AI for Business

AI Hardware Stocks Slide as Anthropic Slowdown Talk Spooks Investors

AI infrastructure stocks took a hit Monday as investors reacted to calls for slowing AI model development, echoing the software sell-off earlier this year that became known as the "SaaSpocalypse."

Share:

AI infrastructure stocks took a hit Monday as investors reacted to calls for slowing AI model development, echoing the software sell-off earlier this year that became known as the "SaaSpocalypse."

Chipmakers and data-center suppliers - including gas turbine makers and electrical equipment vendors - tumbled on worries that demand could weaken if AI development is deliberately throttled. Anthropic CEO Dario Amodei called for a training slowdown over the weekend, citing safety concerns.

Investors would be wise to avoid jumping to conclusions. It's far too early to declare that the bearishness has officially shifted from software to infrastructure.

Earlier this year, concerns that AI would destroy enterprise software businesses proved overblown, though sellers still pushed those stocks down too far. Markets tend to overshoot in both directions.

Despite the alarming headlines, the fundamental case for AI hardware remains intact. Amodei called for a training slowdown, not an inferencing one. Demand for compute driven by existing models and new agentic applications continues to grow.

China's state-backed Global Times characterized Amodei's proposal as a "Cold War playbook" tactic aimed at slowing Chinese AI research so U.S. labs can pull ahead. Beijing is unlikely to participate. Amodei himself acknowledged China probably wouldn't join.

Open-weight models complicate any coordinated slowdown. Anyone can access, download and modify them, and capable models are being developed outside the control of any single regulatable entity. China has been a major hub of open-weight development.

That's why Monday's "AI apocalypse" trade likely won't have the same staying power as the software sell-off. Two other factors loom: the midterm elections in November and the possibility of an imminent Federal Reserve rate hike.

The bottom line: Monday is not a day to overreact. The world remains hungry for more compute, even based only on currently available model capabilities.

Source: CNBC

Ready to Modernize Your Business?

Get your AI automation roadmap in minutes, not months.

Analyze Your Workflows →