Bank of England Chief Warns AI Market Crash Could Spread Across Global Economy

Artificial intelligence could contribute to a global economic downturn if the current boom in AI investment ends in a sharp market correction, Bank of England Governor Andrew Bailey has warned G20 finance ministers and central bank governors. Bailey, writing in his capacity as chair of the Financial Stability Board, said financial markets remain vulnerable to a potentially disorderly correction that could spread across borders if valuations of AI-related companies fall sharply. His warning comes as investors have poured enormous amounts of money into artificial intelligence companies, semiconductor manufacturers and hyperscale technology companies, pushing valuations to extraordinary levels. Bailey's concerns highlight a growing debate about whether the AI boom is creating a financial bubble that could eventually threaten the wider global economy.

The concern is not simply that AI companies could lose value. Bailey warned that the financial system could amplify the impact of a downturn because of the interaction between high stock-market valuations, increased investor borrowing and market concentration. A relatively small group of major technology companies now accounts for a substantial share of global equity-market value, while those companies are also increasingly invested in one another and dependent on the same AI infrastructure ecosystem. Bailey said this cross-investment between AI companies and large technology infrastructure providers could make a future correction more severe than a normal technology-sector sell-off. If investors begin simultaneously reducing their exposure to AI-related assets, losses could spread through financial markets and affect companies, lenders and investors far beyond the technology industry.

The extraordinary pace of AI investment has become one of the defining features of the current global economy. Technology companies are spending vast sums building data centres, buying advanced processors, developing large AI models and expanding cloud-computing capacity. Semiconductor manufacturers have experienced an enormous increase in demand, while companies involved in data-centre construction, networking equipment, electricity generation and energy infrastructure are also benefiting from the AI boom. Investors have responded by placing substantial amounts of capital into the companies expected to profit from this expansion. The result has been a powerful cycle in which rising valuations make it easier for companies to raise money, which then allows them to invest even more heavily in AI infrastructure.

Bailey's warning is particularly notable because it comes from a central banker whose primary responsibility is financial stability rather than technology policy. The Bank of England has previously warned that artificial intelligence could contribute to a stock-market bubble, increase cybersecurity vulnerabilities and create new forms of financial risk. The latest message to the G20 suggests that those concerns are becoming increasingly international in nature. Financial systems are closely interconnected, meaning a major correction in US or Asian technology markets could quickly affect investors, pension funds, banks and businesses in Europe and other regions. A global downturn would be particularly damaging if an AI-related market sell-off occurred at the same time as other economic shocks.

Another major concern is the possibility that increasingly autonomous AI systems could be used to attack financial institutions. Modern frontier AI models are becoming capable of performing more complex tasks with less human supervision, including analysing software, identifying vulnerabilities and interacting with digital systems. Bailey warned that financial institutions should prepare for cybersecurity incidents involving simultaneous disruption across multiple firms. A coordinated or AI-assisted cyberattack could potentially spread quickly through highly interconnected banking and financial networks, disrupting payment systems, trading platforms or other critical infrastructure. The fact that AI systems can operate at machine speed means that attacks could potentially develop faster than traditional human-led security teams are able to respond.

The cybersecurity issue has become more urgent as businesses increasingly give AI systems access to internal software and digital tools. Companies are experimenting with AI agents that can write and execute code, interact with websites, access corporate databases and complete multi-step tasks. These capabilities can dramatically increase productivity, but they also create new security vulnerabilities. A compromised or poorly controlled AI agent could potentially move between systems much faster than a conventional cyberattack. Earlier this month, more than 100 companies and technology organisations, including Google, Microsoft, Anthropic and OpenAI, called for stronger international cyber defences as AI capabilities continue to advance.

Bailey's concerns also reflect the unusual concentration of money flowing into a relatively small group of technology companies. Nvidia, which supplies many of the advanced chips required to train AI models, recently became the world's most valuable public company, reaching a valuation of approximately $5.1 trillion according to Sky News. Other companies involved in cloud computing, AI models and semiconductor manufacturing have also reached extremely high valuations as investors anticipate rapid future growth. Such valuations can be justified if AI generates enormous productivity improvements and profits, but they could become vulnerable if businesses fail to monetize their investments at the pace markets currently expect.

A sudden slowdown in AI spending could therefore have effects far beyond technology stocks. The AI industry has become increasingly important to the broader economy because billions of dollars are being invested in data centres, energy infrastructure, telecommunications networks and semiconductor manufacturing. Construction companies, equipment suppliers, utilities and financial institutions are all becoming connected to the AI investment cycle. If major technology companies suddenly reduce spending because investors question the profitability of AI, the effects could ripple through these supporting industries. Economists are therefore watching not only the share prices of technology companies but also the debt financing and capital expenditure supporting the AI boom.

The warning comes at a particularly sensitive moment for global financial markets. Bailey also referred to the volatility generated by energy shocks associated with the US-Iran war, which has affected oil prices and added to inflation concerns. Higher energy prices can put pressure on households and businesses while simultaneously complicating central-bank decisions over interest rates. If an AI market correction occurred alongside an energy shock, elevated inflation and fragile sovereign-debt markets, the resulting combination could be significantly harder for policymakers to manage. Bailey's message to G20 officials was therefore essentially about preparing for multiple risks that could interact rather than treating AI as an isolated technology issue.

The Financial Stability Board has called for international cooperation to address these emerging risks. Bailey wants policymakers and regulators to develop appropriate measures for the safe and responsible release and deployment of advanced AI models. This could include stronger cybersecurity standards, better information-sharing between financial institutions, improved oversight of AI-related financial exposures and international agreements covering the development of increasingly capable models. Because AI companies and financial institutions operate across multiple jurisdictions, regulators face a difficult challenge: rules introduced by one country may have limited effectiveness if companies can simply move operations or services to another market.

At the same time, Bailey's warning does not amount to a prediction that AI will inevitably cause a financial crisis. The governor is highlighting a potential vulnerability rather than declaring that a crash is certain. AI could still generate enormous productivity gains, support new industries and contribute to economic growth. Companies are developing AI systems that could improve healthcare, manufacturing, scientific research, logistics and professional services. The challenge for investors and policymakers is determining whether current valuations accurately reflect those future benefits or whether markets have become excessively optimistic about how quickly AI will transform the global economy.

For technology companies, the message is that growth must increasingly be matched by evidence of sustainable returns. The enormous capital spending required to build AI infrastructure means investors will eventually demand clearer evidence that those investments are generating revenue and profits. If that happens gradually, markets may adjust without serious disruption. If expectations suddenly collapse, however, companies with high valuations and significant debt exposure could experience much sharper declines. Bailey's comments suggest that regulators want financial institutions to be prepared for that possibility before it becomes a systemic problem rather than after it has already happened.

The debate over AI's economic impact is therefore moving into a new phase. For years, the central question was whether artificial intelligence would transform industries. Increasingly, policymakers are asking a second question: what happens if financial markets move too quickly in pricing that transformation? Bailey's warning to the G20 places AI alongside traditional financial stability concerns such as excessive leverage, concentrated investment and cross-border contagion. As AI continues attracting extraordinary levels of capital, the ability of governments, financial institutions and technology companies to manage those risks may become just as important as their ability to develop the technology itself.