For the past two years, the International Monetary Fund and nearly everyone else described the artificial-intelligence boom as a force for good in the world economy: a surge of investment that could lift productivity, offset the drag from high energy prices and, in the right hands, pull growth upward. On Wednesday, the fund’s leader quietly reframed it as something closer to a hazard.
In a speech in Singapore, IMF Managing Director Kristalina Georgieva grouped the unbalanced AI boom with energy shocks and record public debt as the risks that now threaten the global economy, according to Reuters. She was previewing the IMF and World Bank Annual Meetings, which open next week in Bangkok, and her message was that the three forces now move together, each amplifying the others.
The shift in tone matters because it changes what the AI buildout is assumed to be. Where data-center investment was once treated as a growth engine that could help countries weather a shortage of cheap energy, Georgieva cast it as a demand shock that is itself fuelling inflation. The world, she said, is being pulled in two directions at once: a negative energy supply shock from the Middle East conflicts, and a positive demand shock from artificial intelligence that is also pushing prices higher.
The boom, she added, is bypassing most of the world. The gains from AI investment remain concentrated in a small number of countries, while the inflationary and fiscal side effects spread more broadly. That unevenness, in her telling, is part of what makes the boom a risk rather than a simple blessing. A technology that lifts a few economies while heating up the costs paid by everyone else is not the neutral productivity story the fund once sketched.
Debt is the second leg of the risk combination. Georgieva said global public debt is at its highest level since the Second World War and is on course to exceed 100% of global GDP before 2030. She singled out advanced economies, led by the United States, as the worst offenders on debt loads. The Institute of International Finance puts total global debt above $365 trillion. In that context, higher borrowing costs are not an abstract worry; they are the mechanism by which every other problem gets worse.
The third leg is energy. Conflict in the Middle East has kept oil and gas markets under pressure, and Georgieva warned that price pressures could intensify as the northern winter raises demand and countries rebuild depleted reserves. Data centers are now part of that equation. The electricity demands of AI infrastructure add another layer to global energy demand, linking the boom back to the energy shock the fund once thought it might cushion.
The bond market is already registering the strain. Georgieva noted that 10-year sovereign yields in the United States, Germany and Japan now sit at their highest levels since 2007, 2009 and 1996 respectively, and are still climbing. That is the clearest sign that investors are repricing the combination of inflation and debt, and it is the channel through which the fund’s warnings reach the real economy.
For the fund, the concern is not the technology itself but the way it is being financed and the speed at which expectations have run ahead of results. A boom built partly on debt, layered on top of a debt load already at record levels, leaves governments less room to absorb a shock, whether it arrives through energy prices, bond markets or a repricing of the AI trade itself.
The fund’s prescription was cautious. Georgieva urged central banks to keep a prudently hawkish bias, signaling that they should be slow to ease while inflation risks persist. She said policymakers can no longer rely on higher growth alone to fix fiscal problems, and that governments should build protective buffers now, while markets are still calm.
Behind the speech is a larger recalibration. For much of the AI era, the official line at institutions like the IMF has been that the technology’s benefits would arrive faster than its costs. Wednesday’s remarks suggest that assumption is being walked back. If the boom is itself a source of instability, then the correct policy response is not simply to encourage more of it, but to manage its side effects, from the electricity it consumes to the debt used to finance it to the concentration of its rewards.
Georgieva did not abandon the optimistic case. She said IMF research suggests AI, done right, could add half a percentage point a year to world growth, and that regulation and preparation will decide who captures that upside. But the framing has changed, and framing is what these meetings run on. AI has moved from the solution column to the risk column, alongside the oil shock and the mountain of debt, and that reclassification, once made, is not easily undone.
