Anthropic CEO Dario Amodei urges AI slowdown; SoftBank, Kioxia shares plunge

The sell-off follows a weekend of public warnings from the people running America's leading AI labs. Dario Amodei of Anthropic used an essay published on the 12th to ask AI companies to slow down how fast they improve their models, and both Elon Musk and OpenAI's Sam Altman signaled agreement on X. Amodei's own framing was stark: within 6 to 12 months, he wrote, AI agents could "seize control of the entire internet and cause hundreds of billions of dollars in damage." Sam Altman added a concrete decision, saying OpenAI will not hold this year's IPO because of safety concerns. Concerns had already been raised by the resignation of Anthropic researcher Jacob Coxon, who said people developing AI seriously believe it could wipe out humanity by the end of this decade.

The reaction was not uniform. Several US lawmakers want new regulation, while President Donald Trump on the 12th called AI's critics "a very negative force" raising impossible scenarios and said he wants the US to remain the industry's leader. The two governments are expected to discuss AI safety as part of bilateral talks this month, and China's state-run Global Times attacked Anthropic's essay as a "Cold War guide" meant to hold back Chinese technology.

Investors split along similar lines. Michael Burry called the warnings "hype and bluster" and an excuse to hide either uncontrollable growth or slowing growth. Saxo Bank's Charu Chanana said the warnings could remain a drag on AI and semiconductor shares in the short term, while T. Rowe Price's Sebastien Mallet argued the bigger issue is who ultimately earns a return on the new capital spending — noting that AI will almost certainly change the world, but that does not mean every investment being made today will produce attractive returns. How far the pullback runs is likely to hinge on whether more lab leaders repeat Amodei's call or step back from it.