Hugging Face raises $100m Series C

Hugging Face, the platform for hosting and sharing open machine-learning models and datasets, raised $100 million in a Series C round led by Lux Capital, with Sequoia and Coatue joining as new investors alongside existing backers. TechCrunch reported the round valued the company at $2 billion. The company said it had grown from 30 to more than 120 staff over the preceding year, and that its hub then hosted around 100,000 pre-trained models and 10,000 datasets used by more than 10,000 companies.

Unlike the frontier labs building proprietary models, Hugging Face’s business was infrastructure: a place for anyone to upload, discover and run models trained by others, plus open-source libraries (Transformers, Datasets, Tokenizers) that had become close to a default toolchain for working with them. The company said the new funding would go toward research, open-source development, and what it called “responsible democratization” of machine learning, citing bias, privacy and energy consumption as challenges it wanted to address rather than leave to individual model builders.

The round is a marker of how quickly “open” had become investable as a strategy in its own right, months before ChatGPT made generative AI a mainstream story. Hugging Face’s position as the central distribution point for open-weight models grew substantially over the following two years, as it became the default release channel for open models from Meta, Mistral and others.