🚨 FIVE COMPANIES ARE SPENDING $700 BILLION ON AI THIS YEAR. THE ELECTRICITY DOES NOT EXIST TO RUN IT.
Amazon. Google. Meta. Microsoft. Oracle. Combined capex in 2026: $700 to $725 billion. In 2024 the same five spent $200 billion. Tripled in two years. Amazon's capex alone exceeds the entire U.S. energy sector's annual spending. But here is what the earnings calls buried. Microsoft has $80 billion in Azure orders it cannot fulfill. Not because of chips. Not because of demand. Because the power grid cannot deliver the electricity to run the data centers already built. Goldman Sachs documents a U.S. data center capacity shortfall of 11 gigawatts today. Widening to 40 gigawatts by 2028. The interconnection queue holds up to 2.6 terawatts of projects waiting to connect. Median wait: four to five years. Power transformer lead times: 128 weeks. The companies are building faster than the grid can follow. The revenue is real. Microsoft A.I. run rate hit $37 billion, up 123% year over year. Google Cloud up 63%. AWS up 28%, its fastest in 15 quarters. But for every dollar of A.I. revenue, more than three dollars of capex is deployed. Q4 2026 earnings decide whether this is the early phase of the greatest technology cycle in history or the largest capital misallocation since the 1990s telecom overbuild. But one trade survives both outcomes. Power. If A.I. succeeds, electricity demand explodes for decades. If A.I. fails, the data centres are already built, the power agreements already signed, and hundreds of billions in committed infrastructure will consume electricity regardless. The GPU depreciates. The kilowatt-hour does not. China controls 70% of rare earth mining and 90% of processing. The magnets inside power transformers flow through Beijing. The Strait of Hormuz blockade is raising energy costs during the largest electricity demand surge in modern history. The screen says $700 billion in A.I. investment. The grid says the electricity is not there. The gap between those two numbers is the trade.