Morgan Stanley told clients this week that softness in U.S. semiconductor stocks could be an early indication that market gains are becoming broader, and that money may rotate into a wider set of sectors. In a note dated Monday, the brokerage singled out AI "hyperscalers" - technology companies spending heavily on data centers - as a likely destination for flows if the AI-driven cycle shifts away from chipmakers.
The firm pointed to major technology companies such as Alphabet and Amazon, noting both have committed billions to expand AI infrastructure. Those large-scale investments have powered sharp gains in semiconductor share prices, the note said, but clear evidence that AI products will generate sufficient returns to justify the outlays has not yet emerged.
Morgan Stanley added that there could be "more capex discipline in the near-term," suggesting companies might rein in further spending. The brokerage also said hyperscaler stocks have already passed a period of relative underperformance, implying they could be positioned to regain investor attention as the market reallocates.
The bank observed that Alphabet, Amazon, Meta Platforms and other big tech names experienced heavy selling in June. At the same time, the Philadelphia SE Semiconductor index climbed 11% last month. That momentum faded more recently: Morgan Stanley noted the chip index has fallen over 11% in the last two weeks. The Roundhill Magnificent Seven ETF - used as a proxy for the seven largest Wall Street technology companies - has recovered some of its earlier losses.
Beyond company-specific dynamics, Morgan Stanley said broader market conditions are also influencing the shift. Investors paring back expectations for additional interest-rate increases from the U.S. Federal Reserve, together with a decline in crude oil prices, are contributing to the rotation out of the previously red-hot semiconductor trade.
Impacted sectors: Hyperscalers and big-cap tech, semiconductors, consumer discretionary, transports, and biotechnology are directly referenced as affected by the described rotation.