Jefferies warns US AI capex risks capital destruction
Jefferies strategist Chris Wood warns that the US AI boom could end in major capital destruction as heavy investment meets falling prices and growing competition from Chinese models. US hyperscalers have indicated capital expenditure of about $695 billion in 2026, rising to a projected $870 billion in 2027. Wood argues that some spending may be “malinvestment,” especially as companies rely more on debt to fund data centres, chips and other infrastructure.
Usage and pricing data underline the challenge: Chinese AI models processed 36.39 trillion tokens in the week ending July 19, 2026, compared with 7.39 trillion for leading US models. Some Chinese models cost roughly one-quarter as much per token as US alternatives. Wood says lower prices and market-share gains could weaken US AI providers’ margins and make current spending plans unsustainable.
For investors, key indicators include hyperscaler capex guidance, AI token usage and prices, and the balance between debt and cash financing. The warning could also affect semiconductor and hardware suppliers if infrastructure spending slows. The article has no direct cryptocurrency catalyst, but any wider tech-sector sell-off or credit concerns could influence crypto sentiment through risk appetite.
Neutral
The article does not identify a direct cryptocurrency catalyst, so its immediate effect on crypto prices is likely to be indirect. If investors interpret the warning as evidence of excessive AI spending, debt exposure or weakening technology-sector returns, risk appetite could fall and add short-term pressure to crypto assets, which often move with broader risk markets. A pullback in AI infrastructure spending could also weigh on semiconductor and technology shares, potentially feeding into wider market sentiment.
However, the article focuses on AI investment economics rather than crypto-specific issues such as regulation, exchange flows, token adoption or blockchain activity. It therefore does not support a clear directional call for digital assets. Traders can monitor technology equities, credit spreads, hyperscaler capex guidance and broader risk indicators for signs of spillover. Longer term, the impact on crypto will depend on whether any tech-sector correction becomes a wider liquidity or credit event; absent that, the news is more likely to remain a sector-specific concern than a sustained crypto driver.