Wall Street gauges impact of a possible AI development slowdown on data-center sector
Investors are assessing how a deceleration in AI model training could affect companies tied to U.S. data-center expansion, after Anthropic’s chief executive suggested a slower pace for frontier models.
Wall Street analysts are weighing the consequences of a potential slowdown in artificial-intelligence model development for firms linked to the U.S. data-center boom. The concern grew after Anthropic CEO Dario Amodei proposed moderating the pace of frontier-model training, a move that prompted a sell-off in AI-infrastructure stocks on Monday. GE Vernova fell nearly 9%, Caterpillar slipped more than 4%, Vertiv dropped close to 8% and Oracle declined almost 4%.
Equity analyst Rishi Jaluria of RBC Capital Markets said a slowdown in model training would likely pose a headwind for Oracle’s cloud-infrastructure business, which has been a key driver of its recent performance. Other companies that could feel the impact include server manufacturers Dell and Hewlett Packard Enterprise, as well as energy-equipment providers such as GE Vernova, Caterpillar and Vertiv, which have invested heavily in power solutions for AI servers.
Investors are also monitoring the debt market, where hyperscalers such as Amazon and Alphabet have recently raised billions of dollars. Sources familiar with private discussions said any new debt issuances this fall are expected to carry higher interest rates as fixed-income investors seek greater compensation for perceived risk.
"Fixed income investors are demanding more reward," one source said.
The broader data-center sector remains dependent on sustained demand for AI chips and compute capacity, and a significant delay in model development could affect revenue streams for a range of hardware, software and infrastructure providers.