Introduction
Hugging Face runs the largest hub for open-source AI models, datasets, and applications, a platform that has become a default gathering point for the machine-learning community in much the way GitHub became the default home for open-source code. Founded in 2016 by French entrepreneurs Clément Delangue, Julien Chaumond, and Thomas Wolf, the company began as a chatbot app before pivoting toward open-source infrastructure after its Transformers library found rapid adoption among developers.
What distinguishes Hugging Face from most other prominent AI companies today is its funding posture: the company has not raised venture capital in nearly three years, has publicly turned down a large investment offer from Nvidia, and continues to grow primarily by selling enterprise features on top of its free, open platform rather than by raising ever-larger rounds to fund frontier-model research.
What the company does
Hugging Face’s platform lets AI researchers and developers host, share, and discover machine-learning models, datasets, and demo applications, hosting millions of public models and hundreds of thousands of datasets covering text, speech, image, and video AI. Its Transformers library provides a standardized set of APIs for testing and training a wide range of models, which helped establish Hugging Face as common infrastructure across the open-source AI community. On top of this free platform, the company sells managed, hosted inference and enterprise features such as single sign-on and private model hosting to companies that want to run open-source models in production without managing that infrastructure themselves.
Who it serves
Hugging Face’s user base includes millions of individual AI researchers and developers using the platform for free, alongside enterprise customers — including some of the largest AI infrastructure companies in the world, such as Amazon, Nvidia, and Microsoft — that pay for managed inference, hosting, and enterprise governance features. According to the company’s own figures, its platform hosts millions of models and datasets and serves tens of millions of AI builders worldwide.
Company background
Clément Delangue, Julien Chaumond, and Thomas Wolf founded Hugging Face in New York City in 2016, initially building a chatbot application aimed at teenagers before pivoting toward open-source machine-learning infrastructure after open-sourcing their Transformers library in 2018, which quickly gained traction among AI developers. The company’s most recent funding round, a $235 million Series D in 2023 led by Salesforce with participation from Google, Amazon, and Nvidia, valued Hugging Face at $4.5 billion. In late 2025, according to reporting by the Financial Times, Hugging Face turned down a roughly $500 million investment offer from Nvidia at a $7 billion valuation, choosing instead to avoid having a single dominant investor with outsized influence over its direction. As of mid-2026, the company says it still has capital remaining from its 2023 round and has no near-term plans to raise new funding or pursue a public listing, with Delangue describing the company’s approach as deliberately more capital-efficient than most of its AI-industry peers.
Product and AI capabilities
Hugging Face’s core technical contribution is standardizing access to a huge and rapidly growing catalog of open and openly licensed AI models spanning language, speech, vision, and robotics, rather than building its own frontier models to compete directly with labs such as OpenAI or Anthropic. Its enterprise business centers on Inference Endpoints and other managed-hosting products that let companies deploy open-source models with production-grade reliability, alongside governance features aimed at regulated enterprise customers. The company has continued to deepen partnerships with major AI infrastructure providers, including hosting Nvidia’s open robotics models through its LeRobot project and supporting integration with cloud AI platforms such as Microsoft’s Azure AI Foundry, positioning itself as neutral, model-agnostic infrastructure that AI labs, cloud providers, and enterprises all build on top of rather than compete against.
Key developments
Hugging Face’s most recent funding round, a $235 million Series D at a $4.5 billion valuation, closed in 2023 with Salesforce as lead investor. In late 2025, the company turned down a roughly $500 million investment offer from Nvidia at a $7 billion valuation, according to Financial Times reporting, to avoid ceding outsized influence to a single investor. As of 2026, the company has confirmed it has not raised new capital in nearly three years, continues to fund its growth through enterprise revenue, and has said an IPO is not a near-term priority, even as CEO Clément Delangue has expressed longer-term enthusiasm about eventually giving its large developer community a path to become shareholders.
Why it matters
Hugging Face’s deliberate distance from the capital-intensive frontier-model race is a useful counterpoint to the dominant AI-industry narrative of ever-larger funding rounds chasing ever-larger models. By focusing on being the shared, neutral infrastructure layer that the rest of the AI industry — including some of its own investors and competitors — builds on top of, Hugging Face offers a different theory of how to build a durable AI company: through community adoption and enterprise services revenue rather than compute-intensive model training.
Sector context
Within Brel’s enterprise software coverage, Hugging Face functions as foundational, model-agnostic infrastructure that other AI vendors build on top of, distinct from application-layer companies such as Writer, which trains its own proprietary models for specific enterprise use cases. Its open-source hosting model also complements governed enterprise AI platforms like Dataiku, which integrates with model providers including Hugging Face’s ecosystem rather than hosting models directly.
Sources and references
This profile draws on Observer’s reporting on Hugging Face’s funding history and its decision to turn down Nvidia’s investment offer, supplemented by Sacra’s independent research on the company’s revenue and business model.
- Observer — Delangue interview on Nvidia investment (2026)
- Observer — Hugging Face monetization strategy (2026)
- Sacra — Hugging Face business model research