Hyperscalers face $730 billion in 2026 AI spending, but stock correlation with chipmakers hits a 4.5-year low.
Four major hyperscalers are projected to spend $730 billion on AI infrastructure in 2026, yet investor confidence in returns is waning. The 30-day correlation between AI spenders and semiconductor stocks has dropped from +0.78 to near zero, the lowest in 4.5 years.
Alphabet and Meta have recently increased their 2026 capital expenditure forecasts, joining Microsoft and Amazon in a record-setting spending cycle. Combined off-balance-sheet AI obligations for these firms now total $1.65 trillion, raising questions about sustainable cash flow.
Investors are shifting focus from infrastructure builders to companies that can convert spending into profitable returns, signaling a potential pivot in the AI investment landscape.