AI’s energy use is big, growing fast, and very hard to measure or…

Sarah J · 2026-01-22T08:56:38.000Z

Key learnings: There’s major uncertainty: companies don’t disclose enough, and the tech changes so quickly that current and future energy use are both hard to pin down. Because of that uncertainty, governments and firms are planning for worst‑case demand, fast‑tracking new power projects, often fossil‑fuel based, which can undermine climate targets. The energy impact of individual user queries (like saying “please” and “thank you”) is tiny compared with large‑scale AI deployment. The real environmental problem comes from industry‑driven, economy‑wide integration of AI and ever‑larger models, not from small user behaviours. Better AI policy depends on much better, more transparent data on actual energy use and how it changes over time. https://oilprice.com/Energy/Energy-General/The-Hidden-Energy-Costs-of-Artificial-Intelligence.html

Member Login Get Access

AI’s energy use is big, growing fast, and very hard to measure or… | SEINET