$800 Billion AI Bet With No Revenue to Back It

Hyperscalers will spend over $800 billion on AI infrastructure in 2026—matching the combined capex of all non-tech S&P 500 companies combined. The problem: the revenue needed to justify that outlay doesn't yet exist. On June 23, investors sold AI stocks as they recalibrated whether returns will arrive fast enough to support current valuations. Higher discount rates compress future cash flow projections when the payoff remains uncertain.