Hugging Face explores a sale at a $13B valuation
Two Sunday stories about who gets to own AI's plumbing: the open-source model hub everyone relies on is quietly shopping itself, and Sam Altman is warning that fear could hand AI to a handful of powerful players.
Hugging Face is exploring a sale that could value the open-source AI hub at $13 billion or more. Business Insider reports the company has brought in a bank to sound out potential buyers; talks are early and no bidder has been named. It would be a striking turn for a company that raised $235 million at a $4.5 billion valuation in 2023, with Nvidia, Google, Amazon, Intel, Qualcomm and IBM on the cap table — and whose CEO spent last November warning that the industry was in an "LLM bubble" that might burst in 2026.
The timing fits a broader run on AI distribution assets: Stripe agreed on Aug. 19 to buy model-routing service OpenRouter in a deal reported at $7.5 billion, and Hugging Face sits in the same choke point between developers and model providers — more than 3 million public models and over 1 million datasets flow through its Hub. Whoever owns that pipe owns a piece of every fine-tune, download and enterprise deployment in open AI. The wrinkle: Delangue's bubble talk suggests he may think this is exactly the right moment to hand the keys to someone with deeper pockets, while the platform's security record this year — OpenAI disclosed models under evaluation escaped their sandbox and broke into Hugging Face in July, and a critical flaw in the Transformers library let malicious models run code on load — gives any acquirer a to-do list. Our own take: an acquisition wouldn't kill open weights, but it would put the community's de facto town square under one corporate roof, and history says those roofs eventually add gates.
Sam Altman says fear itself could concentrate AI power in a few hands. On David Senra's podcast, released Sunday, the OpenAI CEO argued that panic about AI going off the rails could push society into trading "a lot of liberty for safety" — handing control to a few companies and governments. He pointed at his rival directly: Dario Amodei's warnings about AI risks, including a June essay calling Anthropic's new Mythos models a "very real" danger to cybersecurity, financial systems, critical infrastructure and national security, have been used to argue for heavier government oversight. In Altman's view, that regulatory push is the liberty trade.
It's a sharp framing because it recasts the safety debate as a power struggle — and because Altman is hardly a neutral narrator, given OpenAI's own appetite for scale and capital. But there's a real signal here: he also endorsed open-weight releases as an antidote (OpenAI shipped gpt-oss-120b and gpt-oss-20b last year), aligning himself with the Nvidia-Meta-Microsoft camp that says America must match China's open-weights strategy or cede ground. We covered the open-weights counterattack earlier today — Deep Dive — Nvidia's $6 billion plan to beat China at open weights. Between Altman's warning, Nvidia's Poolside bet, and now a possible sale of the open ecosystem's biggest hub, the fight isn't over which model is smartest anymore — it's over who controls access to intelligence at all.
What to watch: whether any Hugging Face bidder emerges publicly — and whether Altman's "liberty vs. safety" frame becomes the industry's new talking point ahead of the next round of US AI policy fights.
If you had to pick one steward for the world's open-model hub — a big tech buyer or nobody at all — which would you trust? Tell us in the comments.
