Writing
Napkin Maths for AI at Scale
AI builders will spend about $780bn in 2026, 2.4% of US GDP. Through 2028 the pace of the build-out is set by power and electricians, which take years to add; capital becomes the limit only if AI revenue grows less than about 2.4 times a year.
10 mininteractive
From One Token to a Gigawatt
If most jobs used AI every day, the world would need hundreds of gigawatts of new data centres. Chips and capital could keep pace within a few years; grid connections are the slowest part of the build-out.
8 mininteractive
What Energy, Capital and Compute Say About When AGI Arrives
Electricity, capital and compute are pouring into AI at a pace with no precedent, yet a model’s release date predicts its ability better than any of them: ten times the compute buys about four months. Around 2030, electricity and then money start to limit the biggest training runs.
18 mininteractive