Full story
Deal Talks, Then Agreement
Nvidia has been in talks to acquire Hugging Face, the AI platform for sharing and building with open-source models, with one report valuing the company at more than $13 billion and saying the talks could still fall apart.
“a deal that would value Hugging Face at more than $13 billion”
Business Insider reported that Nvidia and Hugging Face had acquisition conversations in recent weeks, and that Nvidia and Hugging Face did not respond to requests for comment.
TechCrunch later reported that Nvidia has agreed to buy Hugging Face for $12.9 billion, citing The Information and noting that neither Nvidia nor Hugging Face had responded to comment.
The Information’s reported price marked a jump from Hugging Face’s last known valuation, after the platform raised $235 million in 2023 that valued it at $4.5 billion.
TechCrunch also tied the potential purchase to Nvidia’s efforts to protect its position as major closed AI labs build their own AI chips, while open-weight ecosystems give customers alternatives to those labs’ proprietary systems.
Open-Model Stakes and Neutrality
Hugging Face sits at the center of the open-source AI ecosystem, hosting millions of AI models and datasets, and owning the platform could give Nvidia a bigger foothold with developers and potentially drive more workloads onto its chips.
Business Insider warned that Nvidia ownership could complicate one of Hugging Face’s strengths: its neutrality, because the platform supports models and hardware from across the industry, including Nvidia competitors such as AMD and Intel.
TechCrunch described Hugging Face as a “kind of GitHub for the AI era” and said the deal would put Nvidia closer to the developer shelf for open-weight models and benchmarks.
Fierce Network quoted Tekonyx Founder and Chief Research Officer Sid Nag saying the rumored deal is about Nvidia gaining control of a key distribution layer, adding that it could be “potentially negative for ecosystem neutrality.”
Fierce Network also quoted Moor Insights and Strategy CEO and Chief Analyst Patrick Moorhead saying it could be “a way for Nvidia to get more closely aligned with developers,” while Jason Andersen said “There are about 3 MILLION models up there right now” and raised the question of how Nvidia would handle third-party models and data sets.
Who Benefits, Who Watches
The reporting tied the acquisition to a broader wave of big bets on open-weight infrastructure, including Nvidia’s $6 billion agreement with Poolside and Stripe’s $7 billion acquisition of OpenRouter, both framed as ways to reduce dependence on major hyperscalers and frontier labs.
“only 6% of companies use open-weight models”
TechCrunch said open-weight adoption is still limited, citing a survey where only 6% of companies use open-weight models and a separate figure where just 2% of software engineers measured by Jellyfish use them.
TechCrunch also quoted Jellyfish’s Nik Albarran saying open-weight models are primarily used for high-volume, repetitive inference workloads like customer service chats, where tuning can reduce costs.
For Nvidia’s side, Fierce Network reported that Jason Andersen said Nvidia has long contributed to Hugging Face by posting its open-weight models and other testing tools, and that “My guess is that Huggingface is where NVIDIA sees the bulk of its downloads and developer interaction.”
For the open-model ecosystem, Fierce Network captured the concern that an important neutral marketplace could become an Nvidia-centered distribution channel, even as supporters argued the deal could accelerate open-model adoption by bringing more capital, compute capacity, and enterprise support.