Nvidia makes staggering $12.9 billion bet to protect AI empire

Nvidia (NVDA) built its artificial intelligence empire by selling the chips that companies need to train and run increasingly powerful AI models.

Now CEO Jensen Huang is spending almost $13 billion to position Nvidia where millions of developers find, build, and deploy those models.

One of Nvidia’s largest acquisitions, Hugging Face cost $12.93 billion. Hugging Face shareholders receive $11.9 billion and up to $1 billion in equity-based incentives to retain employees. The deal should close in the first half of 2027, pending regulatory approval.

The price is sky high, but Nvidia isn’t purchasing another semiconductor startup.

Hugging Face operates an open AI platform used by more than 18 million developers and over 200,000 companies, giving Nvidia a direct connection to one of the world’s largest AI developer communities.

That might explain why Wall Street is not fleeing from the price tag.

Nvidia pays billions for access to AI’s developers

Hugging Face has emerged as a key marketplace and cooperation tool for open AI.

Nvidia says its customers share 3 million-plus models, 500,000 datasets and 1 million apps. With Hugging Face, organizations and developers are able to find, tweak, and deploy existing models instead of building each model from scratch.

That provides Nvidia something it can’t achieve from just selling more GPUs: closer proximity to the developers determining whose models and infrastructure they will employ.

The company is already the biggest source of open models and data on Hugging Face, with more than 500 models and 250 open datasets on the platform.

The deal might significantly enhance that connection.

But Nvidia is making a strange vow.

Hugging Face will remain accessible to rival processors, clouds, models, and frameworks, the business adds. No Nvidia hardware is needed to create or deploy using Hugging Face. The platform will continue to support other semiconductor suppliers, the SEC filing clearly indicates.

That’s important, since part of Hugging Face’s appeal is its neutrality.

Too much of a lean to Nvidia’s hardware might alienate the company from the developer community it just paid almost $13 billion to buy.

The broader potential may be more subtle: Keep Hugging Face open while making Nvidia’s tech easier and easier for those millions of developers to use.

Nvidia makes a $12.9 billion move that few investors saw coming.

Bloomberg / Getty Images

Wall Street backs Nvidia’s massive AI bet

At least one Wall Street business isn’t backing away from Nvidia following the transaction.

Needham reiterated its Buy rating and $300 price target on Nvidia following the Hugging Face announcement. The rating is notable because Nvidia is spending heavily to acquire a company outside its traditional chip business.

More AI:

Hugging Face was valued at $4.5 billion in 2023, according to Reuters, meaning the Nvidia deal values the company at almost three times that previous private valuation.

That premium implies that Nvidia sees strategic worth well beyond Hugging Face’s present financial profile.

It also comes as Nvidia confronts a long-term threat from some of its largest clients. Microsoft, Meta, and other tech firms are creating proprietary AI chips to save costs and lessen dependency on Nvidia GPUs. The deal might help Nvidia diversify its position as big clients create their own processors.

It also makes Hugging Face a viable hedge fund for you.

Nvidia would own a platform higher in the technology stack that touches millions of developers regardless of model, even if AI hardware becomes more competitive.

Nvidia’s $12.9 billion gamble goes far beyond chips

The real issue is, why does Nvidia need to possess Hugging Face in the first place?

Nvidia already dominates the market for strong AI accelerators. You acquire a hardware-agnostic development platform, and with that, you receive the risk of integration and the potential for customers to raise concerns about its neutrality.

But CEO Jensen Huang appears to be betting that the next phase of AI will not be won by merely building the fastest computers.

The open models allow corporations to customize AI without relying solely on proprietary systems from companies like OpenAI and Anthropic. And as those models improve, the platforms where developers locate and deploy them may become more powerful.

Hugging Face is home to the developer community working on open models, and the infrastructure and technical expertise from Nvidia may improve its reliability, model validation, inference, and deployment capabilities, Huang added.

The strategic reasoning creates an interesting conundrum. Nvidia says it’s acquiring Hugging Face for $12.93 billion and tells developers they won’t need to use Nvidia hardware.

If such openness keeps Hugging Face’s massive developer community intact, Nvidia may not have to force anybody to use its technology. It only has to be the most obvious and appealing alternative for those developers when they finally require processing power.

It’s what makes this transaction more than just another AI deal. Nvidia already has a prime place at the bottom of the AI stack with its GPUs.

Now Huang is spending almost $13 billion for a seat considerably closer to the developers that determine what gets built on top of them.