AMD, Intel, Nvidia: AI Profit Fears Hit Chip Stocks

The chip sector sold off Monday as traders weighed whether the AI boom's huge capital outlays will ever translate into bottom-line earnings growth. AMD fell more than 3% and Intel dropped 2.5%, while Nvidia managed a 1.8% gain after closing Friday near session lows.

AMD has an Alpha Score of 45, a "Mixed" label that reflects the uncertainty around its MI300 accelerator line. The stock has been volatile since its Q1 earnings call, where executives projected the data-center GPU business would hit $4 billion in revenue this year but declined to give a timeline for when that would turn profitable. Nvidia, with a 71 Alpha Score and a "Moderate" label, has held up better on the strength of its CUDA software moat and a forward P/E that is still below 30 after the correction. Intel at 20 and "Weak" is the laggard; its foundry losses ran to $7 billion last year, and the new chip design unit has not yet produced a shipping product that could offset them.

Traders said the selloff was less about any single company's news and more about a rotating concern that has been building for months. AI infrastructure spending hit a record $60 billion in the first quarter across the big U.S. cloud providers, yet none of them have reported a meaningful improvement in cloud margins. The hyperscalers buy chips by the tens of thousands, but their customers–enterprise AI adopters–have been slow to deploy applications that generate recurring revenue. At some point, the argument goes, either demand catches up to the capital spending or the spending slows.

Tuesday brings no chip-specific catalysts. The broader market will watch the 10-year Treasury yield, which rose to 4.48% Monday, adding to the pressure on the high-duration names that have driven the AI trade. The two-year yield sits at 4.70%, and if that gap narrows, the valuation premium on growth stocks becomes harder to justify. AMD and Intel are both leveraged to higher yields through their balance sheets; Nvidia, with no debt and $35 billion in cash, is less exposed on that front.

The next real test for the group is Nvidia's annual GTC conference in March, where Jensen Huang typically unveils the next-generation GPU architecture. Until then, the AI trade will sit on a data diet of hyperscaler capex plans and enterprise software earnings. Both are due in April.