Hugging Face Co-founder Estimates Open-Model Compute at ~$1 Billion
Hugging Face co-founder Thomas Wolf estimates that open-weight models on the Hub represent nearly $1 billion in training compute, contributed by some 30+ labs including Meta, NVIDIA, Alibaba, DeepSeek, Moonshot and Mistral—up about 25x in two years.
2026-10-18 ~ 2026-10-18 · 3 related posts
- Hugging Face founder: $1B of training compute shared free on the Hub, up 25x since 2023 — Thom_Wolf · 2026-10-18
- Labs Shared ~$1B of Training Compute Free on Hugging Face, Up From $40M in 2023 — mervenoyann · 2026-10-18
1 near-duplicate retellings: Thom_Wolf