Morgan Stanley's quantitative strategy team is warning that the recent sell-off in semiconductor and data center-related stocks is not yet over, with further downside risks still looming over the medium term, and it advises investors to use any rebound opportunities to trim these positions.
According to a report from the trading desk, on October 27, Morgan Stanley's Ukyo Haraguchi quantitative strategy team pointed out that during the third to fourth week of October, rising concerns about financing and credit risks in AI infrastructure investment triggered a sharp sell-off in semiconductor and data center-related stocks.
Although these sectors may see technical rebounds in the short term, the team believes the medium-term correction has not yet fully played out and recommends "increasing the pace of position reduction."
The report simultaneously recommends increasing allocations to low-volatility stocks and financially sound value shares, while viewing AI application companies as a medium-to-long-term structural growth opportunity, and it screened out 50 Japan-listed stocks with high correlation to the AI adopters basket for investor reference.
This strategy framework shows that market focus is shifting from upstream capital expenditure and infrastructure in the AI supply chain toward downstream AI adoption and applications.
Sell-Off Exceeds Expectations, Medium-Term Risks Remain
Morgan Stanley's quant team had already cautioned in its October 9 report titled "Navigating the Heightened Challenges of the Late-Stage AI Rally" that against a backdrop of rising interest rates and elevated valuations, investors should not be complacent about the sustainability of the AI rally, and it recommended tactically adding to low-volatility stocks while retaining exposure to financially sound value shares to cope with potential reversals in a highly concentrated AI-driven market.
The team said the speed and magnitude of the recent correction exceeded its expectations, but the direction of market rotation — with low-volatility and value factors outperforming — was consistent with its prior judgment, and this trend has been reflected in both Japanese and U.S. markets.
On the valuation front, the report noted that despite a significant pullback, the median forward price-to-earnings ratio of AI and semiconductor-related stocks remains at historically high levels, meaning that price corrections alone are not sufficient to digest valuation pressure.
(Median forward P/E for AI and semiconductor-related stocks over the next 12 months: despite the pullback, valuations remain elevated by historical standards.)
The report also compared the current trajectory of the Philadelphia Semiconductor Index with that of the 2000 internet bubble period, noting that the two exhibit similar price patterns.
(Semiconductor index price movements during the internet bubble period and the current period show similarities.)
AI Theme Divergence: Upstream Under Pressure, Downstream Resilience Stands Out
Morgan Stanley Research maintains three major global AI-related baskets: AI Enablers, AI Infrastructure, and AI Adopters.
The report shows that the first two have greater exposure to semiconductors and data centers and suffered significant declines during the recent correction, while the AI Adopters basket — companies that integrate AI into their own services and operations — had previously moved largely in sync with the first two but has demonstrated clear relative resilience during this round of selling.
(Performance of global AI-related portfolios: despite the recent pullback, AI consumer companies remain resilient.)
The team believes this divergence indicates that the market's focus is shifting: from the upstream of the AI supply chain (capital expenditure and infrastructure) to the downstream (AI adoption and application deployment).
The report notes that the AI rally has so far been dominated by semiconductor and data center-related stocks, but at a time when there is widespread concern about the outlook for these stocks, the possibility is rising that AI application companies will become the next main theme of the global AI trade.
The report also cites the view of Morgan Stanley thematic research analyst Stephen Byrd, recommending a "barbell strategy" — with one end allocated to companies addressing infrastructure bottlenecks (such as power supply) and select semiconductor and infrastructure names, and the other end allocated to early-stage software enablers and AI application companies.
Notably, the enablers category also covers infrastructure and security software providers, and is not limited to semiconductor and data center companies.
Screening Japanese AI Application Stocks: 50 High-Correlation Names
To help investors identify AI application opportunities in the Japanese market, the team conducted a systematic screening of the TOPIX 1000 constituents.
The specific methodology was: calculate each stock's beta to the AI Adopters basket (controlling for TOPIX returns), then subtract its beta to the AI Enablers basket to derive a "net AI adoption beta," and select the top 50 stocks based on this ranking.
(Performance of high net-beta AI adopter stocks.)
From a historical performance perspective, this portfolio of 50 stocks was generally weak in the first half of 2026, partly constrained by market concerns about a "SaaS apocalypse," but performance has improved since the second half of the year.
The team noted that although short covering may have contributed to part of the rebound, combined with the resilience of the global AI Adopters basket, this trend deserves close attention and may be an early signal of the market rotating toward AI application companies.
The 50 screened stocks are mainly concentrated in IT services and other sectors, covering large-cap companies such as Recruit Holdings, Nintendo, NEC, Nomura Research Institute, and BAYCURRENT, as well as small and mid-cap tech companies such as PKSHA TECHNOLOGY, Appier Group, and SANSAN.
Strategy Summary: Reduce Upstream, Increase Downstream, Hold Low Volatility
Overall, the current allocation recommendations from Morgan Stanley's quantitative team can be summarized into three main lines:
First, for AI capital expenditure-related stocks such as semiconductors and data centers, seize rebound opportunities to gradually reduce positions;
Second, continue to increase allocations to low-volatility stocks and financially sound value shares to cope with market volatility;
Third, treat AI application companies as a medium-to-long-term thematic allocation direction, using the net AI adoption beta screening framework to find structural opportunities.
The team emphasizes that AI applications provide a thematic stock-selection perspective distinct from traditional style factors, which can complement the aforementioned factor allocation strategies.
As the AI rally enters a more challenging late stage, distinguishing between the AI capital expenditure side and the AI application side will be key for investors to control risk and capture opportunities.
Comments