Oliver Wyman Models $33 Trillion AI Market Risk

Oliver Wyman has published a detailed assessment estimating that a full collapse in artificial intelligence valuations could eliminate approximately $33 trillion in global market capitalization. This figure exceeds the current annual GDP of the United States, indicating a correction of historic scale. The analysis distinguishes between a standard equity downturn and a more severe credit crisis driven by the leverage used to finance AI infrastructure.

The firm warns that up to $3 trillion of the $6 trillion in planned AI capital expenditures through 2030 may be debt-financed. If projected revenues fail to materialize, these obligations could trigger widespread defaults. Such a scenario would propagate risk from technology firms into the banking sector and broader financial markets, creating a systemic credit event rather than a simple stock price adjustment.

Debt-Financed Infrastructure Creates Credit Risk

The primary concern in the second scenario of the analysis is the amplification of losses through the debt markets. Companies are borrowing heavily to build data centers and acquire computing power, betting on future revenue streams that have not yet been realized. A failure to meet these financial targets would leave lenders holding non-performing assets, potentially freezing credit lines and impacting financial institutions that provided the capital.

This dynamic mirrors the mechanics of the 2008 financial crisis, where mortgage-backed securities amplified initial housing losses into a global banking panic. In the current AI landscape, the concentration of debt in a single technological sector means that a valuation reset could have immediate liquidity implications for the wider financial system, independent of consumer spending or other economic indicators.

Index Funds Drive Passive Exposure

Retirement accounts and index funds hold substantial positions in major AI developers due to market-cap weighting. With large technology firms comprising roughly 30% to 35% of the S&P 500, a significant decline in these stocks would drag down the entire index. This passive exposure means that investors without direct AI holdings still face material downside risk through their core equity portfolios.

The concentration risk is comparable to the pre-2000 dot-com era, where technology stocks represented a similar share of the market. A 40% drop in AI-adjacent equities could reduce the S&P 500 by 12% to 15%, even if other sectors remain stable. This structural link ties the performance of the broader market to the success of a limited number of high-growth firms.

Historical Parallels To Dot-Com Era

Comparisons to the 2000-2002 dot-com crash highlight the potential duration of a correction. The NASDAQ took seven years to recover its peak value, and unemployment remained elevated for nearly four years. While AI companies differ from dot-com firms in their current revenue generation, the reliance on future growth projections for valuation remains a key similarity.

According to coverage from GN auto stocks/technology: tech stocks, the current environment presents a unique risk profile. The scale of infrastructure spending and the depth of institutional ownership create a tighter feedback loop between asset prices and credit conditions. This suggests that a correction, if it occurs, may have more immediate and pervasive economic consequences than previous technology-driven bubbles.