Brookfield CEO Flatt: AI competition “is already slowing down”, infrastructure and computing power cannot keep up with demand

The Zhitong Finance App learned that Bruce Flatt, CEO of Brookfield Corp., said that the AI race is already slowing down because developers are unable to build sufficient infrastructure to keep up with the pace required by AI companies.

Leaders of major AI companies, including Anthropic PBC CEO Dario Amodei and OpenAI CEO Sam Ultrman, recently stated that AI laboratories must slow down technological progress to deal with unpredictable risks and prevent advanced AI models from escaping human control.

Flatt pointed out that even before that, the industry had hit a bottleneck: there was a hard limit on how fast data centers and other infrastructure could be implemented.

“As an industry, we haven't built enough. We couldn't even build a fraction of what people thought we needed.” Flatt said at Brookfield's annual Investor Day on Thursday, “So when they say they plan to slow down, it's actually going to slow down no matter what because there isn't enough computing power to deliver the scale everyone thinks they need.”

Brookfield estimates that AI infrastructure will require more than $7 trillion in capital investment over the next decade. Flatt said the slowdown is a good thing because it will “bring more discipline to the entire system.”

Flatt's judgment is supported by solid data. According to company guidelines compiled by Statista, the total capital expenditure of the four tech giants Meta, Microsoft, Alphabet, and Amazon is expected to reach about 760 billion US dollars in 2026, an increase of nearly four times that of about 200 billion US dollars in 2022. Among them, Amazon plans to spend about 220 billion US dollars, Alphabet up to 205 billion US dollars, Microsoft about 190 billion US dollars, and Meta about 145 billion US dollars.

However, these huge investments are facing hard restrictions on infrastructure construction. What Flatt said about “not being able to build” is not a lack of capital, but rather a physical supply bottleneck.

The shortage of electricity is currently the most prominent bottleneck. The Uptime Institute clearly stated in its 2026 data center forecast report that electricity will become a “defining constraint” for data center growth in 2026 and beyond, AI-driven load growth will increase the already tight grid pressure, and the long cycle of large-scale power deployment will constitute a substantial constraint.