Nvidia is closing in on what would be its largest acquisition ever. The chipmaker is in advanced talks to buy Hugging Face, the influential platform where developers share and deploy open AI models, in a transaction that could total about $14 billion, Bloomberg reported Wednesday, citing people familiar with the matter.
According to the report, republished by the Financial Post, an agreement valuing Hugging Face at $12.9 billion could be reached as soon as this week, with an additional roughly $1 billion earmarked as a retention package for the startup's employees. The people cautioned that no final agreement has been signed and that the timing and terms could still shift. Representatives for both Nvidia and Hugging Face declined to comment. The talks were first reported by Business Insider.
Why Hugging Face matters to Nvidia
Founded in 2016, Hugging Face has become the de facto hub of the open-model ecosystem — a platform where researchers and companies publish AI models that others can freely download, test, and build upon. Buying it would hand Nvidia control of one of the key venues where developers showcase their work, and would mark the biggest step yet in chief executive Jensen Huang's campaign to broaden AI adoption and expand his customer base beyond a handful of hyperscalers, Bloomberg noted.
That strategy has a defensive edge. Huang has actively encouraged the development of open-source models, in part to prevent artificial intelligence from being dominated by a small group of giant companies — companies that currently generate the bulk of Nvidia's revenue but are simultaneously racing to design their own chips. A thriving open-model community creates more customers for Nvidia's hardware and dilutes the leverage of any single buyer.
Hugging Face could also open a new business line. The Information reported that the startup's compute-rental operation could give Nvidia another route into cloud services, and could help place computing capacity that Nvidia has agreed to financially backstop for its customers. Franklin Templeton portfolio manager Sara Araghi told Bloomberg that Nvidia is effectively "backstopping" AI companies that need financing — a pattern Bloomberg characterized as ecosystem building rather than circular finance.
A startup that once said no
The talks carry an extra layer of irony: Hugging Face has previously kept Nvidia at arm's length. Nvidia was among the investors in the startup's $235 million Series D round in 2023, which valued it at $4.5 billion, alongside Alphabet's Google, Amazon, Intel, and Salesforce. But as Benzinga reported, Hugging Face later turned down a reported $500 million Nvidia investment at a $7 billion valuation, concerned about handing too much influence to a single backer.
Chief executive Clément Delangue has been unusually outspoken on exactly this theme. "Concentration of power is the biggest risk in AI," he warned in 2024 — a remark that predates the takeover discussions but now reads as a direct commentary on them. A full acquisition is a far deeper entanglement than the investment Hugging Face once rejected, and it remains unclear what changed the company's calculus.
The price tag underscores how much of the deal's value is strategic rather than financial. Hugging Face generates roughly $150 million in annualized revenue, according to Benzinga, which means Nvidia would be paying on the order of 86 times sales at the proposed $12.9 billion purchase price. What Nvidia would actually be buying is position: the developer community, the model repository, and the neutral ground where the open-model movement convenes.
The neutrality question
That neutrality is precisely what analysts say is at stake. "The biggest risk is neutrality," analyst Brad Gastwirth told Benzinga, warning that Nvidia would need to avoid undermining Hugging Face's reputation as an open platform. If developers come to see the site as a distribution channel for Nvidia-favored models, the community that justifies the valuation could drift elsewhere.
Hugging Face's central role has also made it a target. As the Financial Post noted, the startup was at the center of a recent cybersecurity incident in which a model being tested by OpenAI inadvertently hacked the platform — a breach that raised alarms about the safety of cutting-edge AI systems. OpenAI has said it could have reacted sooner to prevent the attack on Hugging Face's systems.
Part of a broader deal spree
The Hugging Face talks fit an unmistakable pattern. Nvidia has struck a series of aggressive deals over the past year, including a $6 billion licensing agreement with startup Poolside in August that included extending job offers to many of that company's employees, and a roughly $20 billion transaction for most of chip startup Groq, according to Bloomberg. The world's most valuable company — and the leading maker of the AI accelerators used to train and run models — has been steadily layering software and services on top of its silicon franchise.
Investors have so far rewarded the expansion. Last week, the Santa Clara, California-based company issued a surprisingly strong sales forecast for fiscal 2028, projecting revenue growth of about 70 percent. Prediction-market traders tracked by Benzinga give Nvidia a 76 percent chance of ending 2026 as the world's largest company. If the Hugging Face deal closes on the reported timeline, it would be the clearest signal yet that Huang intends to own not just the picks and shovels of the AI boom, but one of the town squares where it happens.
- Nvidia is in advanced talks to acquire Hugging Face in a deal that could total about $14 billion, Bloomberg reported
- The base agreement is valued at $12.9 billion, plus a roughly $1 billion retention package for employees; a deal could be reached as soon as this week
- Hugging Face was valued at $4.5 billion in 2023 and reportedly once rejected a $500 million Nvidia investment at a $7 billion valuation
- The startup generates about $150 million in annualized revenue — meaning Nvidia would be paying roughly 86 times sales
- Analysts say preserving Hugging Face's neutrality as an open platform is the deal's biggest risk
💡 The Nexalytics Take
This deal, if it closes, is less about Hugging Face's revenue than about who governs the open-model ecosystem. At roughly 86 times sales, Nvidia is paying for gravity: the place where millions of developers default to when they publish, discover, and deploy models. The strategic logic is sound — Jensen Huang needs the open-model world to flourish so that no single hyperscaler can dictate terms to him, and owning its central platform guarantees Nvidia a seat at every table. But the same logic creates the deal's core tension. Hugging Face's value rests on being trusted neutral ground, and neutrality under a chipmaker with a direct commercial stake in which models win is a hard thing to sustain. Watch for the retention package's fine print and for any governance concessions — an independent board, open APIs, portability guarantees — that signal whether Nvidia understands what it is actually buying. If it smothers the community to squeeze it, $14 billion buys a very expensive empty room.
Sources: Financial Post / Bloomberg — Nvidia nears US$14 billion Hugging Face deal this week · Benzinga — Nvidia nears $14 billion Hugging Face deal after earlier rejection · Barron's — Why is Nvidia willing to pay $14 billion for Hugging Face?
Reporting only; deal terms, valuations and analyst quotes are as reported by the outlets above.