competition disrupts OLED and memory profitability. We assume weaker macro conditions stall nearterm global consumption of IT products. Morgan Stanley Research 53 M M Risk Reward – Samsung Electronics (005935.KS) KEY EARNINGS INPUTS Drivers 2020 Smartphone Shipment Growth (%) (12.1) DRAM Bit Growth (%) 22.7 DRAM ASP Growth (%) (18.1) NAND Bit Growth (%) 26.7 NAND ASP Growth (%) (1.5) 2021e 2022e 2023e 10.2 5.3 2.0 26.2 19.3 20.0 14.3 (8.6) (18.0) 43.4 40.4 35.0 (5.2) (9.9) (12.0) INVESTMENT DRIVERS Supply-demand outlook for memory and OLED Smartphone market share and margins Capital returns of 50% FCF GLOBAL REVENUE EXPOSURE 0-10% India 0-10% Japan 0-10% Latin America 0-10% MEA 0-10% UK 10-20% APAC, ex Japan, China and India Mainland 10-20% Europe ex UK 20-30% Mainland China 20-30% North America Source: Morgan Stanley Research Estimate View explanation of regional hierarchies here MS ALPHA MODELS 1/5 MOST 3 Month Horizon Source: Thomson Reuters, FactSet, Morgan Stanley Research; 1 is the highest favored Quintile and 5 is the least favored Quintile RISKS TO PT/RATING RISKS TO UPSIDE New technology developments, especially in memory and foldable displays Longer-lasting memory upcycle from hyperscale data center growth Further narrowing of discount to common shares RISKS TO DOWNSIDE Product and memory cycle, including Apple and new Chinese smartphone competition Earnings growth concentration in semiconductors Deepening of discount to common shares MS ESTIMATES VS. CONSENSUS FY Dec 2021e 6,636 EPS (W) Note: There are not sufficient brokers supplying consensus data for this metric Mean Morgan Stanley Estimates Source: Thomson Reuters, Morgan Stanley Research OWNERSHIP POSITIONING Inst. Owners, % Active 82.9% Source: Thomson Reuters, Morgan Stanley Research SUSTAINABILITY & ESG +1.0 -1.0 Indicator of Change 0.27 Disclosure Rate 82% 54 M M Risk Reward – Samsung Electronics (005935.KS) SUSTAINABILITY AND ESG INDICATOR OF CHANGE OVER TIME 1.0 -0.07 -0.18 -0.45 SUMMARY All Env Gov In 2019, Samsung Electronics had an ESG indicator of Change of 0.27, which signals that the majority of its ESG data points in our database improved YoY. Key ESG metrics that improved in 2019 were CO2 emissions, total waste, and recycled water. 0.27 -0.33 -1.0 2015 2016 2017 2018 2019 Environment CO2 emissions (Scope 1 and 2) Energy consumption Total waste Waste recycled Water consumption Water recycled Unit kilo tons CO2 equivalent GWh tons Percentage kilo tons Percentage Dec '17 13,585 23,419 1,146,812 95% 120,618 47.00% Dec '18 15,173 26,028 1,210,521 96% 134,230 46.00% Dec '19 13,800 26,899 1,099,197 95% 134,479 51.00% Social Avg training hrs per employee Employee engagement Employee turnover Hours Percentage Percentage 73.50 82% - 62.20 87% - 67.20 89% 19.50% Governance Avg tenure of the board Female board members Females in management Independent board members Years Percentage Percentage Percentage 4.72 - 13.20% 33.33% 1.15 - 14.20% 44.44% 2.02 20.00% 14.70% 60.00% Indicator of change chart is plotted using the fiscal aligned methodology. View explanation of 'Indicator of Change' methodology here Source: Company Data, Morgan Stanley Research Morgan Stanley Research 55 M M Financial Summary: Samsung Electronics Exhibit 74: Samsung's Financial Summary Income Statement FY Ending Dec. 31 (KRW bn) Sales Cost of sales Gross profit 2020 236,807 144,488 92,319 2021E 268,465 149,353 119,112 2022E 285,301 158,253 127,048 2023E 300,468 170,183 130,285 R&D SG&A Operating profit 21,111 35,213 35,994 22,600 42,818 53,694 19,971 51,354 55,723 15,023 57,089 58,173 Non-operating income Financial income Gain on FX transact/translation Other 14,158 1,974 9,270 2,914 12,382 1,330 8,756 2,296 10,814 1,600 6,000 3,214 10,820 1,600 6,000 3,220 Non-operating expenses Financial expense Loss on Fab transact/transl. Other Recurring profit 13,807 583 9,869 3,355 36,345 9,272 386 6,939 1,947 56,803 9,000 1,000 6,000 2,000 57,536 9,000 1,000 6,000 2,000 59,993 IBT Income tax expense Net Income from continuing business 36,345 9,937 26,408 56,803 14,402 42,401 57,536 14,384 43,152 59,993 14,998 44,995 Net income Less Dividend to Pref Shr Holders Net income less Div Pref Shr Holders ModelWare Net Income ModelWare EPS 26,408 2,465 23,943 23,626 3,958 42,401 2,587 39,814 39,618 6,636 43,152 3,104 40,048 39,852 6,676 44,995 3,167 41,828 40,928 6,856 Key Ratios (%) Growth Sales Gross profit Operating profit Net profit 2020 3 11 30 21 2021E 13 29 49 61 Profitability EBITDA margin Gross margin Operating margin Net profit ROE ROIC Payout Ratio 28 32 39 44 15 20 11 16 9 14 8 13 78 51 Stability Total liabilities to equity Net debt to equity Current ratio (x) 37 32 -37 -36 2.6 2.8 Per-Share EPS (KRW) BVPS (KRW) DPS (KRW) 3,958 46,204 2,994 6,636 51,267 3,144 Source: Company data, Morgan Stanley Research estimates 2022E 6 7 4 2 31 45 20 15 13 12 60 29 -37 3.1 6,676 55,204 3,773 2023E 5 3 4 4 30 43 19 15 12 12 59 27 -38 3.3 6,856 58,209 3,848 Balance Sheet FY Ending Dec. 31 (KRW bn) Cash & cash equivalents St. investment assets Accts & notes receivable Inventory Other Total current assets 2020 29,383 92,442 30,965 32,043 13,383 198,216 2021E 36,790 93,366 35,105 33,122 13,383 211,766 2022E 46,872 94,300 37,306 35,096 13,383 226,957 2023E 54,795 95,243 39,289 37,741 13,383 240,452 Investment assets PP&E Intangible assets Others Total assets 21,855 128,953 18,469 10,744 378,236 22,292 145,162 17,016 7,744 403,981 22,738 155,840 15,532 4,744 425,811 23,193 163,273 14,014 1,744 442,676 Accounts & Notes payable Short-term borrowings Curr. Portion of LT debt Other current liabilities Total current liabilities 21,638 16,553 1,133 36,280 75,604 22,367 16,388 1,133 34,438 74,325 23,700 16,224 1,133 32,687 73,744 25,486 16,062 1,133 31,024 73,705 Long-term borrowings Other liabilities Total liabilities Paid in capital Capital surplus Retained earnings Capital (adjusted) Total shareholders' equity Total Liab & Eqty 2,948 23,735 102,288 1,431 22,050 97,807 1,418 20,972 96,135 1,405 19,949 95,059 898 4,404 271,068 (422) 275,948 378,236 898 4,404 293,131 7,742 306,174 403,981 898 4,404 314,928 9,446 329,676 425,811 898 4,404 334,297 8,018 347,617 442,676 Cash Flow Statement FY Ending Dec. 31 (KRW bn) Net income Non cash items - net Depreciation Changes in WC Others Operating Cash Flow 2020 26,408 41,619 27,116 122 (2,862) 65,287 2021E 42,401 39,043 25,791 (6,325) (7,000) 68,119 2022E 43,152 42,317 29,032 (4,586) (7,000) 73,884 2023E 44,995 41,369 28,051 (4,498) (7,000) 74,865 Acquisition of PP&E (CAPEX) Proceeds from sale of PP&E Change in current financial assets Increase in intangible assets Others Investing Cash Flow (40,272) 377 (20,370) (2,680) 9,365 (53,579) (42,000) 0 (1,362) (1,800) (1,686) (46,847) (39,710) 0 (1,380) (1,800) (1,077) (43,967) (35,485) 0 (1,398) (1,800) (1,024) (39,706) Increase(Decrease) in Debt Dividends Sale(Purchase) of Treasury Stock Other Financing Cash Flow 1,341 (9,677) 0 8 (8,328) (1,689) (20,338) 8,296 0 (13,732) (184) (21,355) 1,704 0 (19,835) (182) (25,626) (1,428) 0 (27,236) Net change in cash Other change in cash Beginning cash Ending cash 3,380 (834) 26,780 29,277 7,540 0 29,277 36,790 10,082 0 36,790 46,872 7,923 0 46,872 54,795 56 M M Risk Reward – SK Hynix (000660.KS) More risk than reward entering late cycle PRICE TARGET W80,000 Our price target is derived by residual-income model and assume a cost of equity of 10.5% (4% risk-free rate, 6.5% risk premium, 1.0 beta) and a terminal growth rate of 5%. Implied valuation would be 1.1x P/B, or 8% lower than mid-cycle average of 1.2x. W164,484 Consensus Price Target Distribution W80,000 W200,000.00 Source: Thomson Reuters, Morgan Stanley Research MS PT Mean Morgan Stanley Estimates RISK REWARD CHART AND OPTIONS IMPLIED PROBABILITIES (12M) KRW 200000 150000 W133,000(++1188.2.222%%) ) Prob (>133,000)~15.9% 100000 50000 W112,500 W80,000(--2288.8.899%%) ) Prob (< 80,000) ~20.7% W62,000(--4444.8.899%%) ) Prob (<62,000)~8.6% 0 AUG '20 FEB '21 AUG '21 AUG '22 Key: Historical Stock Performance Current Stock Price Price Target Source: Thomson Reuters, Morgan Stanley Research, Morgan Stanley Institutional Equities Division. The probabilities of our Bull, Base, and Bear case scenarios playing out were estimated with implied volatility data from the options market as of 10 Aug, 2021. All figures are approximate risk-neutral probabilities of the stock reaching beyond the scenario price in either three-months’ or one-years’ time. View explanation of Options Probabilities methodology here UNDERWEIGHT THESIS ▪ While pricing is still moving higher, the rate of change is approaching peak as supply is catching up to demand. Our cycle indicator has shifted out of 'mid-cycle' to 'late-cycle' for the first time since 2019 and this phase-change has historically meant a challenging backdrop for forward returns. ▪ The main investor debate at present is around the durability of cyclical positioning against a backdrop of peaking prices. History says that stocks move ahead of the data – so by the time (peak) certainty exists, stocks already headed toward bottom. ▪ Our PT of W80,000 implies 2021e P/B of 1.1x, 8% lower than mid-cycle average of 1.2x. For 2022 BVPS estimate, implied P/B will be 0.8x. Consensus Rating Distribution MS Rating 95% Overweight 3% Equal-weight 3% Underweight Source: Thomson Reuters, Morgan Stanley Research Risk Reward Themes Pricing Power: Secular Growth: Self-help: Negative Positive Positive View descriptions of Risk Rewards Themes here BULL CASE W133,000 Memory TAM grows through new tech segments The market capitalizes on Hynix's potential earnings growth from new tech (Big Data, IoT, AI, autonomous driving). Increased memory demand from these new tech segments causes a deep industry undersupply, boosting prices for memory products. Higher operating leverage and margin outlook based on Intel NAND acquisition and our pro forma analysis suggest ROE expansion to 20% in 2021 (vs. our base case assumption of 17%). Extended DRAM upcycle into 1H22. BASE CASE W80,000 More risk than reward entering late cycle While pricing is still moving higher, the rate of change is approaching peak as supply is catching up to demand. Our cycle indicator has shifted out of 'mid-cycle' to 'late-cycle' for the first time since 2019 and this phase-change has historically meant a challenging backdrop for forward returns. Separately, our base case scenario does not include Hynix's acquisition of Intel business yet. We turn more cautious on memory players, given the potential DRAM price peak out in 4Q21. BEAR CASE W62,000 Supply growth accelerates, macro deteriorates Supply growth accelerates, global macro conditions deteriorate because of Covid-19, and China enters NAND: Industry supplydemand dynamics worsen for DRAM and NAND, leading to sharp ASP and margin erosion. China enters the NAND market initially with 64L 3D NAND in 1H22 and moves to a higher node in 2022-23. Morgan Stanley Research 57 M M Risk Reward – SK Hynix (000660.KS) KEY EARNINGS INPUTS Drivers 2020 DRAM Bit Growth (%) 22.0 DRAM ASP Growth (%) (11.4) NAND Bit Growth (%) 41.8 NAND ASP Growth (%) 0.4 2021e 23.4 17.4 52.4 (4.2) 2022e 21.3 (9.8) 34.5 (10.2) 2023e 20.0 (18.0) 35.0 (15.0) CATALYST CALENDAR Date 02 Nov 2021 08 Nov 2021 27 Jan 2022 31 Jan 2022 Event Source: Thomson Reuters, Morgan Stanley Q3 2021 SK Hynix Inc Earnings Release Q4 2021 SK Hynix Inc Earnings Release INVESTMENT DRIVERS Supply/demand outlook for DRAM and NAND. Global demand for data center products (server DRAM, enterprise SSD). GLOBAL REVENUE EXPOSURE 0-10% Europe ex UK 20-30% Mainland China 30-40% APAC, ex Japan, China and India Mainland 40-50% North America Source: Morgan Stanley Research Estimate View explanation of regional hierarchies here MS ALPHA MODELS 2/5 MOST 3 Month Horizon Source: Thomson Reuters, FactSet, Morgan Stanley Research; 1 is the highest favored Quintile and 5 is the least favored Quintile RISKS TO PT/RATING RISKS TO UPSIDE DRAM pricing reverses later than expectations and the uptrend continues into 1H22. Hyperscalers continue to buy DRAM and NAND after quarters of inventory adjustment period. RISKS TO DOWNSIDE 3D NAND mass production is delayed due to technological hurdles. Earlier-than-expected DRAM price downturns on weak end-demand and YoY increase in production bit shipment. OWNERSHIP POSITIONING Inst. Owners, % Active 82.9% Source: Thomson Reuters, Morgan Stanley Research MS ESTIMATES VS. CONSENSUS FY Dec 2021e Sales / Revenue (W, bn) 39,759 43,805 44,703 42,930 EPS (W) EBIT (W, bn) ROE (%) 11,310 12,842 13,577 11,311 12,490 12,724 18.0 12.1 17.0 16,815 14,526 21.6 Mean Morgan Stanley Estimates Source: Thomson Reuters, Morgan Stanley Research SUSTAINABILITY & ESG +1.0 -1.0 Indicator of Change -0.44 Disclosure Rate 71% 58 M M Risk Reward – SK Hynix (000660.KS) SUSTAINABILITY AND ESG INDICATOR OF CHANGE OVER TIME 1.0 -0.13 -0.19 -0.71 -1.0 2015 Environment CO2 emissions (Scope 1 and 2) Energy consumption Total waste Waste recycled Water consumption Water recycled 2016 2017 Unit tons CO2 equivalent Gigajoules tons Percentage Thousand cubic metre Percentage All Env Soc Gov -0.24 -0.44 2018 2019 Dec '17 5,087,704 70,424,815 296,603 85.00% 35,064 61.94% Dec '18 5,914,390 83,978,734 444,433 90.00% 42,358 60.65% Dec '19 6,839,470 85,270,649 511,216 93.00% 84,638 36.66% SUMMARY In 2019, SK Hynix had an ESG indicator of Change of -0.44, which signals that the majority of its ESG data points in our database decreased YoY. It is in the 4th quartile vs. sector peers. Its ESG disclosure rate was 71% in 2019. Key ESG metrics that decreased in 2019 were CO2 emissions, total waste, and recycled water. 2020 data is not available yet, but we believe the company's ESG rate will improve with its efforts to enhance green activities. SK Hynix, in January, issued a green bond worth US$1bn to expand its eco-friendly business such as for water quality management, enhanced energy efficiency, contamination prevention and restoring of environment. It is also working on building of new high-tech wastewater treatment plant and water reuse system. Social Avg training hrs per employee Employee engagement Employee turnover Hours Percentage Percentage 7.80 85.60% 2.20% 9.50 86.10% 2.00% 9.40 - 2.30% Governance Avg tenure of the board Females in management Independent board members Years Percentage Percentage 2.81 25.20% 60.00% 2.56 21.60% 66.67% 2.89 20.30% 66.67% Indicator of change chart is plotted using the fiscal aligned methodology. View explanation of 'Indicator of Change' methodology here Source: Company Data, Morgan Stanley Research Morgan Stanley Research 59 M M Financial Summary: SK Hynix Exhibit 75: SK Hynix's Financial Summary Income Statement (W bn) Sales Gross Profit Operating Expense EBITDA Operating Income Net Non-operating Income - Net Interest Income Income Before Tax Net Income 2020 31,901 10,811 5,797 14,785 5,013 1,224 -226 6,238 4,755 2021E 43,805 20,537 8,046 22,522 12,490 152 -27 12,643 9,349 2022E 53,950 25,768 10,790 25,880 14,977 60 60 15,037 11,278 2023E 55,861 22,804 11,172 24,365 11,631 60 60 11,691 8,769 Margins Gross Margin EBITDA Margin Operating Margin Net Income 33.9% 46.3% 15.7% 14.9% 46.9% 51.4% 28.5% 21.3% 47.8% 48.0% 27.8% 20.9% 40.8% 43.6% 20.8% 15.7% Growth Rates Sales EBITDA Operating Income Net Income 18.2% 30.5% 84.8% 137.0% 37.3% 52.3% 149.2% 96.6% 23.2% 14.9% 19.9% 20.6% 3.5% -5.9% -22.3% -22.3% Cash Flow Statement (W bn) Net Income Depreciation & Amortization Other Non-Cash Items Change in Net Working Capital Operating Cash flow Capital Expenditure Proceeds from sale of PP&E Change in Other Assets Net Cash Used in Invest Activities Net Change in Debt Net Change in Equity Dividends Others Net Cash Flow From Financing Other Cash Flow Net change in cash Beginning cash Ending cash 2020 4,755 9,772 -563 -1,650 12,315 2021E 9,349 10,031 1,858 -2,318 18,920 2022E 11,278 10,902 1,858 -2,154 21,885 2023E 8,769 12,734 1,858 -544 22,816 -10,069 59 -1,831 -11,840 -13,089 59 -1,000 -14,030 -18,325 59 -1,000 -19,266 -21,990 59 -1,000 -22,931 1,252 0 -684 -316 252 -122 0 -800 0 -922 -118 0 -2,635 0 -2,753 -114 0 -3,179 0 -3,293 -56 670 2,306 2,976 0 3,968 2,976 6,944 0 -135 6,944 6,809 0 -3,408 6,809 3,401 Balance Sheet (W bn) Cash & Cash Equivalents Receivables Inventory Other current Assets Current Assets 2020 3,413 4,995 6,136 2,026 16,571 2021E 7,380 6,698 8,426 736 23,240 2022E 7,246 8,249 10,378 -715 25,158 2023E 3,837 8,541 10,745 -2,106 21,018 Investment Assets Property and Equipment Fixed Assets Total Assets 7,344 41,231 54,603 71,174 7,351 45,190 58,360 81,600 7,359 53,514 66,692 91,849 7,366 63,672 76,857 97,875 Trade Payables Short Term Debts Current Portion of LT Debts Total Current Liabilities Long Term Debts Bonds Total Long Term Liabilities Total Liabilities 4,843 180 2,935 9,072 4,527 3,610 10,192 19,265 6,518 198 2,876 10,929 4,482 3,574 10,214 21,143 7,867 217 2,818 12,508 4,437 3,539 10,241 22,749 7,982 239 2,762 12,910 4,393 3,503 10,275 23,184 Paid-in Capital Retained Earnings Total Stockholders' Equities Total Liabilities & SE 3,658 46,996 51,909 71,174 3,658 55,544 60,458 81,600 3,658 64,187 69,101 91,849 3,658 69,777 74,690 97,875 Net Debt 7,839 3,749 3,766 7,059 Source: Company data, Morgan Stanley Research estimates Ratio Analysis (W bn) ROE (AOY) ROE (EOY) ROA ROIC (AOY) Asset Turnover Day's Receivables Day's Inventory Day's Payables Current Ratio Debt to Equity (%) Net Debt to Equity (%) EPS (MW) BVPS (MW) DPS P/E P/BV EV/EBITDA Dividend Yield (%) FCF Yield Share Price (KRW) Shares Outstanding (000 AOY) 2020 9.5% 9.9% 7.0% 5.6% 2021E 16.6% 18.0% 12.2% 12.2% 2022E 17.4% 18.7% 13.0% 13.1% 2023E 12.2% 12.7% 9.2% 9.0% 0.4 57.2 106.2 83.8 0.5 55.8 132.2 102.2 0.6 55.8 134.4 101.9 0.6 55.8 118.6 88.1 1.8 2.1 2.0 1.6 37.1 35.0 32.9 31.0 15.1 6.2 5.4 9.5 6,532 71,275 1,170 12,842 83,018 3,853 15,492 12,045 94,890 102,568 4,648 3,613 11.7 9.0 7.5 9.6 1.1 1.4 1.2 1.1 4.2 3.8 3.3 3.7 1.5 3.3 4.0 3.1 4.4 7.0 4.3 1.1 76,500 116,000 116,000 116,000 728,002 728,002 728,002 728,002 60 M M Disclosure Section The information and opinions in Morgan Stanley Research were prepared or are disseminated by Morgan Stanley & Co. LLC and/or Morgan Stanley C.T.V.M. S.A. and/or Morgan Stanley México, Casa de Bolsa, S.A. de C.V. and/or Morgan Stanley Canada Limited and/or Morgan Stanley & Co. International plc and/or Morgan Stanley Europe S.E. and/or RMB Morgan Stanley Proprietary Limited and/or Morgan Stanley MUFG Securities Co., Ltd. and/or Morgan Stanley Capital Group Japan Co., Ltd. and/or Morgan Stanley Asia Limited and/or Morgan Stanley Asia (Singapore) Pte. (Registration number 199206298Z) and/or Morgan Stanley Asia (Singapore) Securities Pte Ltd (Registration number 200008434H), regulated by the Monetary Authority of Singapore (which accepts legal responsibility for its contents and should be contacted with respect to any matters arising from, or in connection with, Morgan Stanley Research) and/or Morgan Stanley Taiwan Limited and/or Morgan Stanley & Co International plc, Seoul Branch, and/or Morgan Stanley Australia Limited (A.B.N. 67 003 734 576, holder of Australian financial services license No. 233742, which accepts responsibility for its contents), and/or Morgan Stanley Wealth Management Australia Pty Ltd (A.B.N. 19 009 145 555, holder of Australian financial services license No. 240813, which accepts responsibility for its contents), and/or Morgan Stanley India Company Private Limited, regulated by the Securities and Exchange Board of India (“SEBI”) and holder of licenses as a Research Analyst (SEBI Registration No. INH000001105), Stock Broker (BSE Registration No. INB011054237 and NSE Registration No. INB/INF231054231), Merchant Banker (SEBI Registration No. INM000011203), and depository participant with National Securities Depository Limited (SEBI Registration No. IN-DP-NSDL-372-2014) which accepts the responsibility for its contents and should be contacted with respect to any matters arising from, or in connection with, Morgan Stanley Research, and their affiliates (collectively, "Morgan Stanley"). For important disclosures, stock price charts and equity rating histories regarding companies that are the subject of this report, please see the Morgan Stanley Research Disclosure Website at www.morganstanley.com/researchdisclosures, or contact your investment representative or Morgan Stanley Research at 1585 Broadway, (Attention: Research Management), New York, NY, 10036 USA. For valuation methodology and risks associated with any recommendation, rating or price target referenced in this research report, please contact the Client Support Team as follows: US/Canada +1 800 303-2495; Hong Kong +852 2848-5999; Latin America +1 718 754-5444 (U.S.); London +44 (0)20-7425-8169; Singapore +65 6834-6860; Sydney +61 (0)2-9770-1505; Tokyo +81 (0)3-6836-9000. Alternatively you may contact your investment representative or Morgan Stanley Research at 1585 Broadway, (Attention: Research Management), New York, NY 10036 USA. Analyst Certification The following analysts hereby certify that their views about the companies and their securities discussed in this report are accurately expressed and that they have not received and will not receive direct or indirect compensation in exchange for expressing specific recommendations or views in this report: Charlie Chan; Ryan Kim; Shawn Kim; Joseph Moore; Dominik Olszewski, CFA; Gilbert Wong, CFA; Ray Wu, CFA; Daniel Yen, CFA; Kazuo Yoshikawa, CFA. . Global Research Conflict Management Policy Morgan Stanley Research has been published in accordance with our conflict management policy, which is available at www.morganstanley.com/institutional/research/conflictpolicies. A Portuguese version of the policy can be found at www.morganstanley.com.br Important Regulatory Disclosures on Subject Companies As of July 30, 2021, Morgan Stanley beneficially owned 1% or more of a class of common equity securities of the following companies covered in Morgan Stanley Research: Accton Technology Corporation, Acer Inc., ACM Research Inc, Advanced Process Systems Corp, AirTAC International, ams AG, ASML Holding NV, AU Optronics, Catcher Technology, Delta Electronics Inc., Duk San Neolux Co Ltd, Ecopro BM, Ecopro Co Ltd, Flexium, GigaDevice Semiconductor Beijing Inc, GIS Holding Limited, Infineon Technologies AG, Innolux, Kinsus Interconnect Tech., L&F Co Ltd, LandMark Optoelectronics Corporation, LG Display, LG Innotek, Macronix International Co Ltd, Micron Technology Inc., Radiant Opto-Electronics Corporation, Shengyi Technology Co Ltd., Silicon Motion, Solus Advanced Materials Co Ltd, Tong Hsing, Visual Photonics Epitaxy Co Ltd, Western Digital, Xiaomi Corp, Zhen Ding, Zhongji Innolight Co Ltd, ZTE Corporation. Within the last 12 months, Morgan Stanley managed or co-managed a public offering (or 144A offering) of securities of ams AG, Lenovo, Xiaomi Corp. Within the last 12 months, Morgan Stanley has received compensation for investment banking services from AAC Technologies Holdings, ams AG, Lenovo, Micron Technology Inc., Samsung Electronics, Xiaomi Corp. In the next 3 months, Morgan Stanley expects to receive or intends to seek compensation for investment banking services from AAC Technologies Holdings, Advantech, ams AG, ASM International NV, ASML Holding NV, AU Optronics, Catcher Technology, Delta Electronics Inc., Ennoconn Corporation, Foxconn Industrial Internet Co. Ltd., GoerTek Inc, HIKVision Digital Technology, Hon Hai Precision, Infineon Technologies AG, Legend Holdings Corp, Lenovo, LG Electronics, Lite-On Technology, Micron Technology Inc., Nanya Technology Corp., Samsung Electro-Mechanics, Samsung Electronics, SK Hynix, Soitec SA, STMicroelectronics NV, Tokyo Electron, Wistron Corporation, Xiaomi Corp, Yageo Corp.. Within the last 12 months, Morgan Stanley has received compensation for products and services other than investment banking services from AAC Technologies Holdings, Accton Technology Corporation, Acer Inc., ams AG, ASM International NV, Asustek Computer Inc., AU Optronics, BYD Electronics, Compal Electronics, Foxconn Technology, GoerTek Inc, Hon Hai Precision, Infineon Technologies AG, Innolux, Lenovo, Micron Technology Inc., Nanya Technology Corp., Quanta Computer Inc., Silicon Motion, Tokyo Electron, Tripod Technology, Winbond Electronics Corp, Yageo Corp.. Within the last 12 months, Morgan Stanley has provided or is providing investment banking services to, or has an investment banking client relationship with, the following company: AAC Technologies Holdings, Advantech, ams AG, ASM International NV, ASML Holding NV, AU Optronics, Catcher Technology, Delta Electronics Inc., Ennoconn Corporation, Foxconn Industrial Internet Co. Ltd., GoerTek Inc, HIKVision Digital Technology, Hon Hai Precision, Infineon Technologies AG, Legend Holdings Corp, Lenovo, LG Electronics, Lite-On Technology, Micron Technology Inc., Nanya Technology Corp., Samsung Electro-Mechanics, Samsung Electronics, SK Hynix, Soitec SA, STMicroelectronics NV, Tokyo Electron, Wistron Corporation, Xiaomi Corp, Yageo Corp.. Within the last 12 months, Morgan Stanley has either provided or is providing non-investment banking, securities-related services to and/or in the past has entered into an agreement to provide services or has a client relationship with the following company: AAC Technologies Holdings, Accton Technology Corporation, Acer Inc., ams AG, ASM International NV, Asustek Computer Inc., AU Optronics, BYD Electronics, Compal Electronics, Foxconn Technology, GoerTek Inc, Hon Hai Precision, Infineon Technologies AG, Innolux, Lenovo, LG Display, LG Electronics, Macronix Morgan Stanley Research 61 M M International Co Ltd, Micron Technology Inc., Nanya Technology Corp., Quanta Computer Inc., Silicon Motion, Tokyo Electron, Tripod Technology, Winbond Electronics Corp, Yageo Corp.. Morgan Stanley & Co. LLC makes a market in the securities of ASML Holding NV, LG Display, Micron Technology Inc., Silicon Motion, Western Digital. The equity research analysts or strategists principally responsible for the preparation of Morgan Stanley Research have received compensation based upon various factors, including quality of research, investor client feedback, stock picking, competitive factors, firm revenues and overall investment banking revenues. Equity Research analysts' or strategists' compensation is not linked to investment banking or capital markets transactions performed by Morgan Stanley or the profitability or revenues of particular trading desks. Morgan Stanley and its affiliates do business that relates to companies/instruments covered in Morgan Stanley Research, including market making, providing liquidity, fund management, commercial banking, extension of credit, investment services and investment banking. Morgan Stanley sells to and buys from customers the securities/instruments of companies covered in Morgan Stanley Research on a principal basis. Morgan Stanley may have a position in the debt of the Company or instruments discussed in this report. Morgan Stanley trades or may trade as principal in the debt securities (or in related derivatives) that are the subject of the debt research report. Certain disclosures listed above are also for compliance with applicable regulations in non-US jurisdictions. STOCK RATINGS Morgan Stanley uses a relative rating system using terms such as Overweight, Equal-weight, Not-Rated or Underweight (see definitions below). Morgan Stanley does not assign ratings of Buy, Hold or Sell to the stocks we cover. Overweight, Equal-weight, Not-Rated and Underweight are not the equivalent of buy, hold and sell. Investors should carefully read the definitions of all ratings used in Morgan Stanley Research. In addition, since Morgan Stanley Research contains more complete information concerning the analyst's views, investors should carefully read Morgan Stanley Research, in its entirety, and not infer the contents from the rating alone. In any case, ratings (or research) should not be used or relied upon as investment advice. An investor's decision to buy or sell a stock should depend on individual circumstances (such as the investor's existing holdings) and other considerations. Global Stock Ratings Distribution (as of July 31, 2021) The Stock Ratings described below apply to Morgan Stanley's Fundamental Equity Research and do not apply to Debt Research produced by the Firm. For disclosure purposes only (in accordance with FINRA requirements), we include the category headings of Buy, Hold, and Sell alongside our ratings of Overweight, Equal-weight, Not-Rated and Underweight. Morgan Stanley does not assign ratings of Buy, Hold or Sell to the stocks we cover. Overweight, Equal-weight, Not-Rated and Underweight are not the equivalent of buy, hold, and sell but represent recommended relative weightings (see definitions below). To satisfy regulatory requirements, we correspond Overweight, our most positive stock rating, with a buy recommendation; we correspond Equal-weight and Not-Rated to hold and Underweight to sell recommendations, respectively. Stock Rating Category Overweight/Buy Equal-weight/Hold Not-Rated/Hold Underweight/Sell Total Coverage Universe Count 1511 1466 1 518 3,496 % of Total 43% 42% 0% 15% Investment Banking Clients (IBC) Other Material Investment Services Clients (MISC) Count % of Total IBC % of Rating Category Count % of Total Other MISC 421 48% 28% 676 44% 372 42% 25% 654 43% 0 0% 0% 0 0% 84 10% 16% 199 13% 877 1529 Data include common stock and ADRs currently assigned ratings. Investment Banking Clients are companies from whom Morgan Stanley received investment banking compensation in the last 12 months. Due to rounding off of decimals, the percentages provided in the "% of total" column may not add up to exactly 100 percent. Analyst Stock Ratings Overweight (O or Over) - The stock's total return is expected to exceed the total return of the relevant country MSCI Index or the average total return of the analyst's industry (or industry team's) coverage universe, on a risk-adjusted basis over the next 12-18 months. Equal-weight (E or Equal) - The stock's total return is expected to be in line with the total return of the relevant country MSCI Index or the average total return of the analyst's industry (or industry team's) coverage universe, on a risk-adjusted basis over the next 12-18 months. Not-Rated (NR) - Currently the analyst does not have adequate conviction about the stock's total return relative to the relevant country MSCI Index or the average total return of the analyst's industry (or industry team's) coverage universe, on a risk-adjusted basis, over the next 12-18 months. Underweight (U or Under) - The stock's total return is expected to be below the total return of the relevant country MSCI Index or the average total return of the analyst's industry (or industry team's) coverage universe, on a risk-adjusted basis, over the next 12-18 months. Unless otherwise specified, the time frame for price targets included in Morgan Stanley Research is 12 to 18 months. Analyst Industry Views Attractive (A): The analyst expects the performance of his or her industry coverage universe over the next 12-18 months to be attractive vs. the relevant broad market benchmark, as indicated below. 62 M M In-Line (I): The analyst expects the performance of his or her industry coverage universe over the next 12-18 months to be in line with the relevant broad market benchmark, as indicated below. Cautious (C): The analyst views the performance of his or her industry coverage universe over the next 12-18 months with caution vs. the relevant broad market benchmark, as indicated below. Benchmarks for each region are as follows: North America - S&P 500; Latin America - relevant MSCI country index or MSCI Latin America Index; Europe - MSCI Europe; Japan - TOPIX; Asia relevant MSCI country index or MSCI sub-regional index or MSCI AC Asia Pacific ex Japan Index. Stock Price, Price Target and Rating History (See Rating Definitions) Morgan Stanley Research 63 M M 64 M M Important Disclosures for Morgan Stanley Smith Barney LLC & E*TRADE Securities LLC Customers Important disclosures regarding the relationship between the companies that are the subject of Morgan Stanley Research and Morgan Stanley Smith Barney LLC or Morgan Stanley or any of their affiliates, are available on the Morgan Stanley Wealth Management disclosure website at www.morganstanley.com/online/researchdisclosures. For Morgan Stanley specific disclosures, you may refer to www.morganstanley.com/researchdisclosures. Each Morgan Stanley research report is reviewed and approved on behalf of Morgan Stanley Smith Barney LLC and E*TRADE Securities LLC. This review and approval is conducted by the same person who reviews the research report on behalf of Morgan Stanley. This could create a conflict of interest. Other Important Disclosures Morgan Stanley & Co. International PLC and its affiliates have a significant financial interest in the debt securities of ams AG, ASM International NV, ASML Holding NV, Infineon Technologies AG, Legend Holdings Corp, Lenovo, LG Display, Micron Technology Inc., Samsung Electronics, SK Hynix, STMicroelectronics NV, Western Digital, Yageo Corp., Zhen Ding. Morgan Stanley Research policy is to update research reports as and when the Research Analyst and Research Management deem appropriate, based on developments with the issuer, the sector, or the market that may have a material impact on the research views or opinions stated therein. In addition, certain Research publications are intended to be updated on a regular periodic basis (weekly/monthly/quarterly/annual) and will ordinarily be updated with that frequency, unless the Research Analyst and Research Management determine that a different publication schedule is appropriate based on current conditions. Morgan Stanley is not acting as a municipal advisor and the opinions or views contained herein are not intended to be, and do not constitute, advice within the meaning of Section 975 of the Dodd-Frank Wall Street Reform and Consumer Protection Act. Morgan Stanley produces an equity research product called a "Tactical Idea." Views contained in a "Tactical Idea" on a particular stock may be contrary to the recommendations or views expressed in research on the same stock. This may be the result of differing time horizons, methodologies, market events, or other factors. For all research available on a particular stock, please contact your sales representative or go to Matrix at http://www.morganstanley.com/matrix. Morgan Stanley Research is provided to our clients through our proprietary research portal on Matrix and also distributed electronically by Morgan Stanley to clients. Certain, but not all, Morgan Stanley Research products are also made available to clients through third-party vendors or redistributed to clients through alternate electronic means as a convenience. For access to all available Morgan Stanley Research, please contact your sales representative or go to Matrix at http://www.morganstanley.com/matrix. Any access and/or use of Morgan Stanley Research is subject to Morgan Stanley's Terms of Use (http://www.morganstanley.com/terms.html). By accessing and/or using Morgan Stanley Research, you are indicating that you have read and agree to be bound by our Terms of Use (http://www.morganstanley.com/terms.html). In addition you consent to Morgan Stanley processing your personal data and using cookies in accordance with our Privacy Policy and our Global Cookies Policy (http://www.morganstanley.com/privacy_pledge.html), including for the purposes of setting your preferences and to collect readership data so that we can deliver better and more personalized service and products to you. To find out more information about how Morgan Stanley processes personal data, how we use cookies and how to reject cookies see our Privacy Policy and our Global Cookies Policy (http://www.morganstanley.com/privacy_pledge.html). If you do not agree to our Terms of Use and/or if you do not wish to provide your consent to Morgan Stanley processing your personal data or using cookies please do not access our research. Morgan Stanley Research does not provide individually tailored investment advice. Morgan Stanley Research has been prepared without regard to the circumstances and objectives of those who receive it. Morgan Stanley recommends that investors independently evaluate particular investments and strategies, and encourages investors to seek the advice of a financial adviser. The appropriateness of an investment or strategy will depend on an investor's circumstances and objectives. The securities, instruments, or strategies discussed in Morgan Stanley Research may not be suitable for all investors, and certain investors may not be eligible to purchase or participate in some or all of them. Morgan Stanley Research is not an offer to buy or sell or the solicitation of an offer to buy or sell any security/instrument or to participate in any particular trading strategy. The value of and income from your investments may vary because of changes in interest rates, foreign exchange rates, default rates, prepayment rates, securities/instruments prices, market indexes, operational or financial conditions of companies or other factors. There may be time limitations on the exercise of options or other rights in securities/instruments transactions. Past performance is not necessarily a guide to future performance. Estimates of future performance are based on assumptions that may not be realized. If provided, and unless otherwise stated, the closing price on the cover page is that of the primary exchange for the subject company's securities/instruments. The fixed income research analysts, strategists or economists principally responsible for the preparation of Morgan Stanley Research have received compensation based upon various factors, including quality, accuracy and value of research, firm profitability or revenues (which include fixed income trading and capital markets profitability or revenues), client feedback and competitive factors. Fixed Income Research analysts', strategists' or economists' compensation is not linked to investment banking or capital markets transactions performed by Morgan Stanley or the profitability or revenues of particular trading desks. The "Important Regulatory Disclosures on Subject Companies" section in Morgan Stanley Research lists all companies mentioned where Morgan Stanley owns 1% or more of a class of common equity securities of the companies. For all other companies mentioned in Morgan Stanley Research, Morgan Stanley may have an investment of less than 1% in securities/instruments or derivatives of securities/instruments of companies and may trade them in ways different from those discussed in Morgan Stanley Research. Employees of Morgan Stanley not involved in the preparation of Morgan Stanley Research may have investments in securities/instruments or derivatives of securities/instruments of companies mentioned and may trade them in ways different from those discussed in Morgan Stanley Research. Derivatives may be issued by Morgan Stanley or associated persons. With the exception of information regarding Morgan Stanley, Morgan Stanley Research is based on public information. Morgan Stanley makes every effort to use reliable, comprehensive information, but we make no representation that it is accurate or complete. We have no obligation to tell you when opinions or information in Morgan Stanley Research change apart from when we intend to discontinue equity research coverage of a subject company. Facts and views presented in Morgan Stanley Research have not been reviewed by, and may not reflect information known to, professionals in other Morgan Stanley business areas, including investment banking personnel. Morgan Stanley Research personnel may participate in company events such as site visits and are generally prohibited from accepting payment by the company of associated expenses unless pre-approved by authorized members of Research management. Morgan Stanley Research 65 M M Morgan Stanley may make investment decisions that are inconsistent with the recommendations or views in this report. To our readers based in Taiwan or trading in Taiwan securities/instruments: Information on securities/instruments that trade in Taiwan is distributed by Morgan Stanley Taiwan Limited ("MSTL"). Such information is for your reference only. The reader should independently evaluate the investment risks and is solely responsible for their investment decisions. Morgan Stanley Research may not be distributed to the public media or quoted or used by the public media without the express written consent of Morgan Stanley. Any non-customer reader within the scope of Article 7-1 of the Taiwan Stock Exchange Recommendation Regulations accessing and/or receiving Morgan Stanley Research is not permitted to provide Morgan Stanley Research to any third party (including but not limited to related parties, affiliated companies and any other third parties) or engage in any activities regarding Morgan Stanley Research which may create or give the appearance of creating a conflict of interest. Information on securities/instruments that do not trade in Taiwan is for informational purposes only and is not to be construed as a recommendation or a solicitation to trade in such securities/instruments. MSTL may not execute transactions for clients in these securities/instruments. Certain information in Morgan Stanley Research was sourced by employees of the Shanghai Representative Office of Morgan Stanley Asia Limited for the use of Morgan Stanley Asia Limited. Morgan Stanley is not incorporated under PRC law and the research in relation to this report is conducted outside the PRC. Morgan Stanley Research does not constitute an offer to sell or the solicitation of an offer to buy any securities in the PRC. PRC investors shall have the relevant qualifications to invest in such securities and shall be responsible for obtaining all relevant approvals, licenses, verifications and/or registrations from the relevant governmental authorities themselves. Neither this report nor any part of it is intended as, or shall constitute, provision of any consultancy or advisory service of securities investment as defined under PRC law. Such information is provided for your reference only. Morgan Stanley Research is disseminated in Brazil by Morgan Stanley C.T.V.M. S.A. located at Av. Brigadeiro Faria Lima, 3600, 6th floor, São Paulo - SP, Brazil; and is regulated by the Comissão de Valores Mobiliários; in Mexico by Morgan Stanley México, Casa de Bolsa, S.A. de C.V which is regulated by Comision Nacional Bancaria y de Valores. Paseo de los Tamarindos 90, Torre 1, Col. Bosques de las Lomas Floor 29, 05120 Mexico City; in Japan by Morgan Stanley MUFG Securities Co., Ltd. and, for Commodities related research reports only, Morgan Stanley Capital Group Japan Co., Ltd; in Hong Kong by Morgan Stanley Asia Limited (which accepts responsibility for its contents) and by Morgan Stanley Asia International Limited, Hong Kong Branch; in Singapore by Morgan Stanley Asia (Singapore) Pte. (Registration number 199206298Z) and/or Morgan Stanley Asia (Singapore) Securities Pte Ltd (Registration number 200008434H), regulated by the Monetary Authority of Singapore (which accepts legal responsibility for its contents and should be contacted with respect to any matters arising from, or in connection with, Morgan Stanley Research) and by Morgan Stanley Asia International Limited, Singapore Branch (Registration number T11FC0207F); in Australia to "wholesale clients" within the meaning of the Australian Corporations Act by Morgan Stanley Australia Limited A.B.N. 67 003 734 576, holder of Australian financial services license No. 233742, which accepts responsibility for its contents; in Australia to "wholesale clients" and "retail clients" within the meaning of the Australian Corporations Act by Morgan Stanley Wealth Management Australia Pty Ltd (A.B.N. 19 009 145 555, holder of Australian financial services license No. 240813, which accepts responsibility for its contents; in Korea by Morgan Stanley & Co International plc, Seoul Branch; in India by Morgan Stanley India Company Private Limited; in Canada by Morgan Stanley Canada Limited, which has approved of and takes responsibility for its contents in Canada; in Germany and the European Economic Area where required by Morgan Stanley Europe S.E., authorised and regulated by Bundesanstalt fuer Finanzdienstleistungsaufsicht (BaFin) under the reference number 149169; in the US by Morgan Stanley & Co. LLC, which accepts responsibility for its contents. Morgan Stanley & Co. International plc, authorized by the Prudential Regulatory Authority and regulated by the Financial Conduct Authority and the Prudential Regulatory Authority, disseminates in the UK research that it has prepared, and approves solely for the purposes of section 21 of the Financial Services and Markets Act 2000, research which has been prepared by any of its affiliates. RMB Morgan Stanley Proprietary Limited is a member of the JSE Limited and A2X (Pty) Ltd. RMB Morgan Stanley Proprietary Limited is a joint venture owned equally by Morgan Stanley International Holdings Inc. and RMB Investment Advisory (Proprietary) Limited, which is wholly owned by FirstRand Limited. The information in Morgan Stanley Research is being disseminated by Morgan Stanley Saudi Arabia, regulated by the Capital Market Authority in the Kingdom of Saudi Arabia , and is directed at Sophisticated investors only. Morgan Stanley Hong Kong Securities Limited is the liquidity provider/market maker for securities of AAC Technologies Holdings, Lenovo, Sunny Optical, Xiaomi Corp listed on the Stock Exchange of Hong Kong Limited. An updated list can be found on HKEx website: http://www.hkex.com.hk. The information in Morgan Stanley Research is being communicated by Morgan Stanley & Co. International plc (DIFC Branch), regulated by the Dubai Financial Services Authority (the DFSA), and is directed at Professional Clients only, as defined by the DFSA. The financial products or financial services to which this research relates will only be made available to a customer who we are satisfied meets the regulatory criteria to be a Professional Client. The information in Morgan Stanley Research is being communicated by Morgan Stanley & Co. International plc (QFC Branch), regulated by the Qatar Financial Centre Regulatory Authority (the QFCRA), and is directed at business customers and market counterparties only and is not intended for Retail Customers as defined by the QFCRA. As required by the Capital Markets Board of Turkey, investment information, comments and recommendations stated here, are not within the scope of investment advisory activity. Investment advisory service is provided exclusively to persons based on their risk and income preferences by the authorized firms. Comments and recommendations stated here are general in nature. These opinions may not fit to your financial status, risk and return preferences. For this reason, to make an investment decision by relying solely to this information stated here may not bring about outcomes that fit your expectations. The following companies do business in countries which are generally subject to comprehensive sanctions programs administered or enforced by the U.S. Department of the Treasury's Office of Foreign Assets Control ("OFAC") and by other countries and multi-national bodies: Samsung Electronics. The trademarks and service marks contained in Morgan Stanley Research are the property of their respective owners. Third-party data providers make no warranties or representations relating to the accuracy, completeness, or timeliness of the data they provide and shall not have liability for any damages relating to such data. The Global Industry Classification Standard (GICS) was developed by and is the exclusive property of MSCI and S&P. Morgan Stanley Research, or any portion thereof may not be reprinted, sold or redistributed without the written consent of Morgan Stanley. Indicators and trackers referenced in Morgan Stanley Research may not be used as, or treated as, a benchmark under Regulation EU 2016/1011, or any other similar framework. 66 M M INDUSTRY COVERAGE: S. Korea Technology Company (Ticker) Rating (As Of) Price* (08/11/2021) Ryan Kim Advanced Process Systems Corp (265520.KQ) Duk San Neolux Co Ltd (213420.KQ) Ecopro BM (247540.KQ) Ecopro Co Ltd (086520.KQ) Iljin Materials (020150.KS) L&F Co Ltd (066970.KQ) Posco Chemical Co Ltd. (003670.KS) Solus Advanced Materials Co Ltd (336370.KS) Wonik IPS Co Ltd (240810.KQ) Shawn Kim LG Display (034220.KS) LG Electronics (066570.KS) LG Innotek (011070.KS) Samsung Electro-Mechanics (009150.KS) Samsung Electronics (005935.KS) Samsung Electronics (005930.KS) Samsung SDI (006400.KS) Samsung SDS (018260.KS) Seoul Semiconductor (046890.KQ) SK Hynix (000660.KS) Stock Ratings are subject to change. Please see latest research for each company. * Historical prices are not split adjusted. O (04/09/2020) O (04/09/2020) O (08/03/2021) E (02/21/2019) O (02/21/2019) O (12/17/2020) U (04/27/2021) O (05/11/2021) O (09/07/2020) E (03/19/2020) E (11/04/2020) E (07/19/2021) E (05/12/2021) O (11/18/2019) O (11/18/2019) U (05/30/2021) E (06/23/2017) U (04/04/2018) U (08/11/2021) W30,000 W70,200 W311,200 W101,000 W79,700 W123,100 W152,000 W66,900 W47,400 W22,100 W156,000 W213,000 W183,000 W73,100 W78,500 W802,000 W180,000 W17,500 W105,500 INDUSTRY COVERAGE: Greater China Technology Hardware Company (Ticker) Rating (As Of) Price* (08/11/2021) Andy Meng, CFA BYD Electronics (0285.HK) China TransInfo Technology Co Ltd (002373.SZ) Dahua Technology Co. Ltd. (002236.SZ) HIKVision Digital Technology (002415.SZ) Largan Precision (3008.TW) OFILM Group Co Ltd (002456.SZ) Sunny Optical (2382.HK) Wingtech Technology Co Ltd (600745.SS) Xiaomi Corp (1810.HK) Zhongji Innolight Co Ltd (300308.SZ) ZTE Corporation (0763.HK) ZTE Corporation (000063.SZ) Derrick Yang Advantech (2395.TW) AirTAC International (1590.TW) AU Optronics (2409.TW) BOE Technology (000725.SZ) Ennoconn Corporation (6414.TW) GIS Holding Limited (6456.TW) Hiwin Technologies Corp. (2049.TW) Innolux (3481.TW) Lens Technology (300433.SZ) Leyard Optoelectronic Co Ltd (300296.SZ) Radiant Opto-Electronics Corporation (6176.TW) O (08/11/2021) U (05/24/2021) O (05/24/2021) O (11/02/2015) E (08/28/2020) U (04/14/2021) O (01/12/2021) O (07/02/2021) O (04/14/2021) O (09/10/2020) O (06/20/2019) U (07/02/2021) O (01/20/2021) O (01/20/2021) O (12/09/2019) O (09/06/2019) O (01/20/2021) O (12/01/2020) O (01/20/2021) O (02/18/2020) E (07/22/2020) E (11/03/2020) O (12/01/2020) HK$40.80 Rmb15.83 Rmb21.00 Rmb59.98 NT$2,710.00 Rmb7.79 HK$223.00 Rmb107.40 HK$26.85 Rmb41.56 HK$28.60 Rmb36.72 NT$362.50 NT$935.00 NT$20.85 Rmb5.88 NT$209.50 NT$106.50 NT$344.00 NT$19.75 Rmb25.62 Rmb9.28 NT$97.10 Morgan Stanley Research 67 M M TCL Corp. (000100.SZ) Tianma Microelectronics (000050.SZ) Visionox Technology Inc (002387.SZ) Wuhu Token Science Co. Ltd. (300088.SZ) Howard Kao E (06/11/2019) U (01/24/2018) E (02/18/2020) E (12/01/2020) Rmb7.77 Rmb14.61 Rmb9.45 Rmb8.64 Accton Technology Corporation (2345.TW) Acer Inc. (2353.TW) Asustek Computer Inc. (2357.TW) Compal Electronics (2324.TW) Flexium (6269.TW) Guangdong Fenghua Adv. Tech. (Hldg) Co (000636.SZ) Inspur Electronic Information (000977.SZ) Kinsus Interconnect Tech. (3189.TW) Legend Holdings Corp (3396.HK) Lenovo (0992.HK) Pegatron Corporation (4938.TW) Quanta Computer Inc. (2382.TW) Shengyi Technology Co Ltd. (600183.SS) Shennan Circuits Co Ltd (002916.SZ) Tripod Technology (3044.TW) Wistron Corporation (3231.TW) Wiwynn Corp (6669.TW) Yageo Corp. (2327.TW) Zhen Ding (4958.TW) Ray Wu, CFA O (03/19/2021) U (02/01/2018) U (07/23/2021) U (12/18/2019) E (04/27/2018) E (05/12/2021) E (02/12/2019) O (02/06/2020) E (03/29/2018) E (07/26/2018) U (03/26/2021) E (07/28/2021) O (07/22/2020) O (07/22/2020) E (08/11/2020) E (07/26/2018) O (08/01/2019) E (05/12/2021) U (04/24/2020) NT$325.00 NT$26.05 NT$323.00 NT$21.80 NT$120.50 Rmb34.14 Rmb32.40 NT$170.50 HK$11.96 HK$8.02 NT$67.80 NT$76.40 Rmb27.51 Rmb105.36 NT$116.00 NT$27.20 NT$895.00 NT$533.00 NT$100.00 Chroma Ate Inc. (2360.TW) Wuhan Jingce Electronic Group Co Ltd (300567.SZ) Sharon Shih O (07/25/2019) E (07/03/2020) NT$176.50 Rmb66.30 AAC Technologies Holdings (2018.HK) Asia Vital Components Co. Ltd. (3017.TW) Auras Technology Co Ltd (3324.TWO) Catcher Technology (2474.TW) Delta Electronics Inc. (2308.TW) Foxconn Industrial Internet Co. Ltd. (601138.SS) Foxconn Technology (2354.TW) GoerTek Inc (002241.SZ) Guangzhou Shiyuan Electronic Tech Co Ltd (002841.SZ) Hon Hai Precision (2317.TW) LandMark Optoelectronics Corporation (3081.TWO) Lingyi Itech Guangdong Co (002600.SZ) Lite-On Technology (2301.TW) Luxshare Precision Industry Co., Ltd. (002475.SZ) MLS Company Limited (002745.SZ) Sanan Optoelectronics (600703.SS) Sunonwealth Electric Machine Industry Co (2421.TW) Tong Hsing (6271.TW) Visual Photonics Epitaxy Co Ltd (2455.TW) Stock Ratings are subject to change. Please see latest research for each company. * Historical prices are not split adjusted. ++ O (05/14/2021) E (05/14/2021) U (04/23/2021) O (07/13/2017) O (07/10/2019) E (08/16/2016) O (06/24/2021) U (04/28/2021) O (02/25/2020) E (07/05/2021) E (09/28/2020) E (07/31/2020) O (10/24/2016) E (09/10/2018) U (06/17/2019) E (05/14/2021) E (03/18/2019) E (03/20/2020) HK$47.15 NT$72.30 NT$167.00 NT$177.50 NT$280.00 Rmb11.78 NT$63.30 Rmb40.72 Rmb111.77 NT$109.00 NT$233.50 Rmb6.81 NT$62.60 Rmb41.81 Rmb16.78 Rmb40.64 NT$38.10 NT$237.00 NT$123.00 68 M M INDUSTRY COVERAGE: Technology - Semiconductors Company (Ticker) Rating (As Of) Price* (08/10/2021) Dominik Olszewski, CFA ams AG (AMS.S) ASM International NV (ASMI.AS) ASML Holding NV (ASML.AS) Infineon Technologies AG (IFXGn.DE) Soitec SA (SOIT.PA) STMicroelectronics NV (STM.PA) Stock Ratings are subject to change. Please see latest research for each company. * Historical prices are not split adjusted. E (12/10/2020) E (06/19/2019) O (06/19/2019) O (05/28/2020) E (03/12/2021) O (06/19/2019) SFr 18.22 €318.70 €674.30 €35.34 €201.80 €36.43 Morgan Stanley Research 69 © Morgan Stanley 2021 The Americas 1585 Broadway New York, NY 10036-8293 United States Tel: +1 (1) 212 761 4000 Europe 20 Bank Street, Canary Wharf London E14 4AD United Kingdom Tel: +44 (0) 20 7 425 8000 Japan 1-9-7 Otemachi, Chiyoda-ku Tokyo 100-8104 Japan Tel: +81 (0) 3 6836 5000 Asia/Pacific 1 Austin Road West Kowloon Hong Kong Tel: +852 2848 5200
下载完整pdf文档
关注微信
