# Akash sustains 54-55% utilization across 1,000+ GPUs

- **Speaker:** Greg Osuri
- **Attribution:** attributed
- **Said:** 2025-11-04
- **Theme:** GPU Economics — https://www.gregosuri.com/predictions/gpu-economics/
- **Source:** Sponsored Session: Powering PyTorch: Decentralized Training for an Energy-Hungry Future - Greg Osuri (PyTorch)
- **Watch at:** 00:12:32 — https://www.youtube.com/watch?v=T3nA1nAaLss&t=752s
- **Canonical:** https://www.gregosuri.com/predictions/gpu-economics/2025-11-04-akash-sustains-54-55-utilization-across-1-000-gpus-a5s7/
- **Cite as:** Greg Osuri, "Akash sustains 54-55% utilization across 1,000+ GPUs," Sponsored Session: Powering PyTorch: Decentralized Training for an Energy-Hungry Future - Greg Osuri (PyTorch), 4 Nov 2025, 00:12:32. https://www.youtube.com/watch?v=T3nA1nAaLss&t=752s

## Quote — verbatim

> Since enabling GPUs, the network has scaled roughly from about 150 to about a thousand plus GPUs while sustaining roughly 54 to 55% average utilization across a very heterogeneous fleet — that's competitive with large clouds and far above typical on-prem. It means few idle watts and more useful gradient steps per kilowatt hour.

## Context — site annotation, not spoken

Presenting Akash's reverse-auction marketplace as "an autonomous utilization machine"; decentralized fleets hit 50-60% utilization, ~3x on-prem baselines ([00:15:30]).