We developed GATE: a model that shows how AI scaling and automation will impact growth. It predicts trillion‐dollar infrastructure investments, 30% annual growth, and full automation in decades. Tweak the parameters—these transformative outcomes are surprisingly hard to avoid.

After 2 years tracking GPU clusters, we're releasing our dataset of 700+ AI supercomputers. The US clearly dominates countries, and xAI’s Colossus is leading system: 200k AI chips, $7B price tag, and power needs of a medium-sized city. Here are my personal highlights🧵

Excited to share what we've been working on! AI supercomputers are the infrastructure behind training large AI models, but there has been very little public data tracking them. Our paper examining these systems is available now, with the complete dataset following in early May

THE dataset that I've been looking for. The data points include firm, country, which chips they have, total performance, cost...

Our AI Supercomputers database is now live! Explore specs for 700+ training & inference clusters—compute, chip types, power draw, ownership, location, and more.

TIL DeepSeek has the world's most cited paper on longtermism

An approximation I like is that an H100 is roughly a studio apartment