AI Boom-or-Bust Scenarios Loom Large at World Financial Summit

Financial leaders gathering in Asia this week had an urgent message for the world: The global economy is betting the house on artificial intelligence.

That bet has paid off handsomely so far. Surging investment in AI and the infrastructure to run it is powering growth in the U.S. Exports are booming in Asian economies that make computer chips, servers and other equipment needed for the data center build-out. Countries riding the AI wave have weathered the energy crunch caused by the Iran war better than those stuck on the sidelines.

But the scale of the boom—and its expanding reach into more economies—is seeding new worries about the global consequences if the AI mania ends.

It probably won’t just be stock investors that get burned if boom turns to bust, the International Monetary Fund said. The world could be looking at a sharp pullback in growth that hits both rich and poor countries alike, according to one scenario included in the fund’s fall projections.

The scenario wasn’t a forecast. The fund also ran the numbers on other, less-pessimistic scenarios, including one in which AI gives the global economy a major shot in the arm, without signaling which outcome it thinks is more likely.

The economic possibilities—and potential perils—of the new technology dominated discussions among finance ministers and central bankers gathered for the annual meetings of the IMF and World Bank in Bangkok.

The gathering came during a week when investor nerves over rising borrowing costs and AI labs’ profitability hit semiconductor stocks and the bonds of AI infrastructure providers such as CoreWeave. Two Australian data-center companies put the brakes on plans to issue shares to the public, citing market volatility.

Even if a crash can be avoided or managed, powerful AI-assisted cyberattacks on banks and other financial institutions represent a new threat to the stability of the global financial system, the IMF warned. Earlier this month, hackers used a Chinese AI program to attack South Korea’s biggest banks and steal the personal information of 68,000 people.

AI will be transformative, but “it could be a disruptive transition,” Bank of Canada Gov. Tiff Macklem said in Bangkok on Wednesday.

The buzz among the world’s financial-policy elite underscores how AI and the companies behind it have become deeply entwined with almost the entire global economy.

AI is sucking in capital, workers, equipment and raw materials from around the world, yoking the fortunes of major economies to a technology that is becoming more enmeshed in workplaces and daily life.

In the U.S., the money spent on data centers and related infrastructure is expected to top more than $10 trillion in the seven years through 2032, averaging around 3.6% of annual gross domestic product a year, according to estimates by economist Stijn van Nieuwerburgh. A boom on that scale would eclipse historical parallels such as the construction of the railroads, the electricity grid or the internet.

“I’d say the advent of this new technology is probably the most consequential change to the U.S. and the global economy in my lifetime at this point,” Federal Reserve Chairman Kevin Warsh said Friday.

The immediate future may be bumpy, Warsh said. But in the longer run, the technology would be a positive supply shock, he said, raising the economy’s capacity to produce more goods and services without necessarily boosting inflation.

Asian economies are in the midst of an epic export boom, which is masking weakness in consumption and other engines of growth. South Korea’s exports rose more than 80% year-over-year in September, figures from data provider CEIC show. Taiwan’s exports were up more than 60%. China’s exports rose by more than a third. Japanese exports are up by a fifth.

The technology’s supporters say AI will revolutionize industries and drive a productivity boom that will power growth for years. Some say financial unease over the growing debt pile financing the AI boom is misplaced, thanks to the even bigger cash piles at major firms such as Google and Microsoft.

“The big hyperscalers entered this economy quite cash-rich,” said Atsi Sheth, chief credit officer of Moody’s Ratings.

In the optimistic scenario sketched out by the IMF, productivity gains from widespread AI adoption fuel a goldilocks period of robust economic growth and contained inflation.

Another scenario assumes AI lives up to its promises but the build-out runs into supply snafus as chip makers and others struggle to keep up with runaway demand. That pushes up inflation and interest rates in the near term, pinching growth, but ultimately the productivity gains from AI come through.

In the most downbeat scenario, investors sour on AI after it fails to live up to its lofty expectations or AI firms’ profits turn out much weaker than anticipated. Investment falls in the U.S. and Asian exports suffer. Stocks plunge and the dollar weakens.

In the fund’s modeling, the bust could knock as much as 1.3 percentage points off global growth in 2028. The fund said confidence in the technology could also evaporate if sophisticated AI programs run amok or are used in devastating cyberattacks.

IMF Managing Director Kristalina Georgieva said Friday that some level of upset seems inevitable, even if the gains from AI ultimately pan out. People overestimate the benefits of new technology in the short term, while underestimating the gains over the long term, she said.

“We have very high expectations. They are not going to be met because it takes time for diffusion,” she said, referring to the process by which new technologies spread through an economy.

“The risk is: We’re excited, it doesn’t lift productivity immediately. We’re disappointed, we pull out. And then it becomes our problem.”