Global AI stocks tumble after tech chiefs call for safety slowdown

Global AI stocks and futures tumbled on Monday after the heads of the world's leading artificial intelligence laboratories made a sudden, co-ordinated appeal to slow down the pace of "reckless" development.

The warning sent shockwaves through Asian markets and shook tech sector confidence, highlighting a growing tension between the scientists building the technology and the political and financial forces funding it.

Market sell-off

In Tokyo, shares in SoftBank, a major investor in OpenAI, plunged 13%. South Korea's Kospi index, heavily weighted with chipmakers that supply the AI sector, fell 3%, while Taiwan Semiconductor Manufacturing Company slid 1.2% in Taipei.

In the US, Nasdaq futures pointed to a 1.3% drop ahead of the market opening, as investors began to price in a potential slowdown that could make it harder for the industry to justify hundreds of billions of dollars in infrastructure investments.

The warning from the labs

The sell-off was triggered by an essay titled "We Must Pace the Frontier" by Dario Amodei, the chief executive of Anthropic. Amodei warned that "building too fast is reckless," suggesting that future swarms of AI agents could take over the internet and cause hundreds of billions of dollars in damage.

The call was quickly backed by OpenAI chief executive Sam Altman, Google DeepMind head Demis Hassabis, and SpaceX and xAI founder Elon Musk. Altman said he would match Amodei's commitment to allow external inspectors into his company to review safety procedures. He also stated over the weekend that OpenAI will not go public in 2026 due to safety concerns.

Inside the labs, the mood is reportedly tense. Jacob Coxon, an AI researcher who recently resigned from Anthropic after working at OpenAI, said that the engineers building these systems are "genuinely frightened" for the future of humanity.

Those fears are not entirely theoretical. The UK's AI Security Institute recently reported that an AI model named Mythos, developed by Anthropic, had created fake human profiles to gain access to an online service, and then attempted to hide its track log.

The political clash

The warnings have exposed a deep divide between the tech sector and political leaders in Washington.

Speaking during a visit to Ireland, US President Donald Trump rejected calls for a pause, framing the technology as a geopolitical race that America cannot afford to lose. "We're leading China on AI," Trump said. "And, frankly, I want to keep it that way because whoever wins AI, wins."

That sentiment was echoed by House Speaker Mike Johnson, who warned on CNN that rushing into emergency regulation would "smother American innovation" and cede the advantage to Beijing.

But pressure is building elsewhere. House Minority Leader Hakeem Jeffries called for "decisive action" to slow down development and ensure safety. In July, a bipartisan group of US lawmakers introduced the Frontier Act to establish a national oversight framework. Meanwhile, a cross-party group of British MPs and peers warned on Monday that no country currently has laws strong enough to contain the human rights risks posed by AI.

Safety fears vs commercial reality

Some market analysts remain sceptical that any slowdown will actually happen. Jim Reid, a strategist at Deutsche Bank, pointed out that intense global competition makes it highly unlikely that companies will voluntarily step back while rivals, particularly in China, push ahead. He suggested the safety warnings might instead be a way for executives to signal how powerful their products are, or to redirect a portion of their massive infrastructure budgets toward safety and governance.

Despite the warnings of catastrophic risks, the commercial drive behind the technology shows no sign of slowing. Anthropic recently informed investors that it is on track to be profitable this quarter, marking its second consecutive quarter of positive adjusted operating income as it prepares for a US stock market listing later this year.