AI Safety Warnings Knock Chip Stocks Off Their Highs

Nvidia lost nearly 3.5% on Monday, but the sharper damage landed farther down the AI hardware chain. Marvell Technology fell about 7.5%, Intel dropped roughly 5.5%, and Micron Technology slid more than 5% as investors reassessed how much infrastructure the next generation of models will require.

The selling followed an unusual chorus from the people building the technology. Anthropic Chief Executive Dario Amodei argued over the weekend that frontier AI companies should deliberately pace improvements in model capabilities, pairing faster technical progress with independent testing, stronger safeguards and closer coordination with governments. OpenAI CEO Sam Altman endorsed the broad idea, while Elon Musk also backed Amodei’s warning.

The market reaction was less about an immediate cancellation of data-center projects than about the assumptions embedded in semiconductor valuations. Chipmakers have been priced for years of accelerating spending on training systems, networking equipment and high-bandwidth memory. If model releases become slower, or companies spend more time testing before deployment, investors may have to push some of that demand further into the future.

That matters most for suppliers whose growth depends heavily on hyperscale AI capital expenditure. Marvell, which sells custom silicon and interconnect technology, was hit harder than Nvidia, whose earnings are supported by a broader installed base and dominant position in accelerated computing. Micron, Intel and other semiconductor names also carry sensitivity to a change in the pace of server construction.

A second signal came from Altman’s disclosure that OpenAI will not pursue a public listing in 2026, saying an IPO would be ill-advised amid the current safety concerns. That decision removes a major potential catalyst for SoftBank Group, one of OpenAI’s prominent financial backers, and reinforces the sense that the industry’s funding cycle is entering a more complicated phase.

Still, Monday’s decline was not a verdict that AI demand has vanished. The Nasdaq fell only about 0.6%, suggesting investors were rotating within technology rather than abandoning it outright. Nvidia Chief Executive Jensen Huang rejected the idea that AI development should simply stop, and the White House has opposed a pause on competitiveness grounds, particularly with China.

The dividing line for markets is becoming clearer: slower frontier-model releases may hurt chip multiples, but broader AI adoption, inference demand and enterprise software spending can continue even if the race to build the largest model loses momentum.