This is an open-data hub on the energy behind the AI boom: how much power AI data centres use, the ageing gas-turbine fleet they lean on, the nuclear and gas deals hyperscalers are signing, the waste heat they reject, and the power stations that serve them — all backed by the open PowerAtlas plant database and cited public sources.
Median 18y, 955 plants >20y — the fleet AI runs on
Microsoft–TMI, Amazon–Talen → the actual plants
Germany's 10% mandate (1 Jul 2026) + 300 TWh EU potential
~945 TWh by 2030 · why power is the constraint
On-site DC generation, from open data
Built on PowerAtlas open plant data (WRI GPPD / Global Energy Monitor, CC BY 4.0) + cited public sources (IEA, EIA, German EnEfG, announced PPAs). Author: Dmytro Aheiev (ORCID 0009-0001-5512-0291), Inzonex.
Every gas turbine, HRSG and steam line in this fleet — and every kW of data-centre waste heat — is a hot surface. Inzonex modular removable insulation cuts surface heat loss up to 96% while keeping flanges accessible for maintenance.
See Inzonex modular insulation →Externally-sourced figures are attributed and link to primary sources; deal terms are public company/trade-press announcements. Author: Dmytro Aheiev (ORCID 0009-0001-5512-0291).