The International Monetary Fund named an AI-driven market correction its biggest near-term risk on October 13, the same day Asian chip stocks sold off.

Its October Global Financial Stability Report lists three threats that could tighten financial conditions: a correction in AI-related asset prices, a sharp rise in sovereign bond yields, and major cyberattacks. The first of those had already moved markets by lunchtime in Seoul.

Three risks, one of them built on borrowed money

The report, titled "Resilience, Risk, and Transformation," describes markets that have held up on solid growth and strong corporate earnings while the vulnerabilities underneath have grown.[1] The IMF does not predict a crash. It calls the catalysts "plausible" and says they are worth monitoring.[3]

What concerns the fund is how the AI boom is financed. Hyperscalers now account for 10% of global investment-grade bond issuance, up from less than 1% in 2024, according to Reuters.[3] A review of 14 deals since 2023 found about $300 billion in equity investments and $650 billion in debt, committed in exchange for $1.4 trillion in future revenue.[2]

When AI companies invest in one another, debt and equity move in circles. Stress at one link can lower revenue forecasts, which in turn weakens the valuations the debt was built on. The IMF describes that structure as a channel that could turn a price correction into forced selling and higher volatility.[1][2]

"An AI market correction will affect the real economy," Athanasios Vamvakidis, deputy director in the IMF's Monetary and Capital Markets Department, said at a briefing in Bangkok.[3]

He added that today's calm, propped up by earnings and optimism about AI, means that "resilience is not necessarily safety."[3]

Broadcom's CEO called it a casino

On October 12, Broadcom chief executive Hock Tan spoke on a panel at the DBS Asia Future Forward Conference in Singapore. His assessment of enterprise AI spending was blunt. "Most companies out there cannot find any return on investment," he said, adding that "people are just wasting money, spending tokens, trying to understand" the technology.[5]

He likened it to a child at a video arcade. Ten dollars of tokens disappear in ten minutes and nothing is learned.[5]

Tan then turned to the investors. "A lot of investors have no clue what they're investing in, but set very extreme expectations," he said. "It's almost like they are in a casino, gambling."[4][5]

He was not declaring the boom over. Broadcom supplies much of the hardware inside data centers, and Tan said he sees no slowdown in the sector. His worry is the next constraint. Frontier labs will soon compete for the same capital as the infrastructure builders, and he judged that funding "might be the biggest constraint" on growth.[5]

On OpenAI, which missed revenue guidance earlier this year, Tan said $50 billion in annualised revenue by the end of September points to a workable business but says little about profit, because spending on training still exceeds it.[4][5]

The numbers behind the selloff

Asian markets opened lower and kept falling. South Korea's Kospi closed down 6.25%, with Samsung Electronics off 7.4% and the memory maker SK Hynix down 10%.[4]

In Japan, the Nikkei fell 1.8%. Tokyo Electron, which builds chipmaking equipment, dropped 7.15%, and SoftBank Group, an OpenAI investor, lost 8.5%.[4] Taiwan's exchange shed almost 2%, and TSMC fell by a similar amount. Micron slipped 3.9% in the United States after the Nasdaq lost 0.4% on Monday.[4]

Tan's remarks were not the only trigger. The South China Morning Post reported that China's CXMT had made progress on a DRAM architecture that raises memory density, raising the prospect of fresh competition for Samsung, SK Hynix and Micron.[4]

"CXMT's new developments remind investors of the huge competition in the sector," said Chris Beauchamp, chief market analyst at the trading platform IG. He called the improvement good for customers and dangerous for technology stocks "on whom the entire stock market rally seems to rest."[4]

Where the stress would show up

The IMF ran an adverse scenario in which AI investment slows and bond markets strain. Banks holding 22% of global banking assets would end up below regulatory capital limits, up from 18% in last year's test.[2][3]

Outside banking, the exposure has shifted. Hedge funds now hold about $13 trillion in assets, more than three times their 2013 level, and their leverage can force synchronized selling during stress.[1] Pension funds and insurers have raised allocations to illiquid assets and derivatives, while private credit sits in the same chain.[2]

Cyber risk compounds all of it. In more than 70% of reported cases this year, software vendors had almost no time to patch flaws before attackers used them, roughly double the share in 2021, when providers typically had close to 300 days. AI models are speeding up both the attacks and the defenses.[2]

What the fund wants done

The IMF's recommendations are narrow. Put public debt on a sustainable path. Limit leverage and liquidity mismatches in AI and private credit, including capital and concentration requirements on banks exposed to those sectors. Strengthen cyber defenses through better incident reporting and resilience testing.[2]

Reuters reported that the fund also urged policymakers to curb the use of derivatives and margin loans where leverage is "excessive."[3]

For anyone holding AI-linked shares, the practical point is that the risk has become a financing story more than a technology story. The models keep improving. The open question, raised by both the IMF and Tan, is who pays for the next round of data centers, and whether that money comes back if the earnings arrive later than promised.[1][3][5]

Sources

1. IMF, Global Financial Stability Report, October 2026: Resilience, Risk, and Transformation
2. IMF Blog, Global Financial Markets Face Rising Risks From AI, Bond Yields, and Cyberattacks
3. Reuters, IMF cites AI slowdown, spiking yields as looming global market risks
4. The Guardian, Chip stocks slide after Broadcom boss likens AI investing to 'casino'
5. The Business Times, Firms still in search of AI use cases with worthwhile ROI, says Broadcom's Hock Tan