In a recent address, the head of the Bank for International Settlements (BIS) sounded a stark warning about the rapidly escalating race among corporations to pour capital into artificial intelligence (AI) technologies. The official cautioned that the surge in AI‑related capital expenditures is being underpinned by a wave of debt that is often opaque, poorly documented, and difficult for regulators and investors to assess. This hidden borrowing, the BIS chief argued, could sow the seeds of a systemic financial crisis if it is not brought to light and managed prudently.
The speech highlighted how the current enthusiasm for AI mirrors past periods of speculative investment, such as the railway expansion of the 19th century and the dot‑com boom of the late 1990s. In each of those eras, firms rushed to fund projects that promised transformative breakthroughs, but many of those initiatives were financed through debt that was not fully transparent to the market. When the anticipated returns failed to materialise, the resulting defaults and write‑downs triggered broad‑based corrections that rippled through the wider economy. According to the BIS chief, the modern AI arms race is characterised by several distinct features.
First, the scale of spending is unprecedented. Companies across sectors—from finance and healthcare to manufacturing and retail—are allocating billions of dollars to develop, test, and deploy sophisticated machine‑learning models, generative AI tools, and autonomous systems.
This capital outlay is not limited to large, well‑capitalised firms; even smaller start‑ups are tapping into venture capital and public markets to secure the resources needed to stay competitive. Second, much of the financing is being sourced from non‑traditional lenders and shadow‑bank entities that operate with limited regulatory oversight.
These lenders often provide short‑term, high‑leveraged loans that are recorded off‑balance‑sheet or structured in complex ways that obscure the true level of indebtedness. As a result, the aggregate debt tied to AI projects is difficult for central banks, supervisory authorities, and even the firms themselves to fully gauge.
Third, the expected returns on AI investments remain highly uncertain. While some early adopters have reported impressive efficiency gains, cost reductions, and new revenue streams, many projects are still in experimental stages, with unclear pathways to profitability.
The hype surrounding AI, amplified by media coverage and corporate marketing, can create a feedback loop where investors pour money into the sector based on expectations rather than solid financial fundamentals. The BIS chief warned that this combination of massive, hype‑driven spending and hidden debt could lead to a classic "boom‑bust" cycle. If a significant number of AI‑related ventures fail to achieve their projected earnings, companies may find themselves unable to service their obligations, prompting a wave of defaults.
Because the debt is intertwined with broader financial markets—through syndicated loans, corporate bonds, and securitised assets—a wave of distress could quickly spread, affecting banks, investors, and ultimately consumers. To mitigate these risks, the BIS official called for greater transparency and stronger supervisory frameworks.
Key recommendations include: 1. **Standardised Disclosure Requirements**: Firms should be required to disclose AI‑related capital expenditures and the associated financing structures in a clear, comparable format. This would enable regulators and market participants to assess exposure levels more accurately. 2.
**Enhanced Stress‑Testing**: Central banks and supervisory bodies should incorporate AI‑related debt into their stress‑testing scenarios, evaluating how shocks to AI market sentiment or technology performance could impact financial stability. 3. **Regulation of Shadow Banking**: Greater oversight of non‑bank lenders that are active in AI financing would help ensure that the risk is not hidden in complex, off‑balance‑sheet vehicles. 4.
**Prudent Corporate Governance**: Boards and senior executives should adopt rigorous investment appraisal processes for AI projects, focusing on realistic return expectations and robust risk management. The speech also underscored the importance of international coordination.
Because AI development is a global endeavour and capital flows cross borders effortlessly, a fragmented regulatory approach could leave gaps that allow systemic risk to accumulate unnoticed. The BIS, as a forum for central banks, is positioned to facilitate dialogue and harmonise standards across jurisdictions. In concluding remarks, the BIS chief reiterated that while AI holds transformative potential—promising advances in productivity, healthcare outcomes, and climate solutions—it must be pursued with a disciplined financial strategy. The historical lessons from railway mania and the dot‑com bubble serve as cautionary tales: unchecked optimism combined with concealed borrowing can destabilise even the most robust economies.
By demanding transparency, improving oversight, and fostering responsible investment practices, policymakers can help ensure that the AI revolution contributes to sustainable growth rather than sowing the seeds of a future crisis. Overall, the message was clear: the AI capital‑expenditure race is not merely a technological competition; it is a financial phenomenon that requires vigilant monitoring, clear reporting, and prudent risk management to avoid repeating the mistakes of past speculative booms.