In a recent address, the head of the Bank for International Settlements (BIS) sounded an alarm about the rapidly escalating spending on artificial intelligence technologies, emphasizing that the surge in capital expenditures is being underpinned by debt that is often hidden from public view. This lack of transparency, the BIS chief argued, creates a fertile ground for systemic vulnerabilities that could reverberate throughout the global financial system if left unchecked. The warning came at a time when corporations across a wide range of sectors are racing to embed AI into their operations, from automating routine tasks to developing sophisticated predictive analytics platforms.
While the promise of increased efficiency and new revenue streams is compelling, the pace of investment has outstripped the underlying profitability of many projects. Companies are borrowing heavily—sometimes through complex financing structures, off‑balance‑sheet arrangements, or shadow banking channels—to fund AI initiatives before a clear return on investment can be demonstrated. According to the BIS chief, this dynamic mirrors past episodes of speculative excess.
He cited the 19th‑century railway boom, when investors poured massive sums into rail infrastructure with little regard for the actual demand for transportation, and the late‑1990s dot‑com bubble, where internet startups attracted vast amounts of venture capital despite lacking sustainable business models. In both cases, the inflow of capital was driven more by hype and the fear of missing out than by solid fundamentals.
When the speculative fervor faded, the resulting corrections were painful, leading to bankruptcies, job losses, and broader economic downturns. In the current AI context, the BIS chief highlighted several key risk factors. First, the opacity of the debt that finances AI projects makes it difficult for regulators, investors, and even company boards to assess the true level of leverage.
Traditional financial statements may not fully capture liabilities hidden in special purpose vehicles, convertible notes, or other innovative financing tools that have become commonplace in tech‑heavy sectors. This lack of clarity hampers the ability of supervisors to monitor systemic exposure and intervene before problems cascade.
Second, the rapid pace of AI adoption creates a feedback loop where firms feel compelled to invest more simply to stay competitive, regardless of whether the technology delivers immediate value. This "catch‑up" pressure can lead to a race to the bottom, where companies accept increasingly unfavorable financing terms, further inflating debt levels. The BIS chief warned that such a spiral could erode balance sheets and reduce the capacity of firms to absorb shocks, such as a sudden slowdown in demand, regulatory changes, or a broader credit crunch. Third, the interconnected nature of modern financial markets means that distress in one sector can quickly spill over into others.
If a major AI‑focused company were to default on its obligations, the repercussions could affect banks, asset managers, and even sovereign wealth funds that hold exposure to the firm’s debt. The BIS chief emphasized that the systemic risk is not limited to the technology sector alone; it extends to any institution that has taken on opaque AI‑related liabilities, creating a web of hidden interdependencies. To mitigate these dangers, the BIS chief called for a multi‑pronged approach.
Greater transparency is paramount: companies should be required to disclose the nature, maturity, and terms of any financing used for AI projects, including off‑balance‑sheet arrangements. Standardized reporting frameworks could help investors and regulators compare leverage across firms and sectors more effectively.
In addition, prudential regulators should consider introducing macro‑prudential tools tailored to the AI investment boom. For example, counter‑cyclical capital buffers could be calibrated to reflect the level of AI‑related debt in a bank’s portfolio, encouraging lenders to be more cautious when extending credit for speculative technology projects. Stress‑testing exercises that incorporate scenarios of AI market corrections would also provide valuable insights into the resilience of financial institutions.
Policymakers, too, have a role to play. By fostering an environment where AI innovation is aligned with sustainable financing, governments can help ensure that the benefits of the technology are realized without creating undue financial fragility.
This could involve supporting public‑private partnerships that share risk more evenly, offering incentives for firms that adopt transparent financing practices, and investing in research that clarifies the long‑term economic impact of AI. Finally, the BIS chief urged corporate leaders to adopt a disciplined approach to AI spending. Rather than chasing the latest hype, executives should conduct rigorous cost‑benefit analyses, set realistic timelines for return on investment, and prioritize projects that have clear strategic relevance. By aligning AI initiatives with solid business fundamentals, companies can reduce their reliance on opaque debt and contribute to a more stable financial ecosystem.
In summary, the BIS chief’s message is clear: the AI capital‑expenditure race, while promising transformative gains, is being financed in ways that are often hidden and potentially hazardous. Drawing lessons from historical bubbles, he stresses that unchecked hype‑driven spending can lead to widespread economic corrections. Transparency, prudent regulation, and disciplined corporate governance are essential to prevent the current AI boom from evolving into a source of systemic risk for the global economy.