Nvidia acquires Hugging Face in $12.93bn deal

Developing story first seen 2 hours ago

The Verge · 2 hours ago

Nvidia has confirmed it will acquire Hugging Face for $12.93 billion, ending weeks of speculation and bringing one of the most widely used platforms for hosting open-source AI models, datasets and tools under the ownership of the world's biggest AI chipmaker. Chief executive Jensen Huang said the platform would "remain an open platform for the entire AI ecosystem," pledging that developers would still be free to choose their own models, frameworks, clouds and computing platforms, with no requirement to use Nvidia hardware. The deal gives Nvidia a strategic foothold in open-source AI at a time when rivals including OpenAI, Anthropic and Google are developing their own chips, helping the company defend its dominance in AI hardware.

Founded in 2016 and often dubbed the "GitHub for AI", Hugging Face was last officially valued at $4.5 billion after a 2023 funding round in which Nvidia itself invested; the Financial Times has previously reported that the startup rejected a $500 million Nvidia investment last year that would have valued it at $7 billion, reportedly over concerns about having one dominant investor. Reports of talks first surfaced on 23 August, with Business Insider and The Information giving differing accounts of the deal's progress before Nvidia's official confirmation. Hugging Face is said to generate only around $150 million in annualised revenue, a small fraction of the purchase price, though Nvidia's position as the world's most valuable public company gives it substantial resources to pursue the acquisition.

Nvidia confirms $12.93bn purchase of open-source AI hub Hugging Face
Huang vows platform stays open, no Nvidia hardware requirement
Deal dwarfs Hugging Face's $150m revenue and past $4.5bn valuation

New here? Start with this

Nvidia is buying Hugging Face, a company founded in 2016 that has become one of the most popular online hubs for open-source artificial intelligence models, datasets and tools. It is often nicknamed the "GitHub for AI" because developers use it to share and build on each other's work freely, rather than relying on closed, proprietary systems. Nvidia itself is the world's biggest maker of the specialised chips that power AI, and is currently the most valuable public company in the world.

The deal matters because it links Nvidia's hardware dominance with a central meeting point for the open-source AI community, at a moment when several of Nvidia's own customers, including OpenAI, Anthropic and Google, are trying to design their own AI chips and reduce their reliance on Nvidia. By owning Hugging Face, Nvidia gains closer ties to the developers shaping the next generation of AI models. Nvidia had previously invested in Hugging Face in 2023 and had also tried, unsuccessfully, to make a smaller investment before that.

The price being discussed, close to $13 billion, is far higher than Hugging Face's most recent valuation of $4.5 billion and well beyond its reported annual revenue of around $150 million, making the scale of the offer notable in itself.

Both sides, in good faith

The strongest fair case each way — we don't pick a winner.

The case for

Supporters of the deal would argue that Hugging Face gains the capital, stability and computing resources needed to keep growing a platform that has always operated on thin revenues relative to its enormous usage, and that Nvidia has been a credible, long-standing participant in this ecosystem rather than an opportunistic outsider, having invested in the 2023 funding round. They would see this as the market simply recognising the strategic value of open-source AI infrastructure, with deeper backing potentially accelerating tools, hosting capacity and community resources that benefit developers well beyond Nvidia's own interests.

The case against

Sceptics would counter that folding one of the AI world's most widely used, ostensibly neutral open-source hubs into the industry's dominant chipmaker risks compromising its independence, particularly for rival labs and hardware makers who rely on it and may now fear favouritism or reduced impartiality. They would point out that Hugging Face itself previously declined a far smaller investment from Nvidia specifically over concerns about having one dominant investor, suggesting the underlying worry about concentrated influence is one the company itself once took seriously, even as the eventual price proved difficult to refuse.

Full account

Nvidia has agreed to acquire Hugging Face, the widely used repository for open-source artificial intelligence models and datasets, in a deal worth $12.93 billion (about £10.2 billion). The takeover hands the world's dominant AI chipmaker ownership of a platform often likened to "GitHub for AI", where developers and researchers publish and share machine-learning models, tools and data with the wider community. Announcing the transaction, Nvidia chief executive Jensen Huang said the companies would work together to expand Hugging Face's infrastructure and broaden access to AI, while insisting the platform would remain open to the whole ecosystem rather than being tied to Nvidia's own products.

Founded in 2016, Hugging Face has grown into a central hub for the open-source AI community, particularly since the generative AI boom took off. It says it now hosts more than three million models and serves upwards of 18 million users, with over 200,000 companies drawing on models built there for commercial use. Nvidia was already a significant presence on the platform before the acquisition, having contributed hundreds of its own models and open datasets, and had previously invested in Hugging Face as part of a funding round in 2023 that valued the startup at $4.5 billion.

The purchase price marks a steep jump from that earlier valuation and reflects how central open-source AI infrastructure has become to the wider industry. It follows a reported approach last year in which Hugging Face turned down a $500 million investment from Nvidia that would have valued the company at $7 billion, said to stem from reluctance to rely too heavily on a single major investor. News of a possible sale first surfaced in late August, with early reports describing talks with a bank over options for a roughly $13 billion deal, before further accounts suggested a price close to the figure ultimately confirmed. Notably, Hugging Face's own revenue is comparatively modest, estimated at around $150 million a year, underlining that the acquisition is being driven chiefly by strategic positioning rather than existing earnings.

Commentators frame the move as part of a broader contest over the future shape of AI: with rival developers such as OpenAI, Anthropic and Google increasingly pursuing their own chip capabilities, Nvidia's purchase secures it a foothold in open-source software just as closed AI providers push into hardware. Huang has pointed to his own recent public backing of open-source AI and pledged that developers will remain free to choose whatever models, frameworks, clouds or computing platforms they prefer, without any obligation to use Nvidia's own compute. The deal, which remains subject to regulatory approval, would leave Nvidia with an unusually commanding position spanning both AI chips and open AI software—a combination some observers suggest could invite closer competition scrutiny.

Where outlets differ

Source 1 (The Verge) dwells on the deal's financial backstory: the sequence of press leaks from Business Insider and The Information through late August, Hugging Face's modest ~$150m revenue relative to the price paid, its $4.5bn 2023 valuation, and the Financial Times' account of Hugging Face rejecting an earlier $500m/$7bn Nvidia offer over fears of investor dominance.

Source 2 (likely a tech/business outlet) emphasises market-structure implications, describing the deal as letting Nvidia 'corner the market' in both AI hardware and software and explicitly flagging that it is subject to regulatory approval.

Source 2 includes platform usage statistics (18m+ users, 3m models, 200,000+ companies) and details Nvidia's existing contributions to the platform (500+ models, 250+ datasets) plus its full list of other investors (Google, Amazon, Salesforce, AMD, Intel, IBM, Qualcomm) — none of which appear in Source 1.

Source 2 alone cites Hugging Face CEO Clément Delangue's recent comment that China is winning the AI race through its embrace of open models, and references Huang's own open-source advocacy letter.

Source 1 places more weight on the geopolitical/competitive rationale — Nvidia protecting its hardware dominance as closed AI labs (OpenAI, Anthropic, Google) begin developing their own chips.

More coverage

AI Software Space Technology

Read the full article at the source →

Originally published by The Verge as “Nvidia is buying Hugging Face for almost $13 billion”.