Summary
U.S. semiconductor stocks extended a recent downturn into Thursday premarket trading, with investors signaling concern about elevated valuations and the large-scale AI investments being made across the tech sector. Losses were widespread across memory and chipmakers, while Meta Platforms stood out with a notable rally after reports the company is considering a cloud business to monetize spare AI compute.
Market moves
By 04:42 ET, shares of Micron and Western Digital were each down roughly 2.1%. Coherent and Marvell Technology slipped about 2% and 1.8%, respectively, while AMD, Intel and Microchip Technology moved approximately 1% lower.
The latest selling followed a session in which U.S. equities closed marginally lower on Wednesday, led by weakness in technology names although gains in Meta Platforms limited the overall pullback in the S&P 500 and Nasdaq. An industry index for semiconductors plunged 6.3% on Wednesday, making technology the laggard among S&P 500 sectors that day.
Some individual chip-related stocks suffered sharp losses in the prior session. Micron at one point fell as much as 10.6%, while Applied Materials, Lam Research, Allegro MicroSystems and Intel each declined by more than 9%.
Analyst perspective
JPMorgan analyst Nikolaos Panigirtzoglou noted that the persistent outperformance of semiconductor stocks - specifically AI chip and memory producers - relative to hyperscalers has become difficult to justify over the long term. The remark highlights a widening divergence between companies that build compute hardware and those that operate large-scale cloud infrastructure.
Meta’s rally and cloud prospects
Meta Platforms was an exception to the sector weakness, jumping 8.8% after a report that the company is assembling a cloud offering to sell spare AI computing capacity. According to that reporting, Meta is weighing whether to provide access to AI models hosted on its systems or to offer raw computing power for external customers.
Model developers and major tech companies have been scrambling to secure GPU capacity since the recent surge in demand for AI applications. Meta disclosed in April plans to spend as much as $145 billion on capital expenditures this year as it expands data centers and acquires the GPUs required to train and run large AI workloads.
Launching a cloud business would enable Meta to monetize unused capacity, which could be welcomed by investors concerned about its elevated spending plans. At the same time, such a move would position Meta against established cloud providers that currently lead the market for AI infrastructure.
Context and near-term calendar
The renewed volatility in chip and technology stocks emerged ahead of the U.S. monthly jobs report due Thursday. Markets will be closed on Friday in observance of the Fourth of July holiday.
Note: This article reports on premarket price movements and company plans as described in public disclosures and media reporting; it does not introduce new facts beyond those referenced in company announcements and market data cited here.