Key Takeaways

  • The Bank for International Settlements says AI funding routed through non-bank channels could unwind faster than past banking crises
  • Rapidly rising AI capex by hyperscalers like Alphabet, Microsoft, and Oracle is increasing leverage and valuation pressure
  • Analysts say debt-funded AI growth and high equity expectations create conditions that could amplify financial-system spillovers

The Bank for International Settlements delivered a clear message in its June 28, 2026 economic report: the speed and structure of today's artificial intelligence investment boom could turn an ordinary correction into something sharper and more abrupt than the world is used to.

Instead of relying mainly on traditional bank lending, funding for AI projects is now flowing heavily through hedge funds, private credit pools, and a mix of non-bank intermediaries. The BIS chief representative for Asia and Pacific regions noted that these channels tend to run with lighter oversight. They have become central to how AI activity is being financed, which raises systemic risks because liquidity in these markets can dry up quickly during a downturn.

According to the BIS, the scale of AI spending adds another layer of complexity. The five largest hyperscalers are on track to invest more than $1 trillion in AI-related capital expenditures during 2025 and 2026, a data point highlighted in recent coverage by Reuters. Alphabet, Microsoft, and Oracle are prominent among those firms. If even a fraction of these expectations resets, the impact would reach far beyond a handful of technology balance sheets.

BIS Bulletin 120 notes that AI investment has grown to represent a material share of GDP in several advanced economies. Some policymakers have welcomed the trend because it contributes to headline growth. Yet the report also explains that the sustainability of the boom depends heavily on firms delivering the high earnings currently embedded in their valuations. If revenue growth comes in softer than markets hope, valuations will face intense downward pressure.

One emerging concern is leverage. Instead of relying on internal cash generation, more companies are shifting toward debt financing, with private credit playing a rising role. The Bulletin says this pattern is accelerating. For buyers of private credit, the returns can look attractive, but system-level stability faces increasing risk. A drop in cash flow or a pause in AI adoption cycles can tighten liquidity in these structures in ways that propagate quickly across the broader market.

Another point raised by the BIS is the mismatch between equity and debt market pricing. Equity valuations tied to the AI theme have sprinted ahead, while debt markets have been slower to adjust. According to analysis discussed in Morningstar, that gap matters because lenders could be underestimating risk if they assume those equity-driven growth stories will materialize exactly as forecast. If earnings fall short, both asset classes could reprice at the same time, creating a feedback loop that central banks attempt to avoid.

The interconnectedness of today's global financial system means any adjustment could unfold much faster than prior banking crises. Non-bank intermediaries tend to move capital quickly, and they often have shorter redemption or financing cycles. That mix can amplify volatility when sentiment changes, accelerating capital flight.

Industry analysts outside the BIS have been tracking adjacent risks as well. Research from the CNCF community has pointed out that surging demand for training infrastructure is pushing companies to scale cloud environments faster than their governance models. Although this is an operational point rather than a financial one, it adds to the uncertainty around long-term cost curves. Meanwhile, the Harvard Business Review has explored how AI capital allocation is reshaping productivity expectations in ways that may take longer than equity markets tend to price in. When productivity shifts lag investment cycles, valuation pressure can build.

Not every perspective is cautionary. Some economists argue that periods of heavy capital expenditure have historically generated new productivity waves, even if the timing has been uneven. Still, the BIS is placing more weight on downside scenarios for now, mainly because funding structures have changed so quickly. Hedge funds and private credit vehicles can provide flexibility and speed, but they can also react abruptly to perceived risk.

There is also the governance dimension. The BIS report references ongoing work with financial-stability standards, while global AI practitioners often look to frameworks like the NIST AI Risk Management Framework for model governance. These two domains are not identical, but they are becoming more intertwined as AI investment ties operational risk to macro-financial flows. Enterprises deploying large-scale AI might focus on model bias or data lineage, while investors evaluate how those specific operational bottlenecks affect valuations.

Public and private sector leaders are actively adapting to this shift. Regulators in several markets have indicated that monitoring non-bank leverage is becoming a higher priority. Some central banks have been experimenting with supervisory dashboards that map private credit exposures related to AI infrastructure. These tools are still early, but they reflect a recognition that traditional bank-centric indicators might not capture the risk profile of today's capital cycles.

If the market keeps expanding at its current pace, the BIS expects AI investment to remain a substantive contributor to GDP growth. The catch is that this contribution depends on a delicate balance of earnings delivery, credit conditions, and investor confidence. Any material miss on those fronts could trigger the type of synchronized correction that policymakers attempt to avoid.

Enterprises, lenders, and policymakers are increasingly watching the same signals: capex trajectories, credit spreads, and early indicators of AI monetization. The BIS report establishes that understanding the structure of AI financing has become just as critical to market stability as analyzing the capabilities of the technology itself.