Chip Stocks Drop on AI Pacing Proposal

Global chipmakers and memory producers saw share prices slide on Monday following a public proposal by Anthropic CEO Dario Amodei to slow the pace of frontier AI model development. The call, which echoed recent resignations of safety staff at major AI labs, spooked investors who had been betting on a continued acceleration in artificial intelligence infrastructure spending. The sentiment shift immediately impacted the semiconductor sector, which had been attempting to recover from a 19% drawdown from its June peak.

The market reaction was sharpest among memory giants whose recent margin expansion relies heavily on tight supply of high-bandwidth memory. In Asia, SK Hynix and Samsung Electronics fell 6.4% and 4.1% respectively, while Kioxia dropped 6.4% in Japan. European equipment makers and chip designers also opened lower, with ASML down 4.4%, Infineon and ASM International falling over 5%, and STMicroelectronics shedding 3.5%. Analysts noted that these companies are more exposed to demand fluctuations than peers with diversified revenue streams or long-term equipment backlogs.

Memory Makers Face Higher Exposure

Unlike fabless IP providers or equipment makers with multi-year order books, the earnings of SK Hynix, Samsung, and Micron are directly tied to the pricing power of HBM. This makes their financial performance highly sensitive to any pullback in AI capital expenditure. However, current market dynamics suggest that immediate revenue shocks may be mitigated by existing multi-quarter supply agreements and strong inference demand. Analysts argue that unless demand weakens significantly enough to drive up inventories and soften pricing, the margin hit could be delayed.

Industry Spending Remains Robust

Despite the rhetoric around slowing development, actual spending commitments from AI leaders show no signs of deceleration. Reports indicate that Anthropic has agreed to 14.8 gigawatts of computing capacity involving approximately $517 billion in commitments over the next 11 months. The company also expects a second consecutive quarter of adjusted operating profit, with inference gross margins exceeding 80%. This suggests that the current proposal aims to modulate the speed of progress rather than halt investment, a distinction that analysts believe will keep compute demand high.

Strategic Implications for Semiconductor Supply

The proposed pacing also includes continued restrictions on selling advanced chips to Chinese firms and crackdowns on model distillation, linking AI safety to geopolitical trade policies. Bernstein analysts maintain a positive outlook on the sector, noting that current compute capacity is already insufficient for existing model demands. They identify Nvidia, Broadcom, and semiconductor capital equipment manufacturers as preferred plays, arguing that the shift from extreme speed to moderate speed still represents significant growth for the supply chain.