Andrew Bailey, Governor of the Bank of England, has raised alarms about AI-driven financial instability at the G20, warning that bloated startup valuations and excessive leverage across markets could spark the next major financial crisis.
Bailey's core concern centers on a structural vulnerability in the AI ecosystem. Massive capital flows into AI companies have inflated their valuations beyond justified fundamentals. Simultaneously, hyperscalers like Amazon, Google, and Microsoft hold significant stakes in AI startups while these same startups depend on hyperscaler infrastructure for computing power. This cross-investment web creates systemic risk. If one major player fails or stumbles, the interconnected nature of these relationships could trigger a cascade of failures that ripples through global financial markets.
The leverage problem compounds the exposure. AI companies operate with high debt loads while pursuing massive infrastructure buildouts. Rising interest rates make servicing this debt more expensive. Financial institutions have increasingly bet on AI's growth trajectory, embedding AI-related assets throughout the banking system. A downturn in AI investment or a correction in valuations could force rapid deleveraging, similar to patterns that preceded previous financial crises.
Bailey identified cyber threats from frontier AI models as an additional risk factor. Advanced AI systems trained on vast datasets could become targets for sophisticated attacks or might themselves pose security risks if compromised. Unlike traditional financial infrastructure, the security perimeters around cutting-edge AI systems remain poorly defined and inadequately tested at scale.
The regulatory gap compounds these risks. Many countries still lack comprehensive rules governing advanced AI development and deployment. This regulatory vacuum means oversight of AI-related financial exposure remains fragmented. Some jurisdictions impose strict AI governance; others impose nothing. This creates opportunities for regulatory arbitrage where capital flows toward jurisdictions with lighter-touch approaches, concentrating risk rather than distributing it safely.
Bailey's warnings reflect growing concern among central bankers and financial regulators worldwide. The AI boom has accelerated so rapidly that financial institutions and governments struggle to model true systemic risk. Unlike the 2008 financial crisis, where mortgage-backed securities were the primary toxic asset, AI-related risk operates through multiple channels simultaneously: direct investment exposure, indirect infrastructure dependencies, leverage proliferation, and cyber vulnerabilities.
The warning arrives at a critical moment. Private equity continues pouring billions into AI startups. Public markets have rewarded AI-adjacent companies with soaring valuations despite many generating minimal revenue. Competition for computing resources between AI companies and hyperscalers has driven GPU prices higher, increasing capital requirements for market entry. Smaller players face funding pressure while larger players accumulate more leverage to stay competitive.
Bailey's message to G20 finance ministers essentially calls for coordinated action on multiple fronts: valuation discipline, leverage caps, enhanced disclosure requirements, and harmonized AI governance standards. Without intervention, the current trajectory may produce an AI-specific financial bubble that proves impossible to deflate smoothly.
The challenge for policymakers involves moving quickly without stifling legitimate innovation. Unlike traditional financial regulation, AI governance involves novel technologies where long-term impacts remain uncertain. Overly restrictive rules risk pushing development into less-regulated jurisdictions. Insufficient rules invite systemic risk accumulation. Bailey's intervention suggests central banks view the balance as tilted dangerously toward risk.