In a recent address, the chief of the Bank for International Settlements (BIS) sounded an alarm about the burgeoning race to pour capital into artificial intelligence (AI) technologies. While the enthusiasm surrounding AI is understandable—its potential to transform industries, boost productivity, and generate new revenue streams is immense—the BIS chief emphasized that the financing mechanisms behind this surge are worrisome. Much of the spending is being underwritten by debt that is not fully transparent, raising the specter of systemic risk that could ripple through the global financial system.
The warning is rooted in a historical perspective. The BIS official, Pablo Hernandez, drew parallels with two notorious episodes of speculative excess: the 19th‑century railway mania and the late‑1990s dot‑com bubble. In both cases, investors poured massive sums into projects that promised revolutionary change, yet many of those projects were financed through leveraged structures that obscured the true level of indebtedness.
When the anticipated returns failed to materialise, the resulting corrections were severe, leading to bankruptcies, market crashes, and prolonged economic downturns. Applying that lens to today’s AI landscape, Hernandez noted that corporations, venture capital firms, and even sovereign wealth funds are eager to secure a foothold in what they perceive as the next great technological wave.
Companies are racing to acquire AI talent, purchase cutting‑edge hardware, and develop proprietary algorithms. However, the speed of this capital deployment often outpaces rigorous financial scrutiny. Debt instruments—such as corporate bonds, convertible notes, and private placements—are being issued with limited disclosure about the underlying cash‑flow assumptions and the realistic timelines for monetising AI initiatives. One of the core concerns highlighted by the BIS chief is the opacity of these financing arrangements.
Traditional credit assessment tools rely on clear, verifiable data about a borrower’s revenue streams, profit margins, and debt service capacity. In the AI sector, many firms are still in the research and development phase, generating little or no revenue while incurring substantial expenses on talent, data acquisition, and high‑performance computing infrastructure. When lenders or investors cannot accurately gauge the future earnings potential, they may price the debt based on optimistic scenarios that do not reflect market realities.
This creates a hidden layer of leverage that could be difficult to unwind if the sector’s growth stalls. The systemic implications are profound.
Financial institutions that have exposure to AI‑related debt—whether directly through loan portfolios or indirectly via investment funds—could see a rapid deterioration in asset quality if a correction occurs. A cascade of defaults could tighten credit conditions across the broader economy, amplifying the shock. Moreover, because AI technologies are increasingly embedded in critical infrastructure—such as finance, healthcare, and transportation—a disruption in the sector could have knock‑on effects that extend beyond pure financial metrics, affecting operational stability and public confidence.
To mitigate these risks, Hernandez called for greater transparency and prudence in the way AI capital is allocated. He urged regulators, rating agencies, and market participants to demand clearer disclosures about the assumptions underpinning AI‑related debt. This includes detailed breakdowns of projected cash flows, realistic timelines for product roll‑outs, and contingency plans for scenarios where anticipated market adoption falls short.
Additionally, he suggested that banks adopt more conservative lending standards for AI ventures, perhaps by requiring higher capital buffers or by limiting exposure to a single borrower’s AI projects. The BIS chief also highlighted the role of policymakers in shaping a balanced approach. While it is essential not to stifle innovation—over‑regulation could deter the very advancements that promise societal benefits—there is a need for a calibrated framework that safeguards financial stability.
This could involve coordinated guidance on best practices for AI financing, cross‑border information sharing about emerging risks, and the development of stress‑testing models that incorporate AI‑specific variables. In concluding his remarks, Hernandez reminded the audience that technological revolutions have historically been double‑edged swords. The railway boom opened new markets and spurred economic integration, yet it also left a legacy of bankrupt lines and financial panic. The dot‑com era birthed the internet, which now underpins modern commerce, but it also produced a painful crash that wiped out billions in market value.
The current AI wave holds comparable promise, but without vigilant oversight and transparent financing, it could repeat the mistakes of the past. Stakeholders across the spectrum—corporate leaders, investors, regulators, and central banks—must therefore collaborate to ensure that the AI capital surge is built on solid, visible foundations rather than hidden debt.
By doing so, the global economy can harness the transformative power of artificial intelligence while preserving the resilience of the financial system against potential shocks.