causal

GPU coherence caps cluster size per chip generation

A single pre-training cluster cannot exceed roughly 200,000 Hoppers that are coherent, meaning all GPUs share memory and act as one system, so progress requires waiting for the next chip generation to scale further.

causalpending

Speaker

Gavin

Evidence Quote

“after xAI figured out how to get 200,000 Hoppers coherent, you had to wait for the next generation of chips”

Source

Nvidia v. Google, Scaling Laws, and the Economics of AI— Invest Like the Best
Created: 11/20/2026, 2:50:27 AM

Your notes

Loading notes...