Morgan Stanley's Mike Wilson Says the Chip Rally Is Stalling and Names the Sectors He's Buying Instead

The Warning Shot

The semiconductor trade has been the market's golden child for the better part of two years. Mike Wilson, Morgan Stanley's Chief Investment Officer and Chief U.S. Equity Strategist, now thinks the trade is passing its peak. The signal is evident in the data.

The PHLX Semiconductor Index (SOX) dropped 7.9% in a single week in late June 2026, according to MarketWatch. The week before that, it had jumped 7.3%. That kind of violent oscillation, Wilson told clients in a note published June 29, is a sign that "historically high exposure levels in the space" are becoming harder to maintain.

The Nasdaq Composite posted a five-day losing streak through that same late-June stretch, shedding 4.6%. That decline was more than double the S&P 500's drop over the identical period, according to MarketWatch.

The Silver Comparison

Wilson's most pointed warning is the comparison he draws between semiconductor stocks and silver's price dynamics earlier in 2026. He frames both as liquidity-driven commodity rotations rather than new structural phases.

His exact language: semis may be "the latest commodity to boom this year." He sees the chip rally as tracking silver's trajectory with roughly a four-month lag. For semiconductor bulls who have been buying every dip, that framing is significant.

His concern isn't that AI investment is ending. Capital expenditure forecasts for the major hyperscalers — Microsoft, Amazon, Meta, and peers — have been revised sharply upward. Projections now sit at $805 billion in combined spending for 2026 and $1.116 trillion for 2027, according to KuCoin's reporting on Wilson's analysis. The infrastructure build is real and ongoing.

The distinction Wilson is drawing is between the secular trend and the cyclical trade. AI as a platform isn't going anywhere. But the specific chip stocks that led the infrastructure buildout phase may be due to pass the baton.

Hyperscaler Weakness as Leading Indicator

Wilson's rotation thesis rests on a specific observation: hyperscaler stocks have been weak even as those same companies are committing to record capital expenditure. In Wilson's framework, that's the leading indicator.

His reasoning, as reported by MarketWatch: "rotation away from the hyperscalers may be a leading indication that semis — key spending beneficiaries of hyperscalers — could now see a period of underperformance as earnings per share revisions breadth for the space comes up against historical extremes."

Hyperscaler underperformance came first. Chip underperformance, in Wilson's view, followed on schedule.

Heading into Monday's session, the Roundhill Magnificent Seven ETF (MAGS), which holds several major hyperscalers, was down approximately 1.3% for the year through the most recent close. The SOX, despite its recent two-week slide, remained up roughly 78% year-to-date through the same period, according to MarketWatch.

What Wilson Is Buying

On beaten-down hyperscalers, Wilson's team sees an entry point. They flagged in a Monday note to clients that recent speculation about Meta potentially launching a cloud business to sell excess AI computing capacity adds to the group's appeal. His team wrote that hyperscalers "possess attractive optionality within the AI ecosystem: strong core businesses, the ability to participate in and lead the agentic application layer development and implementation, as well as an underappreciated cost-cutting lever."

Beyond hyperscalers, Wilson's preferred plays as reported by MarketWatch on July 6:

• Consumer discretionary goods, which he sees as the best risk-reward given falling oil prices and a shift in wallet share from services to goods.
• Biotech (IBB), flagged as one of the most rate-sensitive areas of the market. Historically it generates nearly 20% annualized returns in elevated and falling rate regimes. His team's specific call: "policy rate expectations remain overly hawkish this year in the context of our house call for core CPI to stay contained below 3%."
• Transports and regional banks, beneficiaries of what he sees as an underappreciated earnings recovery broadening across the market.

The Case for Staying in Chips

The bull case deserves a fair hearing. Wilson himself does not say the semiconductor cycle is over. His words were explicit: "This doesn't mean the cycle is over." The trillion-dollar-plus capex projections for 2027 represent real demand that has to flow somewhere, and chip companies are the primary suppliers of the hardware required to build AI data centers at scale.

Investors who've bought prior Nasdaq dips over the last two years have generally been rewarded. EPS revision breadth hitting "historical extremes" can stay extreme longer than strategists expect. And a year-to-date SOX gain of roughly 78% through late June, even after the late-June pullback, represents genuine fundamental performance, not pure speculation.

The counterargument is not irrational. It's just that Wilson's framework suggests the easy money in this specific rotation has already been made.

What to Watch Next

Wilson's rotation thesis will be tested on upcoming hyperscaler earnings calls. If Microsoft, Amazon, and Meta maintain or raise their AI capital expenditure guidance while their stock prices stabilize, his case for the leadership shift gains traction. If those companies start trimming capex projections, the chip correction could deepen into something more serious. Correlated risk assets, including crypto, would likely feel it too, given the shared high-beta dynamic that KuCoin's analysis of Wilson's note flagged.

The unresolved question: can hyperscaler stocks recover their footing while the companies are still spending at record pace? The capex numbers say yes. The stock charts, so far, disagree.