The Bank for International Settlements is not in the business of market timing. It is in the business of identifying structural vulnerabilities before they become systemic. So when the BIS flags an artificial-intelligence bust as one of the most alarming threats to global prosperity, commercial real estate capital markets should read the warning not as a tech-sector forecast but as a liquidity scenario.
The BIS is not predicting an AI crash. It is mapping the transmission channels through which a sharp reversal in AI-related investment and equity valuations would hit growth, inflation expectations, and credit conditions. For CRE, those channels matter more than the AI thesis itself.
Here is the capital markets logic. Since late 2023, a meaningful portion of risk appetite in public and private debt markets has been supported by the AI narrative. It has buoyed equity indices, compressed credit spreads, and sustained investor willingness to underwrite long-duration assets. Office owners hoping for a leasing rebound tied to AI tenants, data center developers financing on forward rent assumptions, and lenders extending construction loans on speculative industrial projects have all, to varying degrees, priced in AI-driven demand.
The BIS is saying that scenario is not guaranteed. And in capital markets, the removal of a priced-in assumption is often more disruptive than the arrival of bad news.
What would an AI bust mean for CRE capital? First, credit spreads would widen. The risk premium on real estate debt, already elevated relative to pre-2022 levels, would face another leg up. Lenders would tighten underwriting not because real estate fundamentals deteriorated overnight but because the macro backdrop became less certain. That means lower loan proceeds, higher required debt yields, and fewer refinancing options for maturing loans.
Second, equity capital would become more expensive and more selective. The same institutional allocators who have been rotating into real estate as a yield alternative would face portfolio-level stress from public equity losses. Their willingness to commit fresh equity to illiquid assets would decline. The bid for distressed assets, already narrow, would narrow further.
Third, the data center and industrial segments that have absorbed enormous amounts of debt and equity capital on AI-driven assumptions would face a valuation reset. Not because the buildings are empty, but because the income growth trajectory that justified the basis would no longer be credible. Lenders who financed those assets at peak leverage would find themselves managing exposure to a sector whose underwriting thesis has been called into question.
Who benefits in this scenario? Owners of stabilized, cash-flowing assets with low leverage and long-dated fixed-rate debt. Their capital structures do not depend on a bullish macro narrative. They depend on rent collection. In a credit tightening cycle driven by macro uncertainty rather than real estate fundamentals, those assets become the only safe harbor for cautious capital.
Who is exposed? Sponsors with floating-rate debt maturing in the next 12 to 18 months, particularly on assets where the business plan assumed rent growth from AI-related tenants. Developers of speculative industrial and data center projects with construction loans that require a takeout in a tighter credit environment. And any owner whose refinancing plan depends on credit markets remaining as open as they are today.
The BIS warning is not a forecast. It is a stress test that the market should run on its own portfolio. The question every lender, owner, and investor should ask is not whether AI will bust. It is whether their capital structure can survive a scenario in which it does.