The world’s leading AI executives have urged a reduction in the pace of model development, but the companies responsible for building the infrastructure needed to support artificial intelligence show no signs of concern.
Over the past week, executives from major AI research labs—including Anthropic, OpenAI, and Google DeepMind—have publicly advocated for greater caution, citing security risks and the need for improved safeguards. Anthropic CEO Dario Amodei published an opinion piece arguing that AI progress must be slowed to address safety concerns, while Sam Altman of OpenAI suggested postponing the company’s planned 2026 initial public offering, describing it as “an ill-advised timing” given current safety risks. Elon Musk, whose SpaceX subsidiary is involved in AI development, also supported the call for restraint.
Financial markets reacted quickly to the warnings. AI-related stocks declined globally, with the Nasdaq hitting its lowest point in six weeks. Among the hardest-hit sectors were chipmakers, whose products are essential for training AI models. SoftBank’s shares in Tokyo dropped by more than 13%, as investors grew concerned about a potential slowdown disrupting industry growth.
Read Also: Goldman Sachs Acquires Davie Apartments for $154M
Data centers defy AI slowdown fears
However, within the data center industry, the sector responsible for building the facilities where AI models are trained, there is little evidence of alarm. Industry insiders dismiss the talk of a slowdown as more political posturing than a genuine shift in priorities.
Industry experts argue that frontier AI labs, such as OpenAI and Anthropic, are not the primary drivers of data center demand. The largest consumers of computing power remain the hyperscalers, companies like Microsoft, Amazon, Google, and Meta, which operate cloud services and enterprise workloads that sustain data center activity. These firms are expanding their facilities regardless of whether AI training accelerates or pauses.
Anthony Wanger, who leads the data center investment firm Regnaw Capital, explains that the current push for caution is less about halting AI development and more about establishing better governance frameworks. He notes that frontier labs are evolving into larger, publicly accountable organizations where oversight and risk management are becoming standard practice. “They’re talking about pacing model development. They’re not talking about not doing it,” Wanger said. “Anthropic and OpenAI are growing up. It’s really hard to be a large public company in America.”
Read Also: RFK Campus Parking Garage Plans Revealed
AI training remains just one demand driver
Even if AI training were to stop entirely, the broader data center market would continue to grow. Demand remains strong from AI inference workloads, the computing power required to run AI models in real time, and from traditional cloud services, which show no signs of slowing. Andrew Power, CEO of Digital Realty, the world’s second-largest data center operator, dismissed concerns that an AI slowdown would affect his business. “There’s tremendous digital transformation happening that is not connected to AI,” Power told CNBC. “Frankly, from my business lens, my seat, I think those demand trends, which are massive drivers of our business, have been stifled in these days of AI.”
AI training does contribute meaningfully to data center demand. John Andril, a marketing intelligence manager at Avison Young, estimates that between 10% and 20% of current and projected data center capacity is dedicated to training frontier models. In most industries, losing that level of demand would be devastating. However, the data center sector’s supply constraints make it resilient. The backlog of planned projects is extensive, with most facilities under construction today having been committed to months or even years ago based on existing demand. Tech giants like Microsoft and Amazon have repeatedly assured investors that their infrastructure spending is necessary to fulfill pre-existing commitments. Construction timelines of five to ten years mean much of the future capacity is already secured.
The push for AI restraint has so far come almost exclusively from U.S.-based companies. John Andril of Avison Young highlights the geopolitical context: Washington and Beijing are engaged in a competitive AI race, with each prioritizing advancement over risk mitigation. “Chinese tech companies are not ignorant to the dangers posed by AI, but as of right now, we don’t see tech CEOs over there making similar statements about a potential voluntary pullback, and there’s no reason to believe that Chinese companies or the Chinese government would enter or adhere to such an AI disarmament,” Andril said.