Nvidia Pays $12.93B for Hugging Face — 86x Its Revenue

Nvidia has agreed to acquire open-source AI hub Hugging Face for $12.93 billion, the chipmaker's second-largest deal ever, according to Nvidia's own announcement and reporting from CNBC and CNN on…

Nvidia’s $12.93 billion agreement to buy Hugging Face, confirmed Thursday after weeks of leaks from Business Insider and The Information, is less a software acquisition than an insurance policy. The world’s most valuable chipmaker is spending real money to control the community and distribution layer where open-weight models live, just as rivals like OpenAI, Anthropic, Amazon, and Meta chase their own custom silicon to escape Nvidia’s GPU tax.

The math is the tell. Hugging Face was generating roughly $150 million in annualized revenue, per The Information as cited by The Verge — meaning Nvidia is paying something like 86 times revenue for a platform that, by its own CEO Clément Delangue’s account to CNBC, only recently decided it needed “more resources, more scale, more visibility.” That’s a striking reversal from a year ago, when Hugging Face reportedly turned down a $500 million Nvidia investment at a $7 billion valuation over fears of a dominant single backer, according to the Financial Times. Whatever changed Delangue’s mind — the summer outreach to Jensen Huang, or the OpenAI-agent hacking incident that CNN and CNBC both note forced Hugging Face to lean on an open Chinese model for defense — Nvidia ended up paying nearly double that rejected price for full ownership instead of a stake.

Nvidia isn’t just buying market share in open-source AI, it’s buying the argument that open beats closed — with the platform where 18 million developers currently make that case for free.

Structurally, this is Nvidia’s second-biggest deal on record after the $20 billion Groq asset purchase in December, per CNBC, and it fits a pattern: build a moat not just in chips but in every layer developers touch, from Nemotron models to now the repository itself. Huang’s pledge that “Nvidia compute will not be required” to use Hugging Face is the kind of promise regulators and open-source purists will be watching closely — CUDA’s proprietary lock-in is the obvious precedent for how “open” commitments can erode once the acquirer needs the leverage.

“Together, we will scale Hugging Face’s platform, strengthen its infrastructure and expand access to AI for developers and institutions worldwide,” Nvidia CEO Jensen Huang said in his announcement. “Hugging Face will remain an open platform for the entire AI ecosystem… Nvidia compute will not be required to build on or deploy through Hugging Face.”

The Verge AI

Read the full story at The Verge AI →

The Data Commenter, in your inbox

Data markets, alt data, and the AI training-data economy. No spam, unsubscribe anytime.

Discussion lives in the inline notes attached to article passages.