European tech stocks hit six-week low while Asian tech shares fell sharply after Anthropic chief executive Dario Amodei called on the AI industry to "slow down," warning that building the technology too fast is reckless. He cautioned that a swarm of AI agents could one day cause hundreds of billions of dollars in damage by taking over the internet. Sam Altman and Elon Musk both backed his call.
South Korea's KOSPI index dropped 3.7%, with chipmaker SK Hynix down 5.75%. SoftBank, which holds a stake in OpenAI, fell as much as 13% in Tokyo after Altman said ChatGPT's maker won't go public this year. Taiwan Semiconductor Manufacturing Company slipped 1.2% in Taipei.
Some AI experts have pushed back on Amodei's warnings. But investors are already pricing in a slower pace of AI development, and that has consequences for who foots the bill on the industry's data centre buildout. Ipek Ozkardeskaya, senior analyst at Swissquote, said the leases, debt and power commitments tied to that infrastructure don't disappear even if compute demand and revenue growth slow down.
That, she noted, pushes credit risk further into the AI story, especially for heavily leveraged data centre operators and the lenders backing projects built on optimistic demand assumptions, at a moment when interest rates and borrowing costs are expected to climb. The notable part, she added, is that this slowdown isn't being forced by a cash crunch at Big Tech or investor reluctance. It's coming from the people building the models themselves.
Russ Mould, investment director at AJ Bell, comments: "An AI-related sell-off doesn’t create the best backdrop for Anthropic’s planned IPO, which is already rumoured to be delayed by a month to November".
The ripple effects are worth watching closely. Many have leaned on AI tools and cloud infrastructure priced on the assumption that capacity would keep expanding cheaply. A slower buildout could mean tighter supply and higher costs for AI services down the line, particularly for firms renting compute rather than owning it. It's also a reminder to stress-test any vendor or partner whose business model depends on continued aggressive AI infrastructure spending, since credit troubles at a data centre operator or cloud provider can disrupt service for the small businesses that depend on them.