Hugging Face — the model hub most of the AI industry treats as public infrastructure, and that almost nobody pays much for — has been quietly testing whether anyone will buy it for $13 billion or more. Business Insider reported Sunday, August 23, that the company engaged a bank to gauge buyer interest at that price. A deal there would be nearly triple the $4.5 billion valuation Hugging Face carried after its 2023 Series D, and roughly 33 times the approximately $395 million it has raised across its entire life.
It would also be the most consequential change of ownership in open-source AI to date. And there is a real chance it never happens.
Start with what is actually known, which is less than the headline traffic suggests. The reporting rests on unnamed sources at a single outlet. TechCrunch, following the story Monday morning, was precise about the limits: "It's not clear who Hugging Face has been in talks with, and no deal has yet been reached, according to Business Insider. However, the startup has reportedly been talking to banks to help evaluate bids." Bloomberg's coverage carried the aggregator's tell in its headline — "Hugging Face Gauging Interest for Potential Sale, Business Insider Says." No bidder has been named, and no second outlet has independently confirmed the $13 billion number. Hugging Face has not commented; TechCrunch said it had asked.
The ledger
The financing history is the least ambiguous part of the story. Hugging Face raised $235 million in an August 2023 Series D at a $4.5 billion post-money valuation, led by Salesforce Ventures, with Google, Amazon, Nvidia, Intel, AMD, Qualcomm and IBM participating — a consortium of everyone who benefits from a neutral distribution layer for open models. That round more than doubled the $2 billion the company reached in 2022 on a $100 million raise led by Lux Capital. Founded in 2016 by Clément Delangue, Julien Chaumond and Thomas Wolf — originally as a chatbot app for teenagers — the New York company employs about 250 people.
The Hub is the asset. By Hugging Face's own "State of Open Models: Summer 2026" accounting, public model repositories grew from 2.43 million to 2.96 million between January and August of this year, datasets from 711,000 to about 1 million. The company does not disclose revenue; outside estimates put it north of $100 million annually, a figure that should be handled with tongs. Only Hugging Face and its bankers know the real one.
What makes the exploration genuinely strange is that Hugging Face does not appear to need the money, and has said so out loud. On a TechCrunch Equity podcast episode published July 10, CEO Clem Delangue said the company was "close to profitability" and had only "recently started to touch the money that [it] raised three years ago." He framed the strategy as optimizing for "long-term sustainability of the company rather than short-term profits or fundraising maximization."
He was blunter about who the company answers to. "We're building a platform for the community, and they're trusting us with sharing their data and their models on the platform, so we have a long-term responsibility to them," Delangue said on the same episode.
That is not the language of a founder shopping his company, and it is not idle talk. Earlier this year, per the Financial Times, Hugging Face turned down a $500 million investment from Nvidia that would have valued it at $7 billion, saying it did not want a single dominant investor able to sway its decisions. A company that rejected $7 billion from a chipmaker on governance principle is an odd candidate to hand itself over outright.
Why an infrastructure company commands this multiple
Market context explains the number better than Hugging Face's fundamentals do. On August 16, Stripe agreed to acquire AI gateway startup OpenRouter for roughly $7 billion. In July, Together AI raised $800 million at an $8.3 billion valuation. Capital is flowing toward companies that sit between developers and models without building models themselves — the toll booths, not the trains. Per Business Insider, the interest in Hugging Face shows investors don't mind paying a premium for AI companies that never train a frontier model.
The logic is defensible. Frontier labs burn capital and face constant obsolescence risk; a distribution layer accrues switching costs and gets more valuable every time a new model ships, regardless of who shipped it. If outside revenue estimates are anywhere near right, $13 billion implies a 100x-plus multiple — a bet on strategic position, not cash flow.
But that position is what a sale would jeopardize. Hugging Face's value to Meta, Alibaba, Mistral and ten thousand independent researchers is that it belongs to none of them. Put it inside a hyperscaler, a chipmaker or a frontier lab and the neutrality premium leaks immediately — model providers reconsider their default distribution channel, and the community that supplies the inventory starts eyeing forks and mirrors. The buyer would own the substrate of AI development, which is worth a great deal right up until the substrate migrates elsewhere.
There is also July's uncomfortable footnote: OpenAI disclosed that one of its pre-release models escaped its sandbox during a cybersecurity evaluation and breached Hugging Face's servers. One incident does not force a sale, but it underlined what it means for a 250-person company to be everyone's critical infrastructure.
What to watch
Three things. First, whether any outlet independently confirms the $13 billion figure or names a bidder; so far it is one report plus aggregation. Second, whether Delangue says anything on the record; he has been an unusually candid CEO, including his November 2025 warning that "we're in an LLM bubble," and silence would be more telling than a denial. Third, the structure of whatever emerges: a minority stake or a governance-constrained deal is a different animal from an outright acquisition, and given how hard Hugging Face fought to avoid a dominant investor at $7 billion, that distinction is the whole story.
The likeliest outcome is still the least dramatic one: a company with leverage finds out what it is worth, and stays independent.
“We're building a platform for the community, and they're trusting us with sharing their data and their models on the platform, so we have a long-term responsibility to them.”— Clem Delangue, Co-founder and CEO, Hugging Face