# Distributed training becomes reality this year; state-of-the-art by \~2027

- **Speaker:** Greg Osuri
- **Attribution:** attributed
- **Said:** 2025-08-02
- **Theme:** Decentralized AI — https://www.gregosuri.com/predictions/decentralized-ai/
- **Source:** Solving the AI energy crisis \| Greg Osuri on what it takes to power AI (Changelog)
- **Watch at:** 00:21:50 — https://www.youtube.com/watch?v=61OOQfY5cqU&t=1310s
- **Canonical:** https://www.gregosuri.com/predictions/decentralized-ai/2025-08-02-distributed-training-becomes-reality-this-year-sta-53mo/
- **Cite as:** Greg Osuri, "Distributed training becomes reality this year; state-of-the-art by \~2027," Solving the AI energy crisis \| Greg Osuri on what it takes to power AI (Changelog), 2 Aug 2025, 00:21:50. https://www.youtube.com/watch?v=61OOQfY5cqU&t=1310s

## Quote — verbatim

> There is a lot of work that indicates distributed training will become a reality by end of the year, and by end of the year we will produce a model as good as GPT-3. By end of next year, or maybe going into 2027, there's a good chance that we'll be able to produce a state-of-the-art model fully trained distributed — or decentralized rather.

## Context — site annotation, not spoken

After surveying DiLoCo (DeepMind), Prime Intellect's 32B model, Nous Research's DisTrO, Pluralis's asynchronous swarm training, and Gensyn's fault tolerance.