IMF's Georgieva warns debt-financed AI boom could threaten financial stability

Georgieva's 3 November warning

Kristalina Georgieva delivered the warning at the International Banking Conference 2026, organised by Banco Santander. She shared the stage with Ana Botín, Santander's chief executive president.

She said 60% of jobs in advanced economies will be altered by AI within three to five years. AI will affect 40% of jobs worldwide, she said. She called it "a slow tsunami, but a tsunami nonetheless". The net impact on employment has been slightly positive so far, she added. But the structure of employment is changing. Routine and entry-level roles are already shrinking in the United States. Demand for low-skilled service work is rising.

She named three sources of risk. Financial stability came first. If the investment fervour around AI falters, "there will be serious problems", she said. She also flagged the growing share of AI projects financed with debt. That debt is often long-dated and has a high component of circular investment. "I invest in you, you invest in me, we are all very happy, except that if something goes wrong, that would be a big problem," she said.

Georgieva said the fruits of the AI boom are concentrated in a small group of producer countries. She named the United States, China and India. Europe is absent apart from the Netherlands, she said.

Botín agreed that AI is an opportunity and "a unique event in the history of humanity". She said Europe needs global banks and companies. She also said AI rules should follow "same risk, same rules", including when new agents access accounts.

The debt behind the buildout

At the Hong Kong Monetary Authority's Global Financial Leaders' Investment Summit, Pimco chief executive Emmanuel Roman put a number on the financing question. He said AI-related capital spending could reach around $10 trillion over the next five years. He cited an estimate by General Atlantic chief executive Bill Ford.

Roman said about $550 billion in AI debt had been issued in various forms so far this year. He set that against a roughly $20 trillion global corporate bond market. "At some point in time, it's going to be too much," he said.

Goldman Sachs wrote in a report published on Monday that global data center demand will require about $8.5 trillion in capital spending over the next five years.

Blackstone president Jon Gray sat on the same panel. He said the scale of capital required for AI is creating concentration risks for debt investors. Those risks run across the sector and at individual companies.

Lenders have limited upside, Gray argued. Unlike shareholders, they collect interest and principal rather than growth. He said financing the buildout would need public markets, banks and private capital. Companies are turning to structured transactions to limit the debt they take on. Cash generated by hyperscalers would also fund part of the spending, he said. "The creativity is needed because the scale of capital is enormous," he said.

He stayed bullish on demand. Computing power demand was "running well ahead of supply", he said. Spending on AI models across Blackstone's portfolio companies had risen 20-fold over the past year.

Concentration and a listing test

Verena Ross, who was set to step down as chair of the European Securities and Markets Authority at the end of October, told Politico that a "significant concentration of market value" in a few companies "raises some questions around how to ensure proper diversification of investor portfolios".

Ross spoke to Politico the day after Anthropic filed paperwork to take the company public. That listing could be the biggest in history. It is also a test of investor confidence in the AI boom. She said finance regulators worldwide are watching "the gap between the valuation issue and the macro picture".

Together, the remarks describe a buildout that depends on debt and concentrated equity valuations. Georgieva said a faltering AI investment cycle would cause "serious problems"; Roman said the scale of financing could become "too much"; Ross questioned portfolio diversification.