Tech stocks tumbled across Asia and US futures pointed lower on Monday after some of the world’s most powerful AI bosses called for the industry to slow development over safety fears.
Japan’s SoftBank, one of OpenAI’s biggest investors, plunged as much as 13 per cent in Tokyo, while memory chip maker Kioxia dropped as much as 9.8 per cent.
South Korea’s SK Hynix fell more than five per cent in early trading and Samsung Electronics dropped nearly four per cent, helping drag the benchmark Kospi lower. Taiwan’s TSMC fell 1.2 per cent.
The sell-off followed a weekend intervention from Anthropic chief executive Dario Amodei, who warned that AI companies needed to “slow the pace” at which they improve their most powerful models.
OpenAI boss Sam Altman and Elon Musk both backed the call, marking a rare point of agreement between three of the industry’s fiercest rivals.
US markets looked set to follow Asia lower. Nasdaq 100 futures were down around 1.3 per cent on Monday morning, while S&P 500 futures fell 0.6 per cent and Dow futures edged 0.1 per cent lower.
“Suddenly the headlong rush to develop AI seems to have stopped in its tracks,” said Chris Beauchamp, chief market analyst at investing and trading platform IG.
“For markets this raises the possibility of a slowdown in data centre rollout, or a throttling of demand for chips, both of which would undermine the foundations of the investment thesis in many big-name stocks.”
AI boom hit by safety concerns
AI stocks have powered global markets higher as tech companies pour hundreds of billions of dollars into data centres and increasingly powerful models.
But investors have become more sensitive to any suggestion that the spending boom could slow, particularly as valuations have climbed and companies dramatically expand production capacity.
Samsung and SK Hynix are planning more than $500bn of investment in a new South Korean chipmaking hub, while Kioxia and Sandisk are investing more than $31bn to expand memory chip supply.
“AI valuations assume not only strong demand but also a relentless pace of model development,” Charu Chanana, chief investment strategist at Saxo Markets, said.
She warned memory chips could face an oversupply risk if new capacity arrives before demand.
The latest concerns were triggered by Amodei, who said over the weekend that Anthropic would give independent evaluators greater access to its safety work and urged rival developers to coordinate on slowing the advance of frontier models.
Altman said OpenAI would adopt a similar approach, writing that pacing AI development had become a “primary topic” of discussion at the ChatGPT maker in recent weeks.
He separately ruled out an OpenAI IPO in 2026, telling Fortune that going public while the industry grappled with safety concerns would be “ill-advised”.
The fallout follows former OpenAI and Anthropic researcher Jacob Coxon quitting Anthropic last week after accusing the two companies of “racing straight to self-improving superintelligence and gambling with our lives”.
The UK parliament’s Joint Committee on Human Rights added to the pressure on Monday, warning AI posed “novel and serious human rights risks”, and that governments risked being left behind by the speed of development.
Markets were already nervous ahead of this week’s Federal Reserve decision, while surging oil prices and elevated bond yields have added pressure to technology valuations.
Analysts cautioned that Monday’s sell-off may prove short-lived. Greater use of existing AI models could continue to drive demand for chips and data centres even if companies take longer between major model releases.
“This seems very much a knee-jerk reaction,” Beauchamp said. “AI use is still expected to explode, driving data centre construction even if the wilder experiments are rolled back.”