Jensen Huang and Clem Delangue Frame Nvidia's Hugging Face Bet as Open Source Scale
On 3 September 2026 Nvidia CEO Jensen Huang and Hugging Face CEO Clem Delangue told CNBC the $12.9 billion deal funds open model distribution, not a closed stack. Enterprise teams should read their public remarks as a signal about where compute and model hosting will concentrate.
Classy aggregates publicly available material; we did not conduct a private interview.
Watch the public conversation
What changed
On 3 September 2026, Nvidia founder and CEO Jensen Huang and Hugging Face CEO Clément Delangue appeared on CNBC's Squawk Box to explain Nvidia's agreement to acquire Hugging Face for $12.9 billion. Delangue said the founders approached Huang during the summer because open source AI needed more resources and scale. Huang confirmed Nvidia would keep Hugging Face as an independent, neutral platform inside Nvidia and said open models are a major growth driver alongside closed API labs.
Huang stated Hugging Face hosts roughly 200,000 enterprise customers, 18 million developers, and about 3 million models spanning language, robotics, physics, chemistry, and biology. Delangue said the three founders and the full team will join Nvidia, with a goal of reaching 100 million AI builders on the platform. When asked about talent retention, Huang confirmed a retention plan exists, calling Hugging Face talent "everything."
Why it matters
The conversation is one of the clearest public records of how a compute leader is positioning open weights relative to closed frontier APIs. Delangue tied the deal to Hugging Face's need to defend itself with open models after proprietary APIs could not be used during a security incident earlier in 2026. Huang argued many enterprises, sovereign programs, and science teams still need ownership of models trained on proprietary data, even as Nvidia itself uses rented closed services like Codex and Cursor internally.
For platform and procurement leaders, the interview clarifies intent: Nvidia says Hugging Face compute will not be required to build or deploy through the hub, and Huang framed that as consistent with Nvidia's broader open contributions, including Megatron Core and Dynamo.
Who is affected
Enterprise AI platform owners deciding whether to standardize on rented closed APIs or self hosted open stacks. Open model contributors who depend on Hugging Face hosting, datasets, and community tooling. Cloud and neo cloud vendors competing for training and inference spend tied to open weights. Security and compliance teams watching whether a chip company's ownership changes neutrality claims for a global model registry.
What to do next
Map your current model supply chain against the commitments Huang and Delangue stated on air: independent platform operation, no mandatory Nvidia compute on Hugging Face, and continued open contributions. If your roadmap assumes Hugging Face stays vendor neutral, assign one owner to track post close governance and API terms rather than treating the acquisition headline as proof of continuity.
What to watch
Regulatory review timelines and any published operating principles for Hugging Face inside Nvidia. Delangue named a 100 million builder target; track quarterly community metrics and whether third party cloud integrations remain first class.
Sources
- Primary. CNBC, Transcript: Jensen Huang and Clément Delangue on Squawk Box (3 September 2026). Deal price, team retention, open model rationale, and platform scale figures.
- Primary. CNBC, Full interview video (3 September 2026). Public record for quoted lines and broadcast context.