Citadel founder Ken Griffin has warned that artificial intelligence is “almost certainly” in a bubble, as investors question whether the technology's substantial capital requirements will deliver sufficient returns, according to a report by Bloomberg.

Speaking at the Future of Global Markets Conference in New York on Tuesday, Griffin said companies have struggled to translate investment in AI into measurable productivity gains, raising concerns about the sustainability of current valuations and spending commitments.

“Corporate America has been very uneven in its ability to extract value from its AI models,” Griffin said, adding that businesses are increasingly scrutinising their spending on services from OpenAI and Anthropic.

The comments come amid growing investor concerns over the scale of AI infrastructure investment and the time required for spending on chips, data centres and computing capacity to generate profits.

Anthropic is reportedly preparing for a potential initial public offering, with the Wall Street Journal reporting that the company could spend $750bn on additional computing capacity in the coming years. Griffin said its long-term success would depend on demonstrating that its technology is essential across industries.

Despite his warning, Griffin remained optimistic about AI's economic potential, arguing that technological transformations have historically coincided with speculative bubbles.

“Any radical transformation of our economy is generally intertwined with a bubble,” he said.

Griffin compared the current AI boom with the dot-com era, noting that while speculative excess preceded the market crash, companies that emerged from the downturn went on to dominate their industries for decades.

He also predicted that AI would accelerate the formation of new businesses while disrupting established companies and white-collar employment, describing the coming period as a potential “golden era for entrepreneurs”.