Nvidia acquires Hugging Face and turns into open platform supplier

Nvidia CEO Jensen Huang introduced on his weblog that the corporate can be buying Hugging Face for $12.9bn.

He reassured synthetic intelligence (AI) builders of open fashions and their customers that Hugging Face would stay an open platform for your entire AI ecosystem.

“Builders will select the fashions they need, the frameworks they need, the clouds and inference service suppliers they need and the computing platforms they need,” Huang mentioned. He additionally confirmed that the corporate wouldn’t make Nvidia {hardware} a requirement for constructing or deploying via Hugging Face.

Commenting on the acquisition, Forrester principal analyst Naveen Chhabra mentioned: “Nvidia positive aspects visibility into buyer’s preferences and the AI fashions they use. They’ll see which fashions are trending, what datasets clients are downloading and the architectures which might be gaining traction weeks earlier than they hit mainstream tech information. It turns Hugging Face into a company scouting floor and an unparalleled market intelligence machine.”

Chhabra mentioned that via the acquisition of Hugging Face, Nvidia additionally positive aspects data-driven foresight into developer habits, which offers a roadmap to the place next-generation chip architectures want to move.

Huang claimed Hugging Face would proceed to assist multicloud and multiaccelerator improvement and deployment, enabling mannequin builders and the customers of these fashions to make use of the {hardware} and infrastructure that greatest match their work.

Forrester’s vice-president principal analyst Jeff Pollard mentioned: “By buying Hugging Face, Nvidia positive aspects management over and visibility right into a key AI provide chain layer. The corporate now sits nearer to your entire AI provide chain: compute, fashions, datasets and developer workflows. That creates a possibility to enhance AI safety at scale via higher mannequin transparency, provenance, vulnerability disclosure and guardrails. It additionally concentrates affect over a vital piece of the open AI ecosystem.”

Larger openness is likely one of the important focus factors for Huang. In July, Nvidia – together with a number of different firms – introduced the industry-wide Open Safe AIIiance. On the time, Nvidia described open fashions and open harnesses as “important” as they democratise defensive capabilities and enhance transparency for these tasked with defending IT programs.

Particularly, Nvidia sees open fashions as a strategy to allow cyber defence whereas complementing frontier closed fashions with customisable, localised controls. On the time, Nvidia mentioned: “Open supply allows massively distributed community-driven and self-controlled defence – with no single level of failure.”

Within the weblog, Huang mentioned: “Open weights broaden entry to AI and assist to make sure that AI management is distributed throughout firms, establishments and communities.”

Asserting the Hugging Face acquisition, Huang added that open fashions allow startups, companies, universities and public establishments to  construct on superior capabilities with out coaching each mannequin from scratch: “They allow organisations to match the fitting mannequin to the fitting job. That’s how AI can advance safely, strengthen cyber safety and sovereignty, speed up innovation; and attain factories, hospitals, farms, lecture rooms and Fundamental Road companies world wide.”

Eugene Cheah, CEO and co-founder of serverless inference platform Featherless.ai, described the acquisition as a “GitHub second”, referring to the $7.5bn acquisition of GitHub by Microsoft in 2018.

“This deal is an extremely necessary milestone for the AI {industry},” mentioned Cheah. “In the identical manner that GitHub grew to become a vital a part of the open supply ecosystem underneath Microsoft, Hugging Face is now ideally positioned to grow to be its counterpart for open supply AI. In brief, Hugging Face is now poised to be on the centre of the nexus of AI open supply improvement.”