AI Prediction Index

One prediction, on the record

“If the observed trends continue, the leading AI supercomputer in June 2030 will need 2 million AI chips, cost $200 billion, and require 9 GW of power — the equivalent of 9 nuclear reactors, a scale beyond any existing industrial facility.”

Researcher, Georgetown University; lead author of Epoch AI's 'Trends in AI Supercomputers'

Trend extrapolation; the largest reported system at publication (xAI's Colossus) used 200,000 chips, cost ~$7B and drew ~300 MW.