Wednesday Aug 19
HARDWAREOther

Most AI Datacenter Power Won't Arrive

19AUG
72% VAPOR1,066 GW ASK28% REAL

US datacenters have asked grid operators for 1,066 gigawatts of power. Wood Mackenzie projects only about 28 percent will ever get committed. The AI buildout's pipeline is mostly paper.

The mismatch is structural. Datacenter demand jumped from 23 gigawatts in 2023 to 42 in 2026, but a power plant takes 5 to 10 years to build. A datacenter takes under three.

Developers know it, so they file the same project with multiple utilities and keep whichever connection lands first. Grid queues are full of duplicates and options, not commitments.

For anyone modeling AI capacity, headline gigawatt announcements are now a vanity metric. Signed interconnection agreements are the real number - and they're a quarter of the ask.

full brief & sources

⚡ Why this matters

  • Every AI roadmap quietly assumes power that this analysis says will never exist.
  • It separates the real buildout from the announced one - a distinction worth billions in capex planning.
  • Grid interconnection, not GPUs, is becoming the scarcest resource in AI.

🔍 What happened

  • Aug 12 - Bloomberg reports Wood Mackenzie's analysis of US grid interconnection requests.
  • US datacenters have requested 1,066 gigawatts of power from grid operators.
  • Wood Mackenzie projects operators will commit to only about 28 percent of that.
  • US datacenter electricity demand grew from 23 gigawatts in 2023 to 42 gigawatts in 2026.
  • Power systems take 5 to 10 or more years to deploy; datacenter facilities take under three.
  • Developers duplicate-file projects across multiple utilities, inflating the visible queue.

💬 Smart takes

  • Wood Mackenzie: more than two-thirds of the electricity sought for US AI datacenters is unlikely to ever materialize.
  • Pat Gelsinger, former Intel CEO: the GPU-plus-HBM stack is power-hungry and computationally inefficient - the demand side needs fixing too.
  • Skeptic: projections cut both ways - fast-tracked permitting or behind-the-meter generation could shrink the vapor share quickly.

🧭 Where this goes

  1. Likelyutilities start charging real money for queue positions to flush duplicate filings.
  2. Likelymore labs and hyperscalers sign behind-the-meter nuclear and gas deals to skip the queue.
  3. Possiblea major announced AI campus quietly cancels within two quarters.
  4. Wild Cardfederal preemption of state interconnection queues for strategic AI projects.

🥄 The Spoon Take

Compute roadmaps are being written against power that will never exist. The winners of the next phase aren't the labs with the biggest datacenter announcements - they're the ones holding signed grid connections and their own generation. Power contracts are the new GPU allocations.

🤔 Pushback

The 28 percent figure is itself a projection - a policy shift on permitting or transmission could make it look far too pessimistic.

·Bloomberg - Wood Mackenzie analysis·Data Center Knowledge - August developments·iRecruit - Power project tracker