Hugging Face raises $100 million at a $2 billion valuation

Background

Hugging Face began in 2017 as a consumer chatbot app aimed at bored teenagers, and only later pivoted to natural language processing^[3]. Natural language processing, or NLP, is the branch of machine learning concerned with software that handles human language: sorting text, answering questions, translating, summarising.

The software that carried the pivot was the Transformers library, a package that lets a developer download a pre-trained language model and run it with a few lines of code instead of training one from scratch. At the time of the round the library had about 62,000 stars and 14,000 forks on GitHub^[3]. A star is roughly a bookmark and a vote of interest; a fork is a copy another developer can modify.

The round

On 9 May 2022 the company announced a $100 million Series C round^[1]. Lux Capital led the round^[2]. In venture financing a lead investor usually sets the terms, negotiates the valuation and often takes a board seat. The round valued the company at $2 billion^[1].

Sequoia and Coatue invested for the first time, alongside existing backers including Addition and Betaworks^[1]. Other existing investors at the time included AIX Ventures, Cygni Capital, Kevin Durant and Olivier Pomel^[3].

Two numbers are easy to confuse here. The $100 million is money that goes into the company to hire people and build things; the $2 billion is the price this deal put on the whole company. A valuation is not a bank balance — it is what investors agreed the business was worth per share at that moment, and it moves with the next deal.

The GitHub comparison

TechCrunch framed the company's goal as building the GitHub of machine learning^[1]. The comparison is to the code-hosting site: GitHub is where source code lives, where people file issues, open pull requests and build on each other's work. Applied to machine learning, the same shape would be a place where models and the datasets behind them are published, versioned, discussed and reused rather than kept inside one company.

The star and fork counts on the Transformers library are the part of that comparison that can be measured directly: they record how many people saved the repository and how many made their own copies of it^[3].

What the record shows

A Series C is generally the stage at which investors expect a product that many organisations already depend on, rather than an experiment. The most visible thing Hugging Face had at that point was a library other developers could pull into their own projects, which is a distribution channel built through open source rather than through a sales team^[3].

The concrete record left by the announcement is short: 2017 for the founding, 9 May 2022 for the round, $100 million raised, a $2 billion valuation, Lux Capital as lead, Sequoia and Coatue joining as new investors, and roughly 62,000 stars and 14,000 forks on the Transformers library at the time^[1][2][3].