Hugging Face: the $70M/year anti-OpenAI growing 367% year-over-year
Jan-Erik Asplund
TL;DR: Since launching in 2016, Hugging Face has become one of the most influential companies in AI and the face of an insurgent open-source community going up against the closed-source OpenAI. For more, check out our Hugging Face report and dataset.
Alongside the rise of OpenAI and ChatGPT, which is censored, closed, and runs on proprietary models that have to be accessed through cloud APIs, there’s been the rise of open-source models that are uncensored, can be tweaked, and can be downloaded and run locally, like LLaMa, Mixtral, and Stable Diffusion.
Today, hundreds of thousands of these open-source models exist, with Hugging Face’s model hub as their main distribution channel.
Key points from our research:
- Hugging Face was founded in 2016 as a chatbot for teens, but they found product-market fit in 2018 after they open-sourced the Transformers library they built to run it, hitting 15,000 stars on Github in just one year. Transformers offered developers a set of simple APIs they could use to easily test and train a variety of AI models on their own data—Hugging Face then opened up their platform so that anyone could add their own models (now numbering 500,000) and built cloud-based collaboration tools for teams using them in production.
- Sacra estimates that Hugging Face hit $70M annual recurring revenue (ARR) at the end of 2023, up 367% from 2022, primarily off the growth of lucrative consulting contracts with many of the biggest companies in AI—Nvidia, Amazon, and Microsoft. Compare to OpenAI at $2B ARR at the end of 2023, up 900% from an estimated $200M at the end of 2022, and Anthropic at about $150M ARR at the end of 2023, up 1,400% from $10M at the end of 2022.
- Just as GitHub (NASDAQ: MSFT, ~$1.5B in revenue in 2023) in the early days prioritized user and network growth over monetization—giving away unlimited public repos and charging just $25/mo for private repos—Hugging Face lets users host unlimited models for free, charging only for their hosted inference offering and for enterprise features like SSO. The vast majority of Hugging Face’s ~$70M in annualized revenue today comes from the managed version of their product they’re selling into companies like Amazon, Nvidia, and Microsoft with the much bigger monetization opportunity in the future as a result of being the central collaborative tool for devs building with AI.
For more, check out this other research from our platform:
- Hugging Face (dataset)
- OpenAI (dataset)
- Anthropic (dataset)
- CoreWeave (dataset)
- Lambda Labs (dataset)
- Scale (dataset)
- Samiur Rahman, CEO of Heyday, on building a production-grade AI stack
- Geoff Charles, VP of Product at Ramp, on Ramp's AI flywheel
- Mike Knoop, co-founder of Zapier, on Zapier's LLM-powered future
- Oscar Beijbom, co-founder and CTO of Nyckel, on the opportunites in the AI/ML tooling market
- Cristóbal Valenzuela, CEO of Runway, on the state of generative AI in video
- Thilo Huellmann, CTO of Levity, on using no-code AI for workflow automation
- Dave Rogenmoser, CEO and co-founder of Jasper, on the generative AI opportunity
- Chris Lu, co-founder of Copy.ai, on the future of generative AI
