IMF says it’s too early to call a long-term reversal in AI stocks

The International Monetary Fund is keeping a close eye on the recent turbulence in AI stocks. It just isn’t ready to call it a breakup.

At a regular press briefing in Washington on November 12, IMF spokesperson Julie Kozack said it remains too early to tell whether investors are cooling on artificial intelligence for the long haul. The fund is also tracking a US political shift that could slow the construction of data centers, the physical backbone of the AI boom.

What the IMF actually said

Kozack’s message was cautious. AI equity valuations, she said, are still vulnerable to a rethink of how profitable the technology will be over the longer term.

“That overall view that we have taken still remains the case,” Kozack said.

The fund is watching two separate pressures. One is market volatility in AI-linked shares. The other is political: changes in the US that may make it harder to build the data centers AI companies need to train and run their models.

On that second front, Democratic governors have reportedly committed to taking a fresh look at data center projects. For an industry whose growth plans depend on huge amounts of power, land and permits, a slower approval process is not a minor detail.

The numbers behind the IMF’s caution

The IMF’s worry predates this week’s briefing. Its October Global Financial Stability Report flagged concentrated risks tied to lofty AI valuations and rising debt at the largest tech companies.

The report focused on hyperscalers, the giant cloud providers that operate massive data center networks. According to the IMF, those firms accounted for 10% of global investment-grade bond issuance. In 2024, that share was less than 1%.

The IMF also modeled what happens if the spending spree slows. Under its downside scenarios, a pullback in AI investment could cut global GDP by as much as 1.1% by 2028. US economic output could fall by as much as 1.8%.

Stock market concentration adds another layer. The so-called Magnificent Seven tech stocks now make up more than 35% of the S&P 500, compared with 12% a decade ago. AI-related shares have driven a large share of recent earnings growth.

Why the inflation angle matters too

AI investment and data center construction are seen as major contributors to US inflation. They’re working alongside energy prices that have been pushed higher by the Iran war and the closure of the Strait of Hormuz.

What this means for investors and markets

The IMF’s stance is notable mostly for what it doesn’t say. It isn’t declaring a bubble, and it isn’t calling a top. It’s saying the risks it flagged in October are still in place, and recent volatility hasn’t settled the question either way.

For investors, that reads less like a sell signal and more like a reminder about concentration. With the Magnificent Seven making up over a third of the S&P 500, plenty of portfolios that look diversified are heavily exposed to the same AI theme.

The credit market is the less obvious pressure point. Hyperscalers going from under 1% to 10% of global investment-grade issuance means bond investors now carry meaningful AI exposure too.

Energy is the other variable. As long as the Iran conflict keeps oil and power markets tight, data center growth and inflation will stay linked.