In recent weeks, the head of the Bank for International Settlements (BIS) has sounded a stark warning about the rapid escalation of capital expenditures in artificial intelligence (AI) technologies. The warning focuses not merely on the scale of the spending, but on the opaque nature of the financing that underpins it. According to the BIS chief, the current AI investment surge resembles past speculative bubbles—most notably the railway mania of the 19th century and the internet dot‑com frenzy of the late 1990s—where enthusiasm outpaced actual profitability, ultimately leading to broad market corrections and, in some cases, systemic crises.
The core of the concern lies in the fact that a substantial portion of AI‑related capital is being raised through debt instruments that lack transparency. Unlike traditional corporate bonds or publicly disclosed loans, many AI projects are funded via private placements, convertible notes, or novel financing structures that obscure the true level of leverage. This opacity makes it difficult for regulators, investors, and even the companies themselves to gauge the real risk exposure. When large sums are borrowed on the expectation of future AI‑driven revenue streams that may not materialise, the financial system becomes vulnerable to a cascade of defaults.
Historical analogues provide a useful lens for understanding the potential fallout. During the railway boom, investors poured capital into rail networks with the promise of transforming commerce and connectivity. Yet, many lines were over‑built, revenues fell short, and a wave of bankruptcies followed, shaking banks and investors alike.
Similarly, the dot‑com era saw a flood of venture capital and public market money poured into internet startups whose business models were unproven. When the bubble burst, the repercussions rippled through the broader economy, leading to a recession and a tightening of credit conditions. The BIS chief argues that AI could follow a comparable trajectory if the current hype-driven spending continues unchecked. AI firms are often valued on speculative future earnings rather than on solid cash‑flow fundamentals.
This valuation approach encourages companies to secure large loans, sometimes at favorable rates, based on projected AI breakthroughs that remain uncertain. If those projections fail to materialise—whether due to technical limitations, regulatory hurdles, or slower adoption rates—borrowers may struggle to service their debt, prompting a wave of distress in the banking sector.
Moreover, the interconnectedness of modern finance amplifies the risk. Many AI projects are backed by large institutional investors, sovereign wealth funds, and even central banks that have taken an active interest in fostering technological leadership. When these entities allocate capital through opaque channels, the potential for contagion grows.
A default in one high‑profile AI venture could erode confidence across the sector, prompting a broader pull‑back on credit and a tightening of liquidity. Regulators are therefore urged to enhance disclosure requirements for AI‑related debt. The BIS chief recommends that lenders provide clear information on loan terms, collateral, and the underlying assumptions about AI performance. Standardising reporting would enable market participants to assess the true leverage embedded in AI projects and to identify early warning signs of stress.
In addition to transparency, the BIS calls for a more prudent approach to capital allocation. Companies should be encouraged to align their financing structures with realistic revenue forecasts, rather than relying on optimistic hype. This could involve staggered funding rounds tied to measurable milestones, such as demonstrable improvements in model accuracy, successful deployment in commercial settings, or regulatory approvals for AI‑driven products. The warning also touches on broader macro‑economic implications.
An AI‑driven credit expansion could inflate asset prices, create mismatches between supply and demand for skilled talent, and exacerbate inequality if the benefits of AI accrue only to a narrow group of well‑funded firms. Policymakers need to consider these side effects when crafting incentives for AI development. In summary, the BIS chief’s cautionary message is clear: while AI holds transformative potential, the current race to pour capital into the sector is being financed in ways that could conceal significant debt burdens.
Without greater transparency and disciplined financing practices, the sector risks repeating the boom‑and‑bust cycles of past technological revolutions, with potentially severe repercussions for the stability of the global financial system. Stakeholders—from corporate executives to regulators and investors—must heed this advice to ensure that AI’s growth is sustainable, resilient, and truly beneficial to the broader economy.