One channel partner says Nvidia’s acquisition could boost AI infrastructure sales with enterprises because the steep costs of closed frontier models are prompting such customers to consider open models, many of which are hosted on Hugging Face, as an alternative.
Nvidia’s $12.9 billion blockbuster deal to acquire open model repository Hugging Face has the potential to help the AI infrastructure giant boost enterprise AI adoption and grease the wheels for its robotics business, channel partners told CRN.
In announcing the agreement Thursday, the Santa Clara, Calif.-based company vowed to invest in Hugging Face’s expansion and maintain its status as an open platform that can support any hardware, including those of Nvidia’s competitors.
[Related: Open Models, Harnesses Key To Making AI-Powered Security ‘Sustainable’: CrowdStrike Partners]
The deal is expected to close in the first half of 2027, pending regulatory approval.
In a Thursday blog post, Nvidia CEO Jensen Huang highlighted his company’s years of commitment to the cause of open models and framed the acquisition as a way to expand the benefits of AI to a broader constituency of customers.
“That is how AI can advance safely, strengthen cybersecurity and sovereignty, accelerate innovation, and reach factories, hospitals, farms, classrooms and Main Street businesses around the world,” he wrote.
Nvidia did not address a question posed by CRN in a media briefing about the partner opportunities that the Hugging Face acquisition could enable.
Why Hugging Face Could Help With Enterprise AI Adoption
One area where partners see a lot of potential with the Hugging Face acquisition is enterprise AI adoption, something that Nvidia has spent the past few years discussing as the next stage of growth but what partners say is still in the early innings.
The deal could help Nvidia boost AI infrastructure sales with enterprises because the steep costs of closed models from frontier AI labs like OpenAI amid other issues are prompting such customers to consider open models as an alternative, according to Andy Lin, vice president of strategy and CTO of Houston-based Mark III Systems.
“They’re looking for alternatives to closed frontier models because [such models] are easier to get started and easier a lot of times to use, but obviously they have their drawbacks as far as cost, questions about who owns the data, and outsourcing your business model in some ways. We see that most, if not all large enterprises want some kind of centralized AI factory that runs open models as a 1B option,” he said.
Lin, whose systems integration business has won multiple Nvidia Partner Network awards, said he could imagine Nvidia integrating Hugging Face into its software platforms to improve the way data centers are operated as so-called “AI factories,” in which a plethora of users share compute resources to train and run models.
The operationalization of AI factories has remained a challenge for many enterprises, an issue that has been a top focus for Mark III, according to Lin. And he thinks the Hugging Face acquisition could help further with this cause.
“Hugging Face [and] Nvidia’s NGC image repository are two centralized hubs that will allow these open models to be absorbed into these factories to try to create as frictionless an experience as possible for these enterprises,” Lin said.
While much of Nvidia’s recent revenue growth has come from hyperscalers and neoclouds, the executive said he sees enterprise customers as the next frontier.
“The only way to control your own destiny is to have your own AI factory as an enterprise, and I think that’s something that is starting to dawn on a lot of companies that have relied exclusively on the cloud,” he said.
Hugging Face Move Aligns With Recent Nvidia Sales Hires
The idea that Nvidia could use the Hugging Face acquisition to boost enterprise adoption resonated with C.R. Howdyshell, CEO of Myriad360-owned Advizex in Independence, Ohio, another Nvidia Partner Network award winner.
“Now they have the ability to really execute at speed to really drive adoption,” he said.
Howdyshell also said the acquisition’s enterprise implications line up neatly with Nvidia’s recent hirings of two executives with relevant commercial sales experience: HPE veteran Monica Gille as vice president of global partnerships and Microsoft veteran Nick Parker as executive vice president of worldwide field operations.
“The enterprise space has been slower to adopt, and that’s where they want to be,” he said.
Partner Sees Potential Issue With Prevalence Of Chinese Models
There could be a hitch, however, in the plan to use Hugging Face to boost enterprise AI adoption, according to a director-level employee at a large systems integrator with a U.S. presence.
The director, who asked not to be identified to speak candidly, pointed to the prevalence of open models developed by Chinese developers on the platform, which has become a flashpoint in U.S. discussions about AI security risks and competitiveness.
“Most of them don’t want to dip their toes widely into adopting a model created in China, only to get halfway through a [proof of concept] and end up being shut down or the board gets wind of it or anything else,” he said.
A Potential Boost To Nvidia’s Physical AI Ambitions
Christopher Cyr, CTO of North Sioux City, S.D.-based Sterling Computers, said he thinks enterprises may end up taking a hybrid approach in which they use open models to “do a majority of the work” and let the frontier models “do the heavy lifting.”
“If my consumption is 40 million tokens, wouldn’t it make more sense for me to use 30 million tokens locally on an open model and then use the other 10 [million on] a foundational model?” said Cyr, whose firm won an Nvidia Partner Network award in 2024.
The other big possibility Cyr sees with the Hugging Face acquisition is for Nvidia to tap into the repository’s growing library of robotics models to boost its physical AI business, which is all about using autonomous systems to perform a variety of physical tasks.
The systems integration executive said the process of training a robot in a digital environment and then porting that into a real-world automaton remains a “very complex process,” even with all the tools Nvidia provides.
“I still think there’s a lot of work that needs to be done, and I think that having Hugging Face as a front end for a lot of that will probably make it easier for Nvidia,” he said.
