After the sharp declines in the Philadelphia Semiconductor Index and the Nasdaq, Morgan Stanley says "get ready to add positions"

Title context: After the sharp declines in the Philadelphia Semiconductor Index and the Nasdaq, Morgan Stanley says "get ready to add positions" Text: In the view of top strategists at Wall Street financial giants such as Morgan Stanley, the logic of contrarian position-building and buying on dips in the global AI computing power theme and broadly in technology stocks, which have recently been hit hard by selloffs, is taking shape. However, whether technology can stage a bottoming rebound and then sustain an upward trend has entered a dual verification period of earnings disclosure season and AI financing conditions. Michael Wilson, Chief Equity Market Strategist at Morgan Stanley, wrote emphatically in his latest research report that the valuation adjustment in most stocks that has lasted for several months is nearing its end; the market may complete this round of adjustment quickly, or it may drift lower into November, and investors should prepare to increase equity market risk exposure during this process. Wilson said he still prefers high-quality companies with strong fundamentals, especially sectors such as capital goods where earnings momentum continues to improve substantially. Therefore, this is the latest judgment on the investment opportunities in the equity market after a sharp downturn and adjustment. The institution's bullish view is basically consistent with that of another Wall Street financial giant, Bank of America, namely that after a major adjustment, the valuations of most technology giants, including memory chip giants and leaders in the AI computing power theme, are close to historical cyclical troughs. Fundamental indicators such as core chip prices and momentum, orders, and earnings trends related to AI infrastructure have not shown any corresponding degree of deterioration, and the buy-on-dip opportunity after the pullback appears to have arrived. Currently, new cloud leaders such as CoreWeave have exposed the most sensitive financing link in this round of adjustment. CoreWeave fell 25.21% last week, expanding its year-to-date decline to 14.27%; the yield on its 2031 bonds briefly touched about 16% last week, and the five-year credit default swap (CDS) spread exceeded 1,100 basis points. The five-year cumulative default probability implied by relevant market pricing exceeds 60%. This probability is a market model result that includes a risk premium and is affected by recovery rate assumptions. Nscale, backed by Nvidia, also encountered IPO price pressurerelative to a target valuation of about $30 billion, some investors argued for $16 billion to $19 billion, a discount of about 37% to 47%, and pricing has not yet been completed. Therefore, what the market has recently been re-examining on a large scale is that expensive spending on core hardware components of AI infrastructure such as GPU/TPU/AI ASIC and data center CPUs, as well as large data center construction and power-related spending, occurs first, while revenue must wait for power-on, acceptance, and customer activation; the more expensive the financing, the weaker the project's ability to withstand delays and underutilization. Just as shares in the memory chip sector, which has benefited from an unprecedented AI infrastructure boom, have fallen sharply recently, the Bank of America analyst team believes that the market has already priced in advance a significant slowdown in future price increases, supply expansion, and AI financing pressure. However, the resilience of spot prices, HBM upgrades and iterations, the tiered storage demand brought by the AI agent inference boom, and even the overall AI computing resource demand expanding explosively can still support strong earnings expansion for AI computing power theme stocks for a longer period. It is worth noting that just as the global semiconductor sector plunged, institutions sharply reduced directional allocation exposure, but the semiconductor leveraged ETF long favored by retail investors (that is, the triple-leveraged bullish ETF tracking the Philadelphia Semiconductor IndexU.S. ticker "SOXL") saw huge inflows. Potential covering demand and leveraged funds can be said to have jointly amplified the market's sensitivity to strong fundamental catalysts during the earnings season. The latest bullish views of Morgan Stanley and Bank of America echo the market's latest movestechnology stocks have already shown signs of breathing room, but the damage left by the staged selloff is still evident: last week the Nasdaq 100 fell about 5.5%, and the Philadelphia Semiconductor Index fell for eight consecutive trading days, with a weekly drawdown of about 11.5%, directly entering technical correction territory; as of October 16, South Korea's KOSPI had fallen 17.11% month-to-date, with Samsung Electronics and SK Hynix down 19.74% and 28.38%, respectively. As of the Asian market close on October 19, the KOSPI rebounded 1.76%, but Samsung and Hynix rose only 1.16% and 0.39%, respectively; before the U.S. market open, Nasdaq futures rose about 1%, indicating that buying interest is beginning to return tentatively, and memory chip leaders appear to be leading the AI computing power theme after many days. Morgan Stanley's Wilson calls for "buying the dip": the adjustment is nearing its end, get ready to increase risk exposure Wilson is one of Wall Street's most influential sell-side strategists: he joined Morgan Stanley in 1989, has worked in investment banking, institutional equities, and investment strategy roles, and is currently Chief U.S. Equity Strategist and Chief Investment Officer. He also ranked first in the portfolio strategy category of Institutional Investor's 2022 All-America Research Team, far ahead of other well-known strategists. His influence on the equity market mainly comes from integrating earnings, valuations, interest rates, and institutional positioning into allocation judgments; his positive bullish signal of "get ready to add positions" can usually prompt large institutional investors to reassess the opportunity cost of being overly defensive. Morgan Stanley's Michael Wilson said that as the months-long technology stock correction in some parts of the equity market nears its end, investors should be prepared to increase their portfolios' equity risk exposure. Although the S&P 500 is only about 2% below its all-time high, market valuations have fallen significantly over the past few months, while market breadth within the index has also deteriorated severely. The strategist said this market patterncharacterized by the so-called quality factor outperforming and a sharp decline in the equal-weighted S&P 500 indexis about to end, thereby providing investors with a favorable window to buy stocks on dips. "We recommend investors prepare for the market to complete the adjustment very quickly, or to extend through slower grinding declines into November, because in our view, that is when risk exposure should be increased rather than avoided," Wilson said in a client report. The strategist remains biased toward quality stocks and emphasized that economically sensitive sectors such as financials, industrials, and consumer discretionary have borne one of the main shocks of this round of major valuation adjustment. He said that valuation cuts in some market areas have already created entry opportunities in areas with improving earnings-per-share momentum, such as capital goods and technology. As shown in the chart above, U.S. equity valuations have fallen significantly over the past few months. Wilson's view may reassure some investors. These investors worry that the recent sharp rise in U.S. Treasury yields and high-yield corporate bond yields could undermine the unprecedented AI investment boom that is increasingly reliant on new debt financing. Given that the AI computing infrastructure boom has played an important role in driving this four-year global equity market super bull run, some market participants said they have seen signs of bubble formation. Goldman Sachs strategists also noted on Monday that investor positioning is becoming increasingly defensive; JPMorgan strategists share the optimistic sentiment of Morgan Stanley's Wilson, saying that the current stabilization in bond yields will be "a great help" in supporting the stock market. Broader confidence in the equity outlook mainly comes from increasingly strong earnings growth prospects driven by the AI computing power theme. According to the latest consensus forecasts of Wall Street analysts compiled by Bloomberg Intelligence, overall profits for S&P 500 constituent companies are expected to surge 24.6% year-over-year in the third quarter. Wall Street financial giant Jefferies said recently that, driven by the dual engines of the AI investment boom and AI-related companies' better-than-expected earnings growth, the S&P 500 is expected to soar to 8,000 by the end of 2026 and further reach 9,000 in 2027. Jefferies' core logic is clear and powerful: in a cycle where AI-driven earnings growth exceeds the historical average by more than twofold, fighting the earnings trend is dangerous. Jefferies' base-case forecast of 8,000 for the S&P 500 in 2026 is based on earnings per share (EPS) of $373 (up 35% year-over-year, far above the market consensus of 29%) and a price-to-earnings ratio of 21.5x. A recent JPMorgan research report shows that trading in advance at the share price level for a cooling memory chip cycle can occur at the same time as the global AI computing power supply chain, including memory chip components, maintains a strong earnings super-cycle for a longer period. The JPMorgan Asia-Pacific analyst team led by Jay Kwon accordingly favors Samsung Electronics, SK Hynix, Kioxia, as well as Samsung Electro-Mechanics, ISU, and ISC over the next six months, betting that the memory upcycle will last longer and that demand for related substrates, PCBs, and testing will continue. JPMorgan explained from the industry side's strong earnings trajectory why the memory pullback may create opportunities: market concerns about the sustainability of AI capital expenditure have not necessarily translated equivalently into deterioration in memory orders and earnings. AI inference requires continuous reading of model weights, saving context and KV caches, and handling data demand at different tiers through server memory and SSDs; HBM upgrades will also consume wafer, packaging, and yield-ramp resources, making it difficult for effective supply to keep up quickly with demand. JPMorgan's semiconductor research report on the Korean stock market shows that macro-level concerns about the sustainability of AI investment are clearly disconnected from the memory industry's bottom-up supply and demand data. Therefore, it regards the third quarter as an "expectations reset quarter" and, over the next six months, is core overweight Samsung Electronics, SK Hynix, Kioxia, extending to Samsung Electro-Mechanics, as well as ISU and ISCleading companies in AI PCBs and test sockets, respectively. The "duration trade" here bets that the upcycle will last longer and high profits will be maintained for longer: even if the pace of DRAM/NAND price increases slows, as long as demand for HBM, server DRAM, and enterprise SSDs continues to absorb new capacity significantly and achieve a notable growth trajectory, the rapid profit decline previously priced in by the market may need to be revised. This also explains why memory stocks can fall before earnings peak, yet may rebound early once earnings duration is verified; what determines the trading opportunity is the continuously evolving gap between the actual earnings path and the expectations implied by share prices. Retail investors rush in as valuations hit staged trough levels; the AI-driven technology stock counteroffensive curve still lacks cash flow confirmation The latest judgment from Wall Street legendary investor Howard Marks, co-chairman of Oaktree Capital, provides a contrarian perspective on current pessimism: truly dangerous bubbles are usually accompanied by a loss of vigilance toward risk, while continued discussion of bubbles means investors have not yet fully abandoned discipline. The forward P/E ratios of current leading technology companies are significantly lower than during the internet bubble era, and are also related to the large-scale profits these companies have already generated. His judgment also overlaps with the latest views of Bank of America and Morgan Stanleywhether low P/E ratios can constitute an opportunity depends on whether earnings can be sustained; if the market underestimates the duration of high profits, valuations may recover, while if earnings forecasts are too high, low multiples may merely be a cyclical trap. Howard Marks emphasized that real bubbles are often accompanied by collective neglect of risk, while the current market's constant discussion of bubbles shows that doubt and discipline still exist. Capital behavior also shows this split: Goldman Sachs team statistics show that about $9.5 billion flowed into leveraged ETFs last week, of which about $7 billion entered the triple-long semiconductor ETFthe highly leveraged ETF with U.S. ticker SOXL; at the same time, net exposure of U.S. fundamental long-short funds fell to the lowest historical percentile, while gross exposure remained at 213.9%. All of this means directional bets have contracted, but long and short positions remain large in scale: once earnings weaken pessimistic expectations, adding longs and short covering may jointly drive a rebound. It is worth noting that leveraged ETFs attracting retail capital reflects a strong willingness to buy the dip, but cannot alone prove that long-term institutional funds have completed reallocation. Bank of America's latest research report can be said to provide more specific earnings and valuation evidence for a potential bottoming rebound in AI computing power-led technology stocksbased on market consensus earnings forecasts for 2027-2028, Samsung and Hynix trade at only about 2-3x P/E, showing cyclical trough valuation characteristics, but the actual memory market has not yet seen the price collapse typical of traditional recessions. The report shows that the spot price of 16Gb DDR5 rose 3% during the week to $60.6, and it is expected that the contract price of related DDR5 will rise about 17% month-over-month in October; even assuming a single-digit year-over-year decline in 2028 earnings per share, Bank of America still expects revenue and operating profit to remain near record levels. In the view of the Bank of America analyst team, the market has overpriced future earnings declines. The underlying strong supporting factors for memory chips and even AI computing resources cited by Bank of America include the introduction of HBM4 and demand for HBM, SOCAMM, and enterprise SSDs driven by upgrades of advanced GPUs and ASICs: model weights, long-context caches, and massive AI inference workloads at high concurrency levels require different tiers of memory and storage, as well as coordination across a broader AI computing system including CPUs, GPUs/TPUs, and high-speed optical interconnects in data centers. Even changes in the configuration of a single GPU cannot be directly equated with a contraction in overall data center storage demand. The interest rate environment and U.S. Treasury yield environment have also improved marginally. Statistics newly compiled by the U.S. Treasury show that from October 9 to 16, the 10-year and 30-year U.S. Treasury yields fell by 15 and 17 basis points, respectively, to 5.09% and 5.43%; before the U.S. market open on the 19th, Brent crude fell back to about $107, and the 10-year yield was about 5.085%. However, JPMorgan pointed out that the recent rise in long-term bonds also includes safe-haven demand from Wall Street funds triggered by the AI selloff, and cannot be directly regarded as the removal of inflation pressure. For the AI computing power supply chain and even technology stocks, a sustained counteroffensive requires three pieces of evidence to reinforce one anotherAI revenue and AI computing orders related to the AI computing power theme from large cloud providers and new cloud players materializing to support capital expenditure, new cloud project delivery and collections improving debt-servicing capacity, and U.S. Treasury yields and corporate credit spreads easing together. When earnings prove that computing power investment can be converted into profits and cash flow, valuation compression is more likely to become the opportunity to add risk assets that Wilson refers to, and institutions with low net exposure will also have more reason to increase risk again.