# Crypto can pool capital to train frontier models (\~$50M for a Llama)

- **Speaker:** Greg Osuri
- **Attribution:** attributed
- **Said:** 2024-11-09
- **Theme:** Decentralized AI — https://www.gregosuri.com/predictions/decentralized-ai/
- **Source:** Why Decentralized AI Needs Cosmos: Greg Osuri of Akash Explains (The Interop)
- **Watch at:** 00:24:46 — https://www.youtube.com/watch?v=3_gGBRqzxrQ&t=1486s
- **Canonical:** https://www.gregosuri.com/predictions/decentralized-ai/2024-11-09-crypto-can-pool-capital-to-train-frontier-models-5-x6ai/
- **Cite as:** Greg Osuri, "Crypto can pool capital to train frontier models (\~$50M for a Llama)," Why Decentralized AI Needs Cosmos: Greg Osuri of Akash Explains (The Interop), 9 Nov 2024, 00:24:46. https://www.youtube.com/watch?v=3_gGBRqzxrQ&t=1486s

## Quote — verbatim

> I think you can train a good Llama equivalent for about $50 million... there's a lot of money in crypto that, if we all pull together, we can train the next trillion parameter model fairly well.

## Context — site annotation, not spoken

Back-of-envelope: Llama 3.1 trained on ~30K H100s for ~2 months; at ~$1/GPU/hr that's ~$30K/hour — affordable for token treasuries (he cites TAO's ~$4B market cap).