AI Lab Leaders Call for Frontier Slowdown, Triggering Global Semiconductor Sell-Off

Top executives at Anthropic, OpenAI, Google DeepMind, and xAI called for a coordinated slowdown in frontier AI development in mid-September 2026, citing concerns about autonomous agent capabilities and cybersecurity risks.

The announcement triggered a sell-off in global semiconductor stocks. Shares of Nvidia, Intel, and Samsung fell sharply as markets feared a slowdown in AI infrastructure spending.

Anthropic CEO Dario Amodei said the pace of frontier model development has reached a point where safety and alignment work cannot keep up. OpenAI CEO Sam Altman echoed those concerns, saying the industry may need to pause training runs for new models to ensure adequate safeguards are in place.

President Donald Trump rejected the calls, dismissing the cited risks as a "hoax" intended to stifle economic growth. He framed the AI race against China as a critical geopolitical advantage and said the United States had already implemented sufficient safeguards.

China's foreign ministry labeled the executives' calls "fearmongering" and "malicious competition" aimed at harming China's technological progress. Germany also rejected the idea of halting development, saying Europe must continue building its own AI capabilities to preserve digital sovereignty.

Critics suggested the push for a slowdown might be a strategic move by incumbent AI companies to establish competitive advantages through regulation, or to manage the massive capital expenditures associated with AI compute.

Despite the controversy, infrastructure spending remains high. Companies continue to invest heavily in data centers, optical infrastructure, and specialized hardware. South Korea reported record ICT exports of $59.98 billion in August, signaling continued strong demand for AI hardware.