# The endgame is one H100-class GPU per user, refreshed yearly

- **Speaker:** Greg Osuri
- **Attribution:** attributed
- **Said:** 2024-10-02
- **Theme:** GPU Economics — https://www.gregosuri.com/predictions/gpu-economics/
- **Source:** "Will demand for advanced AI chips (GPUs) be 1:1 for every person? We sit down with Greg Osuri to find out." (Block Fuel)
- **Watch at:** 00:05:58 — https://podcasts.apple.com/us/podcast/will-demand-for-advanced-ai-chips-gpus-be-1-1-for-every/id1705468368?i=1000671509211
- **Canonical:** https://www.gregosuri.com/predictions/gpu-economics/2024-10-02-the-endgame-is-one-h100-class-gpu-per-user-refresh-tra3/
- **Cite as:** Greg Osuri, "The endgame is one H100-class GPU per user, refreshed yearly," "Will demand for advanced AI chips (GPUs) be 1:1 for every person? We sit down with Greg Osuri to find out." (Block Fuel), 2 Oct 2024, 00:05:58. https://podcasts.apple.com/us/podcast/will-demand-for-advanced-ai-chips-gpus-be-1-1-for-every/id1705468368?i=1000671509211

## Quote — verbatim

> Assume what the demand is going to be when we onboard 8 billion users, which is going to happen... imagine one H100 per user. And that gets outdated within a year. So then you have newer, bigger models that are more intelligent, that need bigger chips... I haven't even touched the surface of the kind of demand that I see for H100.

## Context — site annotation, not spoken

The episode's titular 1:1 chips-per-person prediction (spans into the 00:06:41 block); he repeats it later: "we need a GPU H100 per human being on the planet" (00:26:00), arguing no single protocol can serve that alone.