AI Slowdown Calls Send Asian Chip and AI Shares Lower

Calls from leading AI figures to slow development on safety grounds landed in Asian markets as a challenge to the growth assumptions supporting the AI build-out. South Korea’s KOSPI fell 3.7% at the start of trading, while SK Hynix, TSMC and SoftBank shares also declined.

Anthropic chief executive Dario Amodei called on the industry to slow down, arguing that building AI too fast is reckless. OpenAI chief executive Sam Altman and Elon Musk backed the call, turning an argument associated with one frontier lab into a more consequential signal for investors assessing the pace of the sector’s expansion.

Safety warnings meet spending commitments

The warnings are not solely abstract. Amodei warned that a future swarm of AI agents could cause hundreds of billions of dollars in damage by taking over the internet. Anthropic has separately reported that some models can perform intelligence and weapons-related tasks once handled by scarce, highly trained experts, and says those findings support safeguards against misuse.

For investors, the immediate issue is less whether a shared slowdown has been agreed than what a material slowdown would do to projected demand. Ipek Ozkardeskaya, a senior analyst at Swissquote, said leases, debt and power commitments remain even if expected computing demand and revenue growth weaken. That could push credit risk toward heavily leveraged data-center operators and the lenders exposed to their projects.

OpenAI’s competing commitments

OpenAI’s position contains the same tension. In a September publication, the company said it may slow or stop development or deployment if it identifies an unacceptable safety risk. Yet that post also described automated AI research as a potential accelerator of research progress. The contrast leaves companies and markets facing the same unresolved test: how quickly safety measures can keep pace with more capable systems.

The market move does not settle whether the safety warnings justify a slower pace, and Amodei’s claims have been disputed by some AI experts. It does show that safety-driven restraint is no longer just a governance question. Investors are treating the prospect of slower progress as a possible constraint on the business model financing the industry’s physical expansion.