Gary Gensler warns an AI investment bust could cause a recession before 2028 election
The former SEC chair points to rapid spending and hard-to-see financing risks. Economist Ed Yardeni argues that investor skepticism could help prevent a crash.
Loading page…
The former SEC chair points to rapid spending and hard-to-see financing risks. Economist Ed Yardeni argues that investor skepticism could help prevent a crash.
Listen to this story
Former SEC chair Gary Gensler says opaque financing and fast-growing credit exposure could make an AI infrastructure downturn especially painful for debt-heavy neoclouds, even if major technology companies weather it. He sees the potential damage as closer to the dot-com bust than the 2008 mortgage crisis, but cautions that a burst bubble does not guarantee a recession and says he cannot predict the timing or severity. Ed Yardeni offers a less bearish view, pointing to chipmaker valuations that have cooled relative to the broader market.
CNN cited estimates putting annual AI infrastructure spending at about $800 billion.
Gensler said Big Tech was spending five times as much on AI in 2026 as it had three years earlier.
Gensler warned that circular financing and off-balance-sheet arrangements make risks difficult to assess, and that lenders may have extended too much credit too quickly.
An AI investment bust could push the economy into recession before the November 2028 presidential election, former Securities and Exchange Commission chair Gary Gensler warned in an interview published by CNN on November 16. Calling the boom “already a bubble,” he singled out debt-heavy providers of AI computing power as particularly vulnerable.
Gensler, who led the SEC from 2021 until 2025 and is now a professor at MIT, distinguished his concern about financial markets from his view of AI itself. He said he believes in the underlying technology but worries that investors are mispricing it. His warning concerns the capital committed to the buildout, rather than a rejection of the technology.
He expects a potential downturn to be comparable in magnitude to the early-2000s dot-com bust, but less harsh than the mortgage crisis that led to the 2008 financial meltdown. That is his forecast, not an established outcome. CNN also noted that a bursting investment bubble does not necessarily produce a recession, let alone a severe one.
The scale of investment is central to Gensler’s argument. CNN put annual AI infrastructure spending at about $800 billion by some estimates. Gensler said Big Tech companies were spending five times as much in 2026 as three years earlier. Companies have been pouring money into data centers, advanced chips and other equipment needed to support AI development.
He drew parallels with railroads, electricity and the internet: transformative technologies whose infrastructure attracted heavy investment ahead of anticipated revenue. In his account, the danger arrives when that investment cycle plateaus and then retreats. The historical comparison is about the financing of new infrastructure, not whether the inventions ultimately proved useful.
Gensler also questioned how clearly investors can see the risks. He pointed to circular financing and complex arrangements kept off company balance sheets, saying they make the amount of risk difficult to assess. His concern extends beyond the companies spending the money to the lenders supplying it: he expects history to show that credit flowed too freely.
I think history will show that they loaned too much money, too fast. And that’s the nature of these types of booms that become bursted bubbles.
Gary Gensler, speaking to CNN
The businesses he identified as most exposed were neoclouds—specialized cloud providers that rent large amounts of AI computing power to developers. Some have built their operations on substantial debt, leaving their business models vulnerable to sudden shifts. Gensler expects the major companies to survive a bust, while these providers could face business failures and bankruptcies.
Ed Yardeni, president of Yardeni Research, offered CNN a different reading. He described the AI trade as a “self-correcting bubble,” in which periodic declines release pressure without a full collapse. He called Gensler’s recession scenario plausible but argued that widespread concern about a bubble can itself restrain the excess that would make a crash more likely.
His evidence centered on valuations. Yardeni said chipmakers’ price-to-earnings ratios—a comparison of share prices with corporate profits—had fallen below the broader market’s. He argued that current valuations did not show the irrational exuberance of 1999–2000. For him, falling prices can be evidence of an adjustment already happening, rather than proof that a larger collapse must follow.
Gensler’s election-year forecast still came with a limit on precision. “I can’t say exactly when and how stormy,” he told CNN, while urging investors to prepare for rougher conditions.
Loading discussion...
Join the conversation
Explain whether skepticism can keep investment disciplined.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.