TS Lombard scenario analysis: AI bubble burst could trigger a 30% plunge in the S&P 500 and $20 trillion in market value wiped outbut don't rush to short!

In a scenario analysis released Thursday, TS Lombard's Dario Perkins predicted that if AI capital expenditure collapses, U.S. stock market capitalization would evaporate by roughly $20 trillion, triggering $125 billion to $190 billion in credit losses and bringing about a mild recessionbut investors can protect themselves by diversifying into global equities and buying bonds at the right time. Perkins modeled the financial shock of a sharp drop in AI capital expenditure, projecting that the S&P 500 would fall 30%, tech stocks would drop 40%, and other sectors would decline 20%. "Market concentration means that a downturn in AI-related spending would deal a heavy blow to U.S. stocks," Perkins wrote. "This would trigger massive losses and credit stress in the tech sector." Credit losses would be concentrated in the weaker links of the AI ecosystemneocloud providers such as CoreWeave, independent model developers such as OpenAI and Anthropic, and construction suppliersrather than in hyperscalers. The latter have strong balance sheets and should be able to withstand the shock. By the end of 2027, the non-hyperscaler portion of the AI ecosystem is expected to hold about $545 billion in outstanding bonds, $920 billion in bank loans, and $650 billion in private credit exposure. Despite the staggering wealth destruction, Perkins expects only a mild recession similar to those of 1990 or 2001. The end of the AI data center boom would reduce U.S. GDP by about 2 percentage points and cut roughly 800,000 construction-related jobs. "But as far as recessions go, we don't think this one will be particularly severe," he noted, citing healthy private-sector balance sheets and ample room for the Fed to cut rates. On Thursday, renewed concerns about AI sustainability, rising costs, and commercial returns drove investors out of stocksespecially tech stocksand into government bonds as a safe haven. This followed warning signals from both Meta Platforms and IBM after their earnings reports. For investors navigating this uncertainty, Perkins recommends patience and diversification rather than aggressive shorting. "We don't know exactly when this AI capex boom will end," he admitted. "Investors who short U.S. tech stocks (or extend duration) too early will definitely underperform." The recommended strategy is to maintain normal risk exposure through a diversified international equity portfoliowhich should perform reasonably well if the boom continues and outperform if it collapses. "This bubble is different from the dot-com bubble because it is more concentrated in the U.S. and the AI supply chain," Perkins explained, noting that non-U.S. markets have not become concentrated bets on a single theme. The report said government bonds should be added "only after signs of fragility emerge"; long-term rates are expected to decline significantly as the Fed responds to the AI-induced demand shock. Meanwhile, the dollar "will ultimately weaken" because the damage is concentrated in the U.S. financial system.