IMF warns AI market tumble poses risk to world financial stability

The International Monetary Fund has warned that a sharp fall in artificial intelligence share prices could spread across world markets and disrupt global financial stability.

In its latest Global Financial Stability Report, released at meetings in Bangkok, the fund grouped an AI market correction alongside climbing government bond yields and cyberattacks as the three immediate triggers that could tighten credit conditions and shake economies worldwide.

High valuations and rising debt

The report points to stretched share valuations and heavy corporate borrowing as key pressure points. The largest tech companies, known as hyperscalers, have recorded falling free cash flow because of the cost of building computer centres and buying specialised processors. To fund that construction, businesses are turning more heavily to debt.

The fund also flagged circular financing arrangements, where hardware suppliers, cloud providers and software developers invest in each other or commit to buy one another's services. Athanasios Vamvakidis, deputy director of the monetary and capital markets department at the fund, said international exposure to American technology investments is at an all-time high. If corporate earnings fail to match expectations, that exposure means a drop in market value would reach well beyond the United States.

Under a separate downside projection in the fund's economic outlook, a sudden retreat in AI spending could reduce global output by 1.1 per cent by 2028.

What this means for Australian businesses

Australian small and medium firms do not need to follow daily stock market swings, but the financing pressures behind big tech do filter down to everyday software buyers.

• Software price rises are likely. Tech providers under pressure to show real cash flow will focus harder on turning users into paying accounts, raising subscription fees and reducing free tiers.
• Avoid building critical workflows around a single startup. Early-stage firms that rely on continuous venture capital or credits from bigger partners may alter terms or close quickly if funding dries up.
• Keep track of running costs. Review automated tasks and cloud usage monthly to ensure your business only pays for computing capacity that saves actual staff hours or generates revenue.
• Stay model-neutral where possible. Structure your internal tools and data storage so you can switch between different AI models without having to rebuild entire operational workflows.