AI spending boom could trigger sharper downturn than dotcom crash: Professor Aswath Damodaran

The artificial intelligence boom is being fuelled by an unprecedented surge in capital expenditure that could make any eventual downturn far more painful than the dotcom crash, according to valuation expert Aswath Damodaran.

“This has been the biggest infrastructure run-up I’ve ever seen… the amount of money that’s being put into AI capex is immense,” he said, warning that “when the correction comes, the pain will be more intense.”

A key risk, he added, lies in the nature of funding. “The problem with the AI capex boom is not only is it immense, but a big chunk of it is funded with debt,” he said, warning that any downturn could trigger defaults with wider economic consequences. “That pain doesn’t stay restricted. It spills over into the rest of society.”


Unlike the late-1990s dotcom bubble, which was largely equity-funded and involved minimal traditional capital expenditure, the current AI cycle is marked by heavy investment in infrastructure. “The dotcom boom and bust… there was very little traditional capex or even R&D driving it,” he noted. In contrast, today’s AI expansion involves vast spending on data centres and computing capacity.

Damodaran also pointed to a structural shift among large technology firms.

Companies once considered asset-light are becoming capital-intensive due to AI infrastructure demands. “They’re changing their entire character… I’ve got to value them differently than I did,” he said, adding that firms may be venturing into unfamiliar territory with heavy, long-term investments.

He contrasted this with a more cautious approach.

“We undervalue restraint in business… this might be one of those markets where restraint is a good feature,” he said, suggesting that avoiding early overinvestment could prove advantageous.

This raises the risk of wider economic fallout. “There’s a very real chance that if there’s a correction… that pain doesn’t stay restricted. It spills over into the rest of society,” he said, drawing parallels with 2008, when excessive lending magnified the crisis.

He also talked about “big market delusion”, where vast market projections combine with investor overconfidence to drive excessive risk-taking.

“Put overconfidence and a big market together, you can see the big market delusion play out,” Damodaran said, pointing to a cycle of overreach followed by inevitable correction. “There will be an overreach. There will be a correction… but there’ll be a few winners that come out of this space.”

He emphasised that inflated expectations around AI’s total addressable market, often pegged in the tens of trillions, are encouraging indiscriminate investing. “You are in a sense investing and hoping the big market delusion is not a delusion,” he said, cautioning against broad-based bets such as AI-focused exchange traded funds.

He also pointed to a fundamental shift in large technology companies, which are becoming more capital-intensive due to AI investments. “They’re changing their entire character,” he said, comparing them to businesses moving from asset-light models to infrastructure-heavy operations. Investors, he added, must now focus more on capital expenditure and depreciation rather than just margins.

On semiconductor firms such as Nvidia and Micron, Damodaran cautioned that their fortunes hinge on the eventual size and profitability of the AI market. If expectations of a multi-trillion-dollar market fail to materialise, “there’s a heck of a lot of cleaning up to do.”

Ultimately, Damodaran argued that the scale of AI’s promise carries broader societal implications. “If AI replaces people… the stories we’re telling about 10, 15, 20 trillion markets are actually terrifying,” he said, warning that large-scale job displacement could follow if such projections materialise.