# Globally distributed consumer-GPU clusters viable within two years

- **Speaker:** Greg Osuri
- **Attribution:** attributed
- **Said:** 2023-09-28
- **Theme:** GPU Economics — https://www.gregosuri.com/predictions/gpu-economics/
- **Source:** Building the Super Cloud of GPUs with Akash Founder Greg Osuri \| EP \#102 (Frictionless Podcast by Logan Jastremski)
- **Watch at:** 00:41:51 — https://www.youtube.com/watch?v=GVrfHDg30-M&t=2511s
- **Canonical:** https://www.gregosuri.com/predictions/gpu-economics/2023-09-28-globally-distributed-consumer-gpu-clusters-viable-vocw/
- **Cite as:** Greg Osuri, "Globally distributed consumer-GPU clusters viable within two years," Building the Super Cloud of GPUs with Akash Founder Greg Osuri \| EP \#102 (Frictionless Podcast by Logan Jastremski), 28 Sept 2023, 00:41:51. https://www.youtube.com/watch?v=GVrfHDg30-M&t=2511s

## Quote — verbatim

> Right now it's not the most optimal setup to leverage a globally distributed clusters of 4090s, but in two years I bet you is going to be the case — and that's really where we're heading.

## Context — site annotation, not spoken

Explains cost-of-compute vs. cost-of-communication tradeoff; argues research has optimized for local clusters only because AI is pre-productionization, and maturing open models (Llama 2, Alpaca on 3090s) will unlock distributed topologies.