Multiple international media outlets have reported that NVIDIA is nearing an acquisition of the open-source AI platform Hugging Face, with a valuation of approximately $12.9 to $13 billion. However, as of now, neither party has announced a formal agreement, and negotiations remain subject to change.
Valuation has significantly increased compared to the previous round.
Hugging Face's last public funding round occurred in 2023, when the company raised $2.35 billion at a $4.5 billion valuation, with investors including Salesforce Ventures, GV (a subsidiary of Alphabet), IBM Ventures, and NVIDIA.
If this transaction proceeds at approximately $13 billion, it would signify a significant increase in the company’s valuation compared to its previous round. The report also noted that NVIDIA had previously proposed a $500 million investment at the end of 2025, valuing Hugging Face at around $7 billion, but the offer was not accepted at the time.
NVIDIA values the open-source model ecosystem
At the core of this potential acquisition is NVIDIA’s desire to solidify its dominant position in the AI infrastructure market. As major closed-source labs like OpenAI, Google, Amazon, and Anthropic advance their own custom chips, market reliance on NVIDIA’s hardware is declining.
Under this context, the importance of the open-source model ecosystem is growing. If more developers and enterprises continue to focus on open model training, deployment, and inference, demand for NVIDIA GPUs and related software stacks will remain strong. Hugging Face, as a community and distribution platform for open-source models, is positioned at the heart of this ecosystem.
This year, Hugging Face's management has publicly supported the open-model approach multiple times. CEO Clem Delangue has called on the U.S. government to support open models rather than restrict the spread of related technologies in public interviews.
Or use this as an opportunity to re-enter the cloud services market
Beyond the model ecosystem, cloud computing is also a key aspect of this transaction. The report notes that NVIDIA has previously scaled back its own cloud business, DGX Cloud, while Hugging Face already has the capability to help developers rent computing power and run models.
By integrating Hugging Face, NVIDIA can more quickly enter the AI cloud services and computing power distribution market without having to build an entire platform from scratch. This also helps it further connect chip sales, model hosting, and developer services.
Another practical consideration is compute capacity absorption. NVIDIA previously committed to supporting large-scale cloud compute agreements for its customers. If some customers fail to fully utilize their contracted compute resources, NVIDIA must find new recipients for this excess capacity. With Hugging Face, this idle compute capacity can now be resold to platform users.
Hugging Face's revenue is still small but growing rapidly.
Hugging Face's annualized revenue recently reached approximately $150 million, up from about $100 million two months ago, indicating rapid growth. The company's management also stated last month that the business is nearing profitability.
However, at a price nearing $13 billion, this remains a high-multiple acquisition. Such a valuation is uncommon for an AI platform with relatively limited revenue, highlighting the buyer’s strong emphasis on its position within the ecosystem.
Integration in the AI infrastructure sector has also accelerated recently. The report noted that payment company Stripe acquired model routing startup OpenRouter earlier this month, which had a valuation of approximately $1.3 billion during its May funding round.
If NVIDIA ultimately completes its acquisition of Hugging Face, the global AI industry's competitive focus may shift further from pure model capabilities to an integrated battle over chips, open-source ecosystems, cloud computing power, and developer access points.