Nvidia Bets $12.9 Billion on the People Building AI.

Nvidia Bets $12.9 Billion on the People Building AI.

Nvidia is spending $12.9 billion to buy Hugging Face, one of the most important platforms in the open-AI movement, in a deal that could give the world's biggest AI chipmaker a much bigger foothold in the software and developer ecosystem built around its chips.

Nvidia announced the agreement on Thursday, putting $12.93 billion on the table for the New York-based company whose platform has become a major meeting point for developers, researchers and companies building artificial intelligence systems.

Hugging Face is not an AI chatbot in the mould of ChatGPT. It is closer to a giant workshop for the people making the technology behind such systems. Developers use the platform to find, share, test and modify AI models, datasets and applications.

More than 18 million developers, researchers and creators use Hugging Face, according to Nvidia. The platform hosts more than three million models, 500,000 datasets and one million applications, while more than 200,000 companies use its tools; making the acquisition about considerably more than adding another technology company to Nvidia's portfolio.

For years, Nvidia has dominated the hardware side of the AI revolution. Its graphics processing units, or GPUs, have become the workhorses used to train and run many of the world's most advanced AI systems.

Hugging Face gives Nvidia something different: a direct relationship with a huge community of people deciding what gets built and how it gets built.
Analysts suggest it could become increasingly valuable as the AI industry moves beyond the initial race to build ever-larger models.

OpenAI, Microsoft, Meta and other major technology companies are developing their own AI chips, partly to reduce their dependence on Nvidia's expensive processors. Nvidia is responding by pushing further into the software and models that sit on top of the hardware.

Hugging Face is particularly important because of its commitment to open-source and open-weight AI models.

Unlike proprietary systems whose underlying technology is largely controlled by their creators, open models can be downloaded, adapted and deployed by developers and organisations under their respective licences. That has made them increasingly attractive to businesses and researchers looking for cheaper or more customisable alternatives to closed AI systems.

Nvidia says that openness will not change after the acquisition.

CEO Jensen Huang said Hugging Face would remain open to the wider AI ecosystem, with developers free to choose their models, frameworks, cloud providers and computing hardware. Nvidia's chips, he said, will not be required to build or deploy through the platform.

Whether developers continue to see it that way may become one of the most important questions surrounding the deal.

Nvidia is not a neutral observer in the AI hardware market. It sells the computing infrastructure on which much of the industry runs, while Hugging Face brings together developers who can influence which models, tools and hardware become widely adopted.

The company already has a substantial presence on the platform. Nvidia says it has released more than 500 models and more than 250 open datasets on Hugging Face.

The acquisition therefore gives Nvidia a stronger position at several layers of the AI stack - from the chips used to power AI systems to the models and developer community using them.

It is also an expensive bet.

Hugging Face was valued at about $4.5 billion in 2023, meaning Nvidia is paying almost three times that valuation less than three years later. The agreement includes about $11.9 billion for Hugging Face investors and up to $1 billion in equity-based incentives for employees joining Nvidia.

For Africa's growing AI community, the significance is less about the size of the cheque than what happens to the open ecosystem that developers increasingly rely on.

African startups, researchers and students often work under tighter computing and funding constraints than their counterparts in Silicon Valley. Open models can offer a way to experiment, customise systems for local needs and build products without having to start from scratch or depend entirely on expensive proprietary AI services.

Hugging Face has consequently become part of the wider infrastructure through which developers around the world access the open-AI ecosystem.

Nvidia's challenge now is to convince those developers that owning the platform will not compromise what made it valuable in the first place.

The transaction is expected to close in the first half of 2027, subject to regulatory and other customary approvals, but the message from Nvidia is already clear.

The company that made its fortune selling the machinery behind the AI boom is no longer content to simply power the revolution. Now, it wants a bigger stake in deciding what gets built on top of it.