1. The IMF AI warning stands, but no verdict on a reversal
The International Monetary Fund said on November 12 that it is too early to determine whether investors are souring on artificial intelligence over the long term. Speaking at a regular press briefing in Washington, IMF spokesperson Julie Kozack said the Fund is watching volatility in AI stock valuations and a US political shift that could slow data center construction, while its earlier IMF AI warning on valuations still holds.
The message was deliberately measured. Kozack did not describe recent turbulence in AI-linked shares as the start of a lasting downturn, nor did she dismiss it. Instead, she restated the Fund's existing view that the valuations of companies tied to artificial intelligence are vulnerable to a reassessment of how profitable the technology will prove to be over the longer run, and said that position has not changed.
For investors looking for a signal from one of the world's most closely watched economic institutions, the takeaway was that the IMF is not declaring a bubble, a crash or a market top. It is saying the risks it identified earlier remain in place, and that recent swings have not yet settled the question of whether enthusiasm for AI is fading.
2. What did the IMF say about AI stocks?
The IMF said AI equity valuations remain exposed to a reassessment of the technology's longer-term profitability, and that it is too early to say whether a broader reassessment of AI investment is under way. Kozack said the overall view the Fund has taken "still remains the case," signalling continuity rather than a fresh alarm.
Her tone was cautious throughout. When pressed on whether recent developments point to a durable change in investor sentiment toward AI, she declined to draw that conclusion. "It's really too early to say whether there's going to be some broader reassessment of investment in AI technology," she said.
That framing matters because the IMF's statements can influence how governments, central banks and large investors interpret market moves. By neither amplifying nor downplaying the latest volatility, the Fund is effectively asking observers to wait for more evidence before treating short-term price swings as a turning point.
3. Two pressures the Fund is tracking
Kozack identified two developments the IMF is monitoring. The first is market volatility in AI-linked equities, where sharp price moves have tested investor confidence in the sector's growth story. The second is political: changes in the United States that could make large data center projects harder or slower to build.
Data centers are the physical foundation of the AI boom. Training and running advanced AI models requires vast computing capacity, which in turn depends on large facilities packed with specialised chips, along with reliable access to electricity, land and local permits. Any obstacle that slows the approval or construction of those facilities can ripple through the plans of the companies investing in them.
Asked specifically about pledges by newly elected Democratic governors to slow data center development, Kozack said the Fund was monitoring the situation. She did not offer an assessment of how much those pledges could affect AI investment, consistent with her broader point that the effects are not yet clear.
The political dimension adds a new variable to an investment cycle that has so far been driven largely by corporate spending decisions. If state-level policy makes it harder to secure power, land or permits, the pace at which companies can convert their AI ambitions into working infrastructure could change, even if their appetite for investment does not.
4. Concerns that predate the briefing
The IMF's caution is not new. In its October Global Financial Stability Report, the Fund flagged concentrated risks stemming from high valuations of AI-related companies and from rising borrowing by the largest technology firms. The November 12 comments built on that analysis rather than replacing it.
A central focus of the report was the hyperscalers, the large cloud computing providers that operate extensive data center networks and account for much of the spending on AI infrastructure. The IMF found that these companies made up about 10% of global investment-grade bond issuance, up from less than 1% in 2024.
That shift shows how much of the AI build-out is now being financed with debt rather than with cash generated by existing businesses alone. Investment-grade bonds are typically held by pension funds, insurers and other conservative investors, so the rapid rise in hyperscaler issuance means AI exposure has spread well beyond the stock market into portfolios that are often regarded as lower risk.
The IMF also examined what might happen if spending on AI slowed sharply. In its downside scenarios, global economic output could be up to 1.1% lower by 2028, and US output up to 1.8% lower. Those estimates are not forecasts of what the Fund expects, but they illustrate how much the broader economy now depends on the AI investment cycle continuing.
5. Why the Fund's view on AI carries weight
The International Monetary Fund is a 190-member-country institution whose core job is to monitor the health of the global economy and the stability of the international financial system. Its assessments feed into the decisions of finance ministries and central banks, and its regular reports are read closely by investors trying to gauge how policymakers see the risks ahead.
The Global Financial Stability Report, published twice a year, is the Fund's main tool for identifying vulnerabilities in markets and in the financial system. When it highlights a theme such as stretched AI valuations or rising corporate borrowing, that signals the issue has moved from a sector story to a question of broader economic stability.
The regular press briefings in Washington serve a different purpose. They give the Fund's spokesperson a chance to respond to current events between major reports, usually by restating or clarifying existing positions rather than unveiling new analysis. Kozack's November 12 remarks fit that pattern: they connected the latest swings in AI stocks and the debate over data centers to the framework the IMF had already laid out, without changing its conclusions.
6. How AI spending feeds into US inflation
The IMF also linked AI to the inflation picture in the United States. Kozack said the Fund views increased demand generated by AI investment and data center construction as one of two major contributors to US inflation.
The other is elevated energy prices, which have been pushed higher by the war involving Iran and the closure of the Strait of Hormuz, a narrow waterway through which a large share of the world's seaborne oil normally passes. Higher energy costs raise prices across the economy, from transport to manufacturing, and they also raise the cost of running power-hungry data centers.
The connection between the two pressures is direct. Data centers consume large amounts of electricity, so tight energy markets make them more expensive to operate while their construction adds to demand for materials, equipment and labour. As long as the conflict keeps oil and power markets constrained, the growth of data center capacity and the path of inflation are likely to remain intertwined.
Kozack added that it is too early to say whether shifts in AI investment will alter the US inflation outlook. That echoes her broader message: the Fund sees the channels through which AI affects the economy, but it is not yet prepared to say how recent changes will play out.
7. What the IMF stance means for investors
The IMF's position is best understood as a reminder about concentration risk rather than a signal to sell. The Fund is not predicting a collapse in AI stocks. It is pointing out that a great deal of market value and economic activity now rests on expectations about a single technology, and that those expectations could be revised.
For individual and institutional investors, that concentration can be easy to underestimate. A portfolio spread across many funds or sectors may appear diversified, yet still carry heavy exposure to the same AI theme if the largest holdings in major indexes are technology companies whose valuations depend on AI growth. A reassessment of AI profitability would therefore reach many investors who do not think of themselves as making a bet on the technology.
The same logic applies in credit markets. With hyperscalers rising from less than 1% to about 10% of global investment-grade issuance, bond investors now hold meaningful AI exposure as well. If the profitability of AI investments disappointed, the effects would not be confined to equity prices; they could also show up in how lenders price the debt of the companies building the infrastructure.
For crypto markets, which have tended to trade alongside technology stocks during periods of shifting risk appetite, the IMF's comments are a reminder that sentiment toward AI has become one of the broader forces shaping demand for risk assets.
8. What comes next
The IMF has made clear that its view will depend on how two things evolve: the behaviour of AI-linked equity markets and the political environment for data center construction in the United States. Statements from newly elected governors and the progress of major data center projects will offer early evidence on the second question.
On the economic side, the interaction between AI-driven demand, energy prices and inflation will remain in focus as long as the conflict involving Iran keeps energy markets tight. Until the picture becomes clearer, the Fund's position is that the risks it flagged in October still apply, and that recent volatility has not yet answered whether investors are stepping back from AI for good.