The International Monetary Fund (IMF) and the Reserve Bank of Australia (RBA) have issued stark warnings about the burgeoning corporate borrowing funding the artificial intelligence (AI) boom, signalling a potential risk to global financial markets. The IMF’s Global Financial Stability report highlighted that hyperscalers, the world’s largest cloud computing and data centre operators, are increasingly reliant on debt. These companies provide massive infrastructure for data storage and AI development, burning through cash for expansion despite strong core earnings.
This rapid increase in AI-related debt is a significant concern. Hyperscalers now represent approximately 10 per cent of global investment-grade corporate bond issuance, a sharp rise from under 1 per cent at the end of 2024. Schroders’ Australian head of credit, Helen Mason, projects this borrowing could reach US$300 billion ($428.4 billion) this year, potentially hitting US$400 billion by 2027.
The sheer volume has attracted a broad spectrum of investors beyond specialist credit funds, making their reaction unpredictable if AI sentiment shifts negatively. The IMF classified this as a “salient” risk, noting the investment surge rests on uncertain returns and could trigger a market correction. The RBA, in its September board meeting minutes, also expressed concerns about fragile investor sentiment in the tech sector, observing investors were accepting low compensation for risk.
Westpac chief economist Luci Ellis acknowledged valid questions about AI company valuations but questioned the RBA’s implied doubt over the technology’s ultimate success. Ellis suggested an “ugly unwind” of the data centre boom is plausible even if AI transforms the world. The IMF further warned that if AI underperforms, the resulting investment unwind could cut global GDP by up to 1.1 per cent by 2028, a risk amplified by leveraged bets on AI shares by hedge funds and retail investors.
