Hugging Face Raises $1B Series E at $10B Valuation
Hugging Face has closed a $1 billion Series E led by Salesforce Ventures and Google at a $10 billion valuation. The funds will go toward open-source model hosting, inference infrastructure, and enterprise tooling.
Original sourceHugging Face announced the close of a $1 billion Series E funding round, pushing the company's valuation to $10 billion. The round was co-led by Salesforce Ventures and Google, both of whom have deep strategic interests in the outcome of enterprise AI adoption. The capital will be directed toward expanding the Hub's model hosting capacity, improving inference infrastructure for both public and private deployments, and building out enterprise-grade tooling for teams running open-source models in production.
The raise is notable because Hugging Face occupies a structurally unusual position in the AI stack: it is simultaneously an open-source platform, a model registry, a managed inference provider, and an enterprise SaaS business. That breadth is both its defensibility and its strategic risk. The company hosts hundreds of thousands of models and datasets, making it the de facto distribution layer for the open-source AI ecosystem — a position that gives it leverage with both developers and enterprise buyers.
For the open-source AI community, the funding carries real stakes. Hugging Face has historically kept core infrastructure free and accessible, and the question now is whether enterprise-tier revenue pressures will shift that calculus. The company's ability to maintain community trust while monetizing at enterprise scale is the central tension its leadership will need to manage over the next funding cycle.
With Google and Salesforce both on the cap table, the company has secured strategic relationships that could accelerate enterprise distribution — but also introduced partners whose own cloud and CRM AI ambitions may not always align with Hugging Face's open-source positioning. How the company navigates those relationships while continuing to invest in the open ecosystem will define whether the $10B valuation holds.
Panel Takes
The Founder
Business & Market
“The moat here is real and specific: Hugging Face owns the distribution layer for open-source models, which means every enterprise team that standardizes on an open model eventually needs to touch the Hub. Having Google and Salesforce on the cap table is a double-edged sword — it accelerates enterprise distribution but introduces strategic investors who will eventually want to route that demand through their own clouds. The business survives if it becomes the npm of AI infrastructure; it doesn't if it becomes a feature inside Google Cloud's managed AI stack.”
The Futurist
Big Picture
“The thesis Hugging Face is betting on: open-source models will be the default choice for enterprise AI deployments by 2028, and whoever owns the hosting, versioning, and inference layer for those models owns the margin. That's a plausible and falsifiable claim — it depends on open-weight models continuing to close the capability gap with closed frontier models, which is a trend line that has held for three years running. The second-order effect that nobody is talking about: if this works, it shifts AI procurement from API subscriptions back toward infrastructure contracts, which is a massive power transfer away from OpenAI and Anthropic toward a much more fragmented, DevOps-style ecosystem.”
The Skeptic
Reality Check
“The valuation requires believing that Hugging Face can charge enterprise margins on infrastructure that the community expects to be free — and that Google, one of its lead investors, won't simply replicate the Hub's core functionality inside Vertex AI once the strategic value is absorbed. The scenario that kills this in 18 months isn't competition from a startup; it's Google deciding the cap table investment was cheaper than building a competitor and then building the competitor anyway. What would change my read: a clear, public breakdown of how much revenue comes from enterprise contracts versus inference API usage, because right now the monetization story is still opaque.”
The Builder
Developer Perspective
“The Hub's API is genuinely one of the better-designed ones in the ecosystem — the model card standard, the dataset viewer, the inference endpoints — these are primitives that compose cleanly rather than a platform you have to adopt wholesale. What I want to see this capital actually fund: inference infrastructure that doesn't route every cold-start through a 15-second spin-up, and dedicated endpoints that don't require a sales call to provision. If the enterprise tooling money goes toward making the DX for self-hosted deployments as clean as the Hub browsing experience, this is a genuine win for developers; if it goes toward a dashboard with a Salesforce SSO integration, I'll be skeptical.”