Training energy doubles every ~2 years; gigawatt-scale training by 2030
Akash Network - Decentralized Cloud Built for AI's Next Frontier (DePIN Connection)
“The amount of energy required to train state-of-the-art models is doubling roughly every 2 years… by 2030 we’ll need about a gigawatt power install capacity to train the state-of-the-art AI if this trend continues, which looks like it’s continuing. Only thing that can provide reliably a gigawatt power is a nuclear reactor.” — 00:17:09
Context: Recapping his congressional testimony (8 months prior) and years of warnings; notes Elon Musk’s Dwarkesh podcast made the same argument. Followed by a long breakdown of why the US grid can’t scale.