Morgan Stanley states that the AI trend is far from over, but it's time for the hyperscale cloud providers to take the lead.
In his latest weekly report, Morgan Stanley's chief US equity strategist Michael Wilson delivered a clear message to clients: reduce holdings in semiconductors and pivot towards hyperscale cloud computing companies. This is not a bearish call on AI, but rather a rotation—the fourth such shift within the current AI investment cycle, according to Wilson.
After a historic rally since late March, semiconductor stocks have recently shown signs of cooling off. A high-beta momentum basket, including memory and chip stocks, recorded its largest two-day decline since the COVID-19 pandemic. Wilson believes this pullback "likely has further to go."
This assessment is not a sudden shift. Wilson originally outlined a framework for a "broadening trade" in his November 2025 annual outlook. The core thesis was that following a rolling recession ending in April 2025, the US economy would enter a new expansion cycle, leading to better-than-expected earnings growth. Market leadership should consequently broaden out from the initial beneficiaries of AI capital expenditure to a wider range of sectors.
This broadening trade was interrupted in February 2026 by the Iran conflict. Soaring oil prices and a market repricing of Fed rate hikes stalled the diffusion, allowing semiconductor stocks to regain sole dominance on the back of the AI compute narrative. Now, with oil prices retreating and inflation expectations stabilizing, Wilson believes the conditions are ripe once again.
Analogous to "Silver's Peak": Memory Chips Pose the Greatest Risk
Wilson drew a specific analogy in his report: the trajectory of semiconductors is highly reminiscent of silver. He cites two reasons: first, both have experienced parabolic price surges; second, both are heavily tied to commodity markets, which are historically prone to sharp booms and busts.
The firm first introduced this analogy in early June, and it now appears to be playing out. Wilson further noted that this correction is likely to be led by the Memory sub-sector—as it is the "most commodity-like" segment within the semiconductor complex, characterized by high price elasticity and rapid reversals.
The pronounced sell-off in semiconductor stocks following Micron's earnings report, in Wilson's view, confirms that the market has begun to focus on the "peak rate of change in earnings revisions" as a core concern.
Meta's Statement Ignited the Fuse
The immediate catalyst for this rotation was an announcement from Meta. Last week, Meta announced it would begin selling its excess compute capacity to external customers. This move signaled to the market that the growth rate of capital expenditure by hyperscalers may be reaching an inflection point.
Wilson wrote that the severe performance divergence between hyperscale cloud providers—such as Microsoft, Google, Amazon, and Meta—and semiconductor stocks is fundamentally unsustainable. This is because chip demand is ultimately reliant on the capex appetite of these cloud providers. Historical patterns show that when this divergence reaches extremes, a "mean reversion" often occurs: cloud providers either lower capex guidance or announce a shift in direction, triggering a correction in semiconductor stocks. Meta's announcement provided precisely this rationale.
It's crucial to note that Wilson explicitly stated this does not signal the end of the AI capex cycle, but rather a meaningful mid-cycle reset and rotation. In his words: "This is a peak in the rate of change, not the peak of the entire capex cycle." In fact, since ChatGPT's launch in November 2022, there have been three such mid-cycle corrections; this marks the fourth.
The Rationale for Favoring Cloud Over Chips Now
Wilson's logic is that hyperscale cloud providers—like Amazon AWS, Microsoft Azure, and Google Cloud—have underperformed in recent months, yet their fundamental support remains robust. He lists three reasons:
First, their core businesses are solid. Cloud providers have strong underlying businesses that are not entirely dependent on the AI capex narrative.
Second, they hold a unique position in the AI application layer. Wilson believes the market undervalues cloud providers' leading advantage in the "development and deployment of the agentic application layer."
Third, overlooked cost-cutting potential. Wilson terms this "an underappreciated cost-cutting lever."
Meanwhile, a "high capex/sales" factor tracked by Morgan Stanley, after a year of strong performance, is now showing signs of peaking. Cloud stocks have already underperformed and digested this pressure, whereas chip stocks may only be beginning to face it.
Broadening Trade Resumes: It's Not Just About Cloud
Beyond hyperscale cloud providers, Wilson identified other preferred sectors for a broadening market:
Consumer Discretionary Goods is Wilson's top pick. The logic hinges on a shift in consumer spending from services back to goods, improving pricing for goods, and strong earnings-per-share revisions. He views this as "the most compelling expression" of the broadening earnings story.
Transportation also stands to benefit from the onset of an economic expansion cycle.
Biotechnology serves as a proxy for interest-rate-sensitive sectors. Historical data shows biotech has delivered annualized returns near 20% in environments of high but declining interest rates. Morgan Stanley expects core CPI to stay below 3%, and current policy rate expectations remain overly hawkish. A correction in these expectations would directly benefit biotech. Additionally, a heating M&A cycle provides an extra catalyst for the sector.
Macro Backdrop: Falling Oil Prices Stabilize Rates, Creating Fertile Ground
Wilson's broadening thesis has a crucial macro underpinning: a significant drop in oil prices. Lower oil prices help stabilize bond yields, and stable rates are a key driver for broadening market action. The firm's baseline forecast is that falling energy prices, peaking tariff-related inflation, and manageable services and housing inflation will combine to keep the Fed on hold this year, rather than hiking.
The bond market is still pricing in roughly 1.5 rate hikes before the first quarter of next year. Wilson believes a correction in this overly hawkish expectation would be a positive surprise for equities.
He also highlighted Fed Chair Wash's comments at the Sintra conference, stating "inflation risks have subsided" while reaffirming the dual mandate of employment and prices. Combined with last week's weaker-than-expected nonfarm payrolls data, Wilson sees this helping to further dampen hawkish rate expectations, providing support for the broadening trade.
This is a Rotation, Not a Conclusion
Wilson offered a clear summary at the report's conclusion:
"The market is beginning to broaden as the index also moves into a consolidation/correction phase. This is happening."
"Within the AI winners, leadership has rotated for years. This is just the next rotation in the cycle playing out."
"This is just the next rotation—from semis to hyperscalers, and the other broadening trade candidates noted above."
The relative underperformance of semiconductor stocks post-Micron's earnings has made the market aware that "peak rate of change" is now a core focus. Meta's unexpected announcement about selling excess capacity solidified this expectation. Consolidation in the high capex/sales factor may further prompt other cloud providers to lower their capex guidance expectations. All of this is fueling the broadening market narrative.
Comments