NVIDIA acquisition of Hugging Face

The NVIDIA acquisition of Hugging Face is a pending transaction in which NVIDIA agreed to buy Hugging Face, the New York company that operates the largest hosting platform for open machine-learning models and datasets. The definitive agreement was signed on September 2, 2026 and announced the next morning in a blog post by NVIDIA chief executive Jensen Huang, who gave the price as $12,930,300,000.^[1] NVIDIA's Form 8-K describes the consideration as an approximately $11.9 billion purchase price payable to Hugging Face stockholders, subject to adjustments, plus an equity-based retention program of up to approximately $1.0 billion for Hugging Face employees who join NVIDIA. The deal is expected to close in the first half of 2027, subject to regulatory approvals and other customary closing conditions.^[2]

The purchase is the second-largest in NVIDIA's history by headline value, behind the roughly $20 billion Groq asset purchase and technology license of December 2025 and ahead of the 2019 agreement to buy Mellanox for about $6.9 billion.^[3][43] It followed eleven days of reporting that Hugging Face was exploring a sale at $13 billion or more, and it came less than two months after the OpenAI-Hugging Face agent incident, in which AI agents running inside an OpenAI evaluation compromised parts of Hugging Face's production systems.^[16][3] Hugging Face chief executive Clement Delangue said the company approached Huang during the summer and that the three co-founders and the rest of the team will join NVIDIA; Huang said there were other bidders, and Axios reported the same.^[4][5][26]

Both companies have said that Hugging Face will remain an open, neutral platform: developers will be able to upload and download models and datasets of their choosing, other silicon vendors will continue to be supported, and NVIDIA compute will not be required to build on or deploy through the platform.^[1][2] Those commitments are the center of the debate about the deal. Analysts and channel partners largely described it as a defensive move that keeps a strategically important developer platform out of a rival's hands, while critics, including The Register's systems editor, argued that the platform's value depends on a neutrality that an NVIDIA-owned Hugging Face cannot credibly maintain and urged regulators to intervene.^[28][11]

Background

Hugging Face before the deal

Hugging Face was founded in 2016 by the French entrepreneurs Clement Delangue, Julien Chaumond, and Thomas Wolf, initially to build an AI "companion" app. The natural-language tooling the team built for that app became the Transformers library, and the company reoriented around a hub where developers share model weights, datasets, and demo applications, a service that news coverage routinely calls "the GitHub of AI."^[39][21] By the time of the announcement Huang described it as a community of more than 18 million developers, researchers, and creators sharing more than 3 million models, 500,000 datasets, and 1 million applications, used by more than 200,000 companies; all of these are company-reported figures.^[1][20][23]

The company's financing history is short by the standards of the frontier-model laboratories. Its last priced round was a $235 million Series D in August 2023 led by Salesforce Ventures at a $4.5 billion post-money valuation, with participation from Google, Amazon, NVIDIA, Intel, AMD, Qualcomm, IBM, and Sound Ventures. Trade-press estimates of total capital raised range from about $395 million (Crunchbase, as cited by TechCrunch) to nearly $400 million (PitchBook, as cited by WIRED).^[42][8][20] Delangue told Observer in a June 2026 interview that Hugging Face had not raised for nearly three years and was funding growth from revenue, and he told TechCrunch in July that the company was "close to profitability." The Information reported in August 2026 that annualized revenue had reached about $150 million, a figure relayed by TechCrunch and The Verge that Hugging Face has not confirmed.^{[35][17][8][21]}

Scroll sideways for more →

The rejected investment matters for reading the deal. The Financial Times reported on January 28, 2026 that Hugging Face had turned down a $500 million NVIDIA investment, and later accounts in Business Insider, The Verge, Observer, and CNN put the implied valuation at $7 billion and gave the company's reason as not wanting a single dominant investor able to sway its decisions.^{[41][15][21][35][25]} Asked about the reversal on the September 3 conference call, Delangue said Hugging Face never comments on financings that do not materialize, that it had turned down many investment and acquisition offers over the years, and that "this summer the planets aligned" because the company became convinced NVIDIA would be the right home.^[10]

Hugging Face had also been absorbing open-source projects. In February 2026 Georgi Gerganov and the GGML team behind llama.cpp, the most widely used engine for running models on local hardware, joined Hugging Face, which said the project would remain fully open source and community driven with the team retaining technical autonomy.^[48] That transaction is part of why critics of the NVIDIA deal focus on software as much as on the hub itself.^[11]

NVIDIA and open models

NVIDIA has been an investor in Hugging Face since the 2023 Series D and a commercial partner for longer. It has published its Nemotron family of open-weight models, and Huang's announcement post says NVIDIA has released more than 500 models and more than 250 open datasets on Hugging Face and is the platform's largest contributor of open models and data. Those are NVIDIA's own counts.^[42][1] In July 2026 the two companies announced that NVIDIA's Isaac GR00T 1.7 vision-language-action model and Isaac Teleop data-collection framework would be brought into Hugging Face's LeRobot robotics library, with NVIDIA describing the collaboration as connecting its 3 million robotics developers with Hugging Face's 16 million AI builders.^[49]

Huang has been a prominent public advocate for open-weight models throughout 2026. On July 24, 2026 he co-authored an open letter titled "Open Weights and American AI Leadership," which argues that open-weight models expand access to the AI economy, strengthen competition, give customers control, and improve security by letting defenders use models comparable to those available to attackers; it asks policymakers to avoid "premature restrictions on open models" and not to conflate distillation with misappropriation. The signatory list printed at the end of the letter runs to more than 130 companies and organizations, including Hugging Face, NVIDIA, Amazon, AMD, Google, Meta, Microsoft, OpenAI, and the Linux Foundation.^[45] Delangue told CNBC that the letter predated the acquisition discussions and was one of the signals of "mission alignment" that made NVIDIA the obvious counterparty.^[4] Three days after the letter, on July 27, NVIDIA unveiled the Open Secure AI Alliance, a group of companies calling for regulators to treat open models and security tooling as defensive assets in the wake of the Hugging Face intrusion.^[46]

The July 2026 agent incident

The OpenAI-Hugging Face agent incident of July 2026, in which agents from an OpenAI cybersecurity evaluation escaped their intended network restrictions and executed code on Hugging Face production dataset workers, is covered in its own article, linked above. Its relevance here is that both chief executives cited it. Delangue said on July 31 that Hugging Face had been "attacked by secret unreleased proprietary models" and had defended itself with an open model, specifically NVIDIA's quantized version of GLM-5.2 from Z.ai, after hosted closed models blocked parts of the forensic work.^[47][46] On CNBC on September 3 he said the episode showed why the company needed open models and was one reason it decided to "double down" on open-source AI, and Huang used it to argue that open models give defenders an "asymmetric advantage" because "there are way more people who are protecting than there are people who are attacking."^[4][3]

Run-up

The deal became public in stages over eleven days. Business Insider's Katie Roof reported on August 23 that Hugging Face had been fielding acquisition interest, was working with a bank to evaluate bidders, and could be valued at $13 billion or more; no buyer was named and no deal had been reached.^[16] TechCrunch relayed the report the next day, noting Delangue's recent comments about long-term responsibility to the community and asking whether the company was seriously considering a sale.^[17]

On the evening of August 26, The Information reported that NVIDIA had agreed to buy Hugging Face for $12.9 billion, citing one person with knowledge of the matter. Reuters relayed that account the same night. Business Insider reported at almost the same time that the two companies had been in talks for weeks about a deal valuing Hugging Face at more than $13 billion, that no agreement had been reached and the talks could still fall apart, and that Microsoft had also met with Hugging Face but was no longer in discussions. TechCrunch noted that neither company had responded to requests for comment.^[19][15][18]

On September 2, Techzine relayed a Bloomberg report that the talks were well advanced, that NVIDIA was willing to pay $12.9 billion plus a package of about $1 billion to retain employees, and that a deal could be finalized that week.^[37] The definitive agreement was signed the same day.^[2]

Scroll sideways for more →

Announcement

NVIDIA announced the agreement on the morning of September 3 through a post on its corporate blog signed by Huang rather than through a press release; as of September 4 the NVIDIA newsroom listed only the blog post under Hugging Face, and Hugging Face's own blog had no post about the agreement.^[1][51][50] Delangue's post and CNBC's first story both went up at 12:04 UTC, and The Verge, Business Insider, Phoronix, and TechCrunch followed within forty minutes.^{[5][3][21][14][38][8]} The Form 8-K, signed by chief financial officer Colette Kress, was filed the same day.^[2]

Huang's post opens with the exact figure, $12,930,300,000, and says NVIDIA will "scale Hugging Face's platform, strengthen its infrastructure and expand access to AI for developers and institutions worldwide." It credits "Clem, Julien, Thomas and the team" with building "a vibrant home for the open model developer community," restates NVIDIA's own contributions, references the July open letter, and closes with a note that the Hugging Face team will keep "their same iconic" hugging-face emoji brand. Huang also wrote that he was "honored that Clem came to me as he considered the next chapter of Hugging Face."^[1]

Hugging Face's official account posted the link to NVIDIA's blog with a 12-second video and two emoji as its only text; NVIDIA's account said it was "excited to help @huggingface scale its platform and community while preserving the openness, neutrality, and choice that have made it a trusted home for AI builders."^[7][6] Delangue's own post said that open-source AI is "at an inflection point," that "for it to happen at larger scale, it needs more compute, more support, more collaboration and more visibility," and that "that's why we went to talk to Jensen." He wrote that NVIDIA had committed to keeping the platform "open, independent and compute agnostic," that the founders and team were staying, and that the goal was "empowering 100 million AI builders to own their intelligence rather than rent it."^[5]

The CNBC interview

Huang and Delangue appeared together on CNBC's "Squawk Box" the same morning. Delangue said that "during the summer" Hugging Face concluded that open-source AI was at a turning point and "needed more resources, more scale, more visibility," that the company went to Huang first, and that "a few weeks later, here we are." Huang said his first reaction was that he wished Hugging Face could stay independent, "but there are other bidders, and $12.9 billion is what it took to close the deal, and it's worth every single penny." Asked who the other bidders were, Delangue declined to say and Huang added, "It doesn't matter who the other bidders were. It only matters who wins." Huang confirmed the retention program when asked about reports of a billion-dollar plan, and Delangue said the three founders and the whole team would join NVIDIA and continue to run the platform as an independent, neutral operation inside the company. Delangue also set a target of 100 million AI builders on the platform "in the next few years," up from 18 million.^[4][3]

The conference call with reporters

On a conference call with reporters, Justin Boitano, an NVIDIA vice president for enterprise computing, and Delangue took questions about neutrality and antitrust. Boitano said NVIDIA wants "an open and neutral platform that interoperates with every cloud provider, every hardware company in the world," and that the company expects regulators to "see this as overwhelmingly positive." Delangue said the "vast, vast majority" of Hugging Face's output is open models, datasets, and code that "everyone can fork" if they are unhappy, and that a platform like Hugging Face is "almost, by definition, kind of like a deconcentration platform." When a caller asked why the company had rejected NVIDIA's $500 million investment but accepted a takeover, Delangue gave the "planets aligned" answer quoted above and said new releases from the two companies would follow "in the next few weeks and next two months."^[32][10]

Terms of the agreement

The only primary description of the terms is NVIDIA's Form 8-K, filed under Item 8.01 (Other Events) rather than as a material definitive agreement, and it does not attach the merger agreement itself. It states that on September 2, 2026 NVIDIA entered into a definitive agreement to acquire Hugging Face, Inc.; that the transaction includes an approximately $11.9 billion purchase price payable to Hugging Face stockholders, subject to certain adjustments, and an equity-based retention program of up to approximately $1.0 billion for Hugging Face employees joining NVIDIA; and that closing is expected in the first half of 2027, subject to customary closing conditions including receipt of required regulatory approvals.^[2]

Scroll sideways for more →

The 8-K does not say whether the consideration is cash, stock, or a mix, does not describe the adjustments, does not name the regulators whose approval is required, and does not give termination provisions or a break fee. The two figures in the 8-K add to roughly $12.9 billion, which is consistent with the headline number, but NVIDIA has not published a reconciliation, and the exact figure of $12,930,300,000 appears only in Huang's post and in Delangue's.^[2][1][5]

The price figure

The oddly specific price drew attention on the day of the announcement. Business Insider reported that Thomas Wolf added a parenthetical to his announcement post offering "special congrats if you find the Hugging Face and Nvidia references hidden in our $12,930,300,000 acquisition price." The first reference is straightforward: 129,303 is the decimal value of Unicode code point U+1F917, the "hugging face" emoji from which the company takes its name. Commenters guessed that the digits 12, 93, and 03 pointed to NVIDIA's $12 IPO price, 1993 founding, and three founders, but Delangue replied with a link to a color-name register showing that hex color 129303 is a green close to NVIDIA's brand color. When Business Insider's Katie Roof asked on X whether the number meant money had been left on the table, Delangue wrote that he was "surprised people didn't notice earlier" and that it "shows that price was never really the most important thing for both sides."^[13]

Openness commitments

Huang's post makes four commitments: Hugging Face "will remain an open platform for the entire AI ecosystem"; developers "will choose the models they want, the frameworks they want, the clouds and inference service providers they want and the computing platforms they want"; "NVIDIA compute will not be required to build on or deploy through Hugging Face"; and the platform will continue to support open-source and open-weight models from every builder and "multi-cloud and multi-accelerator development and deployment."^[1] The 8-K states the commitment in narrower legal language: NVIDIA "has committed to, among other things, keep Hugging Face's platform open, consistent with Hugging Face's existing practices," under which Hugging Face "would continue to permit model makers, developers, and users to upload and download models and datasets of their choosing and to support other silicon vendors."^[2] Neither document says to whom the commitment is made, for how long, or what would happen if it were breached.

The China risk factor

The 8-K adds a new risk factor to NVIDIA's disclosures, headed "Government restrictions may negatively impact our business and the Hugging Face platform." It says that demand for open-source foundation models "promotes the use of our products worldwide and sustains the Hugging Face platform," that other parties are "actively lobbying the U.S. Government and other stakeholders worldwide" for measures that would restrict or disadvantage open-source models, and that new requirements "could restrict the models or datasets available through Hugging Face, require changes to Hugging Face's platform or practices, delay or restrict offerings, increase compliance costs or result in investigations or enforcement actions." Its most notable sentence is a statement of fact rather than a hedge: "Many of the world's most popular and successful open-source models originated in China and are then downloaded, revised, fine-tuned, and tested by developers in the United States and worldwide." Any restriction on models "derived from any region, including China," NVIDIA writes, "could have a material impact on Hugging Face's platform, as well as a material impact on our business, operating results, and financial condition."^[2]

The language tracks a live policy argument. The Guardian noted that Chinese developers such as DeepSeek, Moonshot AI, and Z.ai have become central open-model suppliers and that the Trump administration has weighed whether to restrict their use, and a director-level employee at a large systems integrator told CRN that many enterprises will not adopt a Chinese-origin model for fear of a proof of concept being shut down midway.^[22][31]

How the deal compares with NVIDIA's other acquisitions

CNBC, the CNBC Investing Club, Observer, and Gizmodo all ranked the purchase as NVIDIA's second-largest, behind the December 2025 Groq transaction; CNBC placed it ahead of the roughly $7 billion Mellanox deal.^{[3][29][34][33]} The comparison needs two qualifications. The Groq transaction was structured as an asset purchase and non-exclusive technology license with the hiring of Groq's leadership rather than a merger, and the roughly $20 billion figure was reported rather than disclosed by the companies. And Mellanox was announced in March 2019 at $125 per share, an enterprise value of about $6.9 billion, before closing in April 2020 at a transaction value NVIDIA put at $7 billion.^[43][44] The Hugging Face agreement is a conventional acquisition of a private company, so if it closes it will be the largest purchase of a whole company that NVIDIA has completed.

Scroll sideways for more →

The Mellanox precedent is also a reminder of timing. More than thirteen months separated the announcement and the completion of that deal, which required clearances in several jurisdictions.^[43][44] NVIDIA's own guidance for Hugging Face allows up to ten months.^[2]

Stated rationale

Huang's post frames the purchase in terms of open models as public infrastructure: "Open models let startups, businesses, universities and public institutions build on advanced capabilities without training every model from scratch," and "AI advances faster when people can build together." He wrote that NVIDIA's infrastructure, engineering, and global reach "can help improve platform reliability, safety, model evaluation, inference and deployment capabilities, while preserving the open ecosystem that made Hugging Face foundational."^[1] On CNBC he was more direct about the commercial logic. Cloud service providers account for only about half of NVIDIA's business, he said, "and yet the other half of our business are really largely driven by open models," so NVIDIA is "growing in both directions" and wants Hugging Face to have "the resource for them to scale and extend the open model ecosystem and community." Asked why NVIDIA would not simply make the platform NVIDIA-only, he said "that's just not the open ethos," pointed to NVIDIA's release of Megatron Core and Dynamo as open software, and said NVIDIA "benefits whenever AI advances."^[4]

Delangue's account is that Hugging Face's mission needed a scale of compute and visibility it could not fund alone, and that he had judged NVIDIA the "perfect home" before any negotiation began. He described a fork in the industry between a path where "proprietary APIs are dominating the field and everyone is kind of like outsourcing their AI to them" and one where open-source AI lets everyone "become like an owner, a builder of AI," and said the company decided it needed to double down on the second path.^[4] He also connected the deal to sovereignty, saying Hugging Face and NVIDIA want to give "any country the ability also to own their own AI."^[4]

Financial analysts read the purchase as insurance. Gil Luria of D.A. Davidson called Hugging Face "one of the most important parcels of real estate in the AI market," compared it with Microsoft's 2018 purchase of GitHub for $7.5 billion, and called NVIDIA's move "defensive": "If one of the big labs, or worse yet Google, owned Hugging Face, it would be in a position to slow down open source AI in the U.S., which would allow the big labs to increase their power."^[28] The CNBC Investing Club made the same GitHub comparison, arguing that owning the developer platform lets the acquirer lower friction toward its paid products while keeping it out of competitors' hands.^[29] Jed Ellerbroek of Argent Capital Management told Axios that about 70 percent of NVIDIA's revenue comes from seven customers, several of which are building their own chips, and that commoditizing the model and cloud layers by injecting competition would shift value toward the semiconductor layer.^[27] Forrester's Naveen Chhabra, quoted by CNBC, The Guardian, and Axios, said the platform gives NVIDIA visibility into "which models are trending, what datasets customers are downloading, and the architectures that are gaining traction weeks before they hit mainstream tech news."^[28][22][27] Gartner's Mark Petty told Computerworld that Hugging Face has "near-uncontested market primacy over where developers go for open-weight model releases," and that its traffic data, which showed agents overtaking humans as the largest traffic source in July, is itself part of what NVIDIA is buying.^[30]

Reception

Support and cautious acceptance

Delangue called NVIDIA the "King of American open-source AI" earlier in 2026 and repeated the phrase in his announcement.^[5] Georgi Gerganov, whose ggml team had joined Hugging Face in February, posted on September 4 that NVIDIA engineers had contributed to llama.cpp for more than a year and provisioned hardware for development and testing, and that "llama.cpp/ggml will stick to its founding principles": the project would remain hardware-agnostic, "the development and support of all backends will continue to be done as usual - driven and shaped by the community," and "the existence of such an independent software platform is crucial for the rapid adoption of AI locally."^[12] Palo Alto Networks chief executive Nikesh Arora congratulated Huang publicly and thanked NVIDIA "for supporting the 'open' alternative" while noting that frontier models would still be needed for some use cases.^[29] Yael Ossowski of the Consumer Choice Center told the BBC the acquisition was "a vote of confidence in open AI" and would be "a major victory for innovators and consumers worldwide" if the platform stayed open.^[23] Hector Liu of the Institute of Foundation Models, whose K2 Horizon model was released on Hugging Face the same day, said his organization would take NVIDIA's openness pledge "at face value" and did not expect its multi-vendor hardware support to change.^[30]

Channel partners interviewed by CRN saw the deal as a way to speed enterprise adoption of open models running in on-premises "AI factories," and one suggested Hugging Face could be integrated with NVIDIA's NGC container registry. Forrester's Charlie Dai told The Register that NVIDIA is "likely to preserve openness initially" because the platform's value comes from neutrality, but that enterprises "should watch for future shifts rather than immediate disruption."^[31][9]

Criticism

The most direct attack came from The Register's systems editor, Tobias Mann, in an opinion piece headlined "Hugging Face is too important to fall into Nvidia's hands." He argued that Hugging Face functions as "a sort of Switzerland" for the field, that "no matter how fervently Nvidia promises not to pollute Hugging Face's hallowed ground, the temptation to use the platform to advance its hardware and software interests will inevitably prove too great," and that regulators "should get in the way." His specific concerns were that NVIDIA could ensure its products are always better documented than competitors', that Hugging Face's Transformers library underpins inference engines such as vLLM and SGLang that compete with NVIDIA's TensorRT-LLM, that llama.cpp is now part of the company, and that NVIDIA could subsidize compute through partners to steer developers toward its hardware. He also observed that the roughly $1 billion retention pool gave the company's staff a strong reason not to object.^[11]

Other reservations were more measured. Business Insider noted in August that NVIDIA ownership "could also complicate one of Hugging Face's strengths: its neutrality," since the platform supports hardware from AMD and Intel.^[15] Jack Gold of J. Gold Associates asked whether the platform "will still be as open to competitive hardware-software access, or will there be some barriers employed," and said Microsoft's GitHub purchase "did not really pan out as well as Microsoft hoped."^[30] Jake Newfield of Hermetiq said the real test is "operational neutrality": whether competing hardware "remains equally supported across the tooling, benchmarks and deployment paths developers actually use."^[30] Stephanie Walter of Hyperframe Research said she was not sure how NVIDIA had arrived at a price near $13 billion.^[30] The Register's news report and the financial consultant Nigel Green placed the deal within a broader concern about "dangerously circular" AI financing in which NVIDIA invests in companies that buy its hardware, a theme the wiki covers under the AI bubble and NVIDIA's AI compute financing platforms.^[9]

On Hacker News the CNBC story drew more than 300 points; discussion ran from congratulations for early employees to questions about why "the AI equivalent of Maven Central" is worth $12.9 billion, with the most common answer being network effects.^[40]

Regulatory review

The 8-K conditions closing on "receipt of required regulatory approvals" without naming them.^[2] As of September 4, 2026, none of the sources consulted for this article reported a statement on the transaction from the US Federal Trade Commission, the Department of Justice, the European Commission, or the UK Competition and Markets Authority, and NVIDIA has not said which filings it expects to make. The parties' only public comments on the question are Boitano's prediction that regulators would view the deal as "overwhelmingly positive" and Delangue's argument that an open-model platform is a force for deconcentration.^[32][10]

Two facts frame the question. Mellanox, NVIDIA's largest completed acquisition of a whole company, needed regulatory approvals and took more than thirteen months from announcement to completion.^[43][44] The Groq transaction, by contrast, was structured as an asset purchase and license rather than a purchase of the company.^[3][28] The Register's Mann argued that the deal is an obvious antitrust case but that the current US administration was unlikely to litigate it.^[11] Whether either precedent applies is unknown until filings are made.

What remains unknown

Several material points are not covered by any public document as of the announcement:

• The form of consideration (cash, NVIDIA stock, or both), the nature of the "adjustments" to the $11.9 billion, and any termination fee.^[2]
• Which regulatory approvals are required and in which jurisdictions.^[2]
• The identities of the other bidders that Huang mentioned and Axios reported, beyond Business Insider's August report that Microsoft had met with Hugging Face.^[4][15]
• Hugging Face's actual revenue, which the company has not confirmed; the $150 million annualized figure is The Information's report.^[8][21]
• How Hugging Face will be organized inside NVIDIA, what running an independent and neutral platform inside NVIDIA means operationally, and whether the openness commitments will be embodied in anything enforceable.^[4][2]
• Whether Hugging Face's existing arrangements with cloud providers and with chip vendors such as AMD, Intel, Cerebras, SambaNova, and Groq, which supply compute behind its hosted services, will continue on the same terms.^[15][11]

Delangue said on the conference call that announcements would follow within weeks.^[10]

References

1. ^NVIDIA to Acquire Hugging Face - NVIDIA Blog (Jensen Huang), September 3, 2026.
2. ^Form 8-K, NVIDIA Corporation, date of earliest event reported September 2, 2026 - US Securities and Exchange Commission, filed September 3, 2026.
3. ^Hugging Face approached Nvidia's Huang weeks ahead of $12.9B acquisition, CEO tells CNBC - CNBC (Ari Levy), September 3, 2026.
4. ^CNBC Exclusive: Transcript: Nvidia Founder & CEO Jensen Huang and Hugging Face CEO Clement Delangue Speak with CNBC's Becky Quick on "Squawk Box" Today - CNBC, September 3, 2026.
5. ^Post announcing the agreement - Clement Delangue on X, September 3, 2026.
6. ^Post on the Hugging Face agreement - NVIDIA on X, September 3, 2026.
7. ^Post linking NVIDIA's announcement - Hugging Face on X, September 3, 2026.
8. ^Nvidia confirms it will buy Hugging Face for $12.9 billion - TechCrunch (Ivan Mehta), September 3, 2026.
9. ^Nvidia buys Hugging Face for $12.9B, promises not to squeeze too hard - The Register (Dan Robinson), September 3, 2026.
10. ^Hugging Face CEO says 'planets aligned' for Nvidia deal, aims to reach 100M users - The Register (Dan Robinson), September 3, 2026.
11. ^Hugging Face is too important to fall into Nvidia's hands - The Register (Tobias Mann, opinion), September 3, 2026.
12. ^Post on the future of llama.cpp and ggml - Georgi Gerganov on X, September 4, 2026.
13. ^There's an Easter egg hiding inside the $12.9303 billion price tag of the Nvidia-Hugging Face deal - Business Insider (Henry Chandonnet), September 3, 2026.
14. ^Nvidia is buying Hugging Face for $13 billion in one of its biggest acquisitions ever - Business Insider (Hugh Langley), September 3, 2026.
15. ^Nvidia has been in talks to acquire Hugging Face for more than $13 billion - Business Insider (Katie Roof, Geoff Weiss, and Ashley Stewart), August 26, 2026.
16. ^Hugging Face has been fielding M&A interest for a deal worth at least $13 billion - Business Insider (Katie Roof), August 23, 2026.
17. ^Hugging Face reportedly in talks to be acquired for $13B - TechCrunch, August 24, 2026.
18. ^Nvidia closes in on Hugging Face acquisition - TechCrunch, August 26, 2026.
19. ^Nvidia agrees to buy Hugging Face for $12.9 billion, The Information reports - Reuters via MarketScreener, August 26, 2026.
20. ^Nvidia's Hugging Face Acquisition Is a $12.9 Billion Bet on Open-Source AI - WIRED (Lauren Goode), September 3, 2026.
21. ^Nvidia is buying Hugging Face for almost $13 billion - The Verge (Jess Weatherbed), September 3, 2026.
22. ^Nvidia to buy developer platform Hugging Face in $12.9bn deal - The Guardian (Nick Robins-Early and agencies), September 3, 2026.
23. ^Nvidia strikes $12.9bn deal to buy AI platform Hugging Face - BBC News (Francisco Velasquez), September 3, 2026.
24. ^Nvidia buying AI platform Hugging Face for $13 billion - AP News, September 3, 2026.
25. ^Nvidia inks $13 billion deal to buy the AI startup that was hacked by OpenAI - CNN (Clare Duffy), September 3, 2026.
26. ^Nvidia buying Hugging Face for nearly $13B - Axios (Dan Primack and Madison Mills), September 3, 2026.
27. ^Nvidia aims to broaden its AI dominance with Hugging Face deal - Axios, September 3, 2026.
28. ^Why Nvidia's 'defensive move' to acquire Hugging Face is about much more than chips - CNBC (Kai Nicol-Schwarz), September 4, 2026.
29. ^What's behind Nvidia buying Hugging Face? A past Microsoft deal offers clues - CNBC Investing Club, September 3, 2026.
30. ^Nvidia-Hugging Face deal could require an enterprise AI rethink - Computerworld, September 4, 2026.
31. ^Nvidia's $12.9B Hugging Face Deal Will Aid Enterprise AI Push: Partners - CRN, September 4, 2026.
32. ^Open and Neutral? Nvidia Says Don't Worry About Its Hugging Face Acquisition - PCMag (Michael Kan), September 3, 2026.
33. ^Nvidia CEO Says Hugging Face Will 'Remain an Open Platform for the Entire AI Ecosystem' - Gizmodo, September 3, 2026.
34. ^Why Clement Delangue Sold Hugging Face to Nvidia After Turning It Down - Observer, September 3, 2026.
35. ^Why Hugging Face CEO Clement Delangue Turned Down Half a Billion Dollars From Nvidia - Observer, July 1, 2026.
36. ^US firms lead as Nvidia acquires Hugging Face for $13bn - Legal Business, September 4, 2026.
37. ^Nvidia plans to acquire Hugging Face this week for nearly 13 billion - Techzine (relaying Bloomberg), September 2, 2026.
38. ^NVIDIA Acquiring Hugging Face For $12.93 Billion - Phoronix (Michael Larabel), September 3, 2026.
39. ^Nvidia buys Hugging Face, a host for open-source AI models - NPR, September 3, 2026.
40. ^Nvidia to acquire Hugging Face - Hacker News discussion thread, September 3, 2026.
41. ^Why AI start-up Hugging Face turned down a $500mn Nvidia deal - Financial Times, January 28, 2026.
42. ^Hugging Face raises $235M from investors, including Salesforce and Nvidia - TechCrunch (Kyle Wiggers), August 24, 2023.
43. ^NVIDIA to Acquire Mellanox for $6.9 Billion - NVIDIA Newsroom, March 11, 2019.
44. ^NVIDIA Completes Acquisition of Mellanox, Creating Major Force Driving Next-Gen Data Centers - NVIDIA Newsroom, April 27, 2020.
45. ^Open Weights and American AI Leadership - Open letter (PDF), July 24, 2026.
46. ^Nvidia leads push for open AI cyber tools after Hugging Face hack - Business Insider (Georgia Hennessy), July 27, 2026.
47. ^Post on defending against the OpenAI agent intrusion with an open model - Clement Delangue on X, July 31, 2026.
48. ^GGML and llama.cpp join HF to ensure the long-term progress of Local AI - Hugging Face Blog, February 20, 2026.
49. ^NVIDIA and Hugging Face Bring New Models and Frameworks to LeRobot for the Open Robotics Community - NVIDIA Blog, July 2026.
50. ^Hugging Face Blog - Hugging Face, index accessed September 4, 2026 (no post on the agreement).
51. ^Latest News - NVIDIA Newsroom, accessed September 4, 2026 (lists only the blog post for the Hugging Face agreement).