The biggest tech companies on the planet are preparing to spend somewhere between $660 billion and $690 billion on AI-related capital expenditures in 2026 alone. Wall Street expects total AI spending to blow past $900 billion, with cumulative outlays crossing the $1 trillion mark when you factor in recent and upcoming years.
The bottleneck money can’t buy its way out of
The core problem is deceptively simple: AI data centers need electricity, and the grid can’t deliver it fast enough. Interconnection delays for power connections often range from two to more than seven years.
AI workloads are driving electricity demand of 30 to over 100 kilowatts per rack. That’s a staggering jump from the previous standard of 5 to 15 kW.
Gartner projects that 40% of AI data centers will be power-constrained by 2027.
