TS Lombard warns AI bust could erase $20 trillion
TS Lombard estimates that a collapse in the AI investment boom combined with a mild recession could wipe $20 trillion from equity markets. Economist Dario Perkins says Wall Street is increasingly focused on when an AI downturn might come and how severe it could be, rather than whether the sector is in a bubble.
Credit stress is already visible among AI infrastructure firms. CoreWeave’s five-year credit default swap spreads rose to 1,100–1,194 basis points in mid-October, implying a roughly 60%–63% default probability. Its unsecured bonds yielded close to 15%, while the company carried about $50 billion in debt. Neocloud stocks fell around 20% over five trading sessions, and CoreWeave shares dropped about 15%–25%.
TS Lombard also estimates that roughly 85% of AI ecosystem revenue is linked to recycled capital spending by large technology companies and related firms. If financing tightens or external customers do not generate sufficient demand, spending and revenue could weaken across the industry at the same time. Perkins cautions that the modeled $20 trillion equity-market decline assumes only a mild recession and does not necessarily imply a broad economic meltdown.
For crypto traders, the main concern is indirect: an AI-driven tech sell-off could increase risk aversion, pressure highly correlated assets and trigger volatility. Credit spreads, refinancing conditions and evidence of real-world AI demand are key indicators to monitor.
Bearish
The assessment is bearish because the article describes a potential unwind in AI-related investment, mounting credit stress at a highly leveraged infrastructure company, and crowded positioning in technology stocks. These factors could prompt broader risk reduction. For crypto markets, the effect is indirect: a sharp tech-sector sell-off can weaken appetite for speculative assets and increase volatility, particularly when equities and crypto are moving in tandem.
In the short term, traders may react to widening credit spreads, falling bond prices and further declines in AI-linked stocks by cutting risk across portfolios. That could weigh on Bitcoin and altcoins, though the article gives no direct evidence of crypto-specific selling or a change in blockchain fundamentals. If the credit warnings ease, external demand for AI proves durable, or markets treat the stress as contained, crypto could be less affected.
Over the longer term, the key issue is whether AI spending generates revenue from customers beyond the technology companies funding the buildout. A sustained funding squeeze or weak external demand could reinforce a wider risk-off cycle; a successful shift to real, recurring customer demand could reduce the concern. Past episodes of technology-sector stress have often affected crypto through liquidity and investor sentiment rather than direct exposure. Traders should therefore watch credit indicators, equity volatility, broader liquidity and crypto-specific flows. TS Lombard’s $20 trillion figure is a modeled equity-market scenario—not a forecast of crypto losses or a certainty that a crash will occur.