Dept of Justice reportedly probes Nvidia over $13 billion Hugging Face AI platform acquisition

The DOJ is investigating how Nvidia's proposed acquisition of Hugging Face could affect competition in AI. It is not simply examining the purchase price. Officials are asking whether ownership of Hugging Face would give Nvidia influence over an important route that competitors use to distribute models and tools.

Hugging Face operates an open-source platform for hosting and testing AI models. It also sells AI infrastructure. If Nvidia owned that platform, it could potentially connect model distribution, infrastructure, and Nvidia's own products more closely. The article describes this as a possible way to lock in a key distribution channel.

The investigation is still an inquiry, not a finding that Nvidia broke the law. Nvidia declined to comment on the reported questions but has said Hugging Face will continue supporting different chip and AI providers. The acquisition is expected to become official in the first half of 2027.

Hugging Face provides an open-source platform for hosting and testing new AI models. That makes it useful during several stages of development. Developers can place models on the platform, test them, and make them available to others. The platform therefore helps turn experimental AI work into tools that people can access.

Its importance is not limited to storage. The article also says Hugging Face sells AI infrastructure. That gives the company a role in the systems supporting AI models, as well as in the platform where models are hosted and tested. These combined functions help explain why the DOJ is examining Nvidia's proposed purchase.

The article does not provide user numbers or describe every service Hugging Face offers. It does establish that the platform is open-source and important enough to prompt competition concerns. Nvidia says it will continue supporting a range of chip and AI providers after the acquisition.

The proposed acquisition is valued at almost $13 billion. Nvidia agreed to buy Hugging Face in September, although the article does not specify the year in that sentence. The deal is expected to officially complete in the first half of 2027.

The price suggests Hugging Face is strategically important to Nvidia beyond being a normal software purchase. Hugging Face hosts and tests AI models, offers an open-source platform, and sells AI infrastructure. Those activities place it near developers and customers building AI systems. Nvidia may therefore gain a stronger position across several parts of the AI technology stack.

The price alone does not prove that the acquisition will harm competition or that Nvidia will restrict access. Those are issues for regulators to examine. Nvidia has confirmed that Hugging Face will continue supporting a range of chip and AI providers once the buyout is complete.

A distribution channel is the route through which products reach users. Hugging Face provides such a route for AI models by hosting and testing them on an open-source platform. Developers and companies can use that environment to make models available and connect them with supporting infrastructure.

If Nvidia owned the platform, it would control an important point between model creators and the people using their models. The DOJ is concerned that Nvidia could use this position to lock in distribution for its own products. The article does not say Nvidia has done that. It describes the possibility as the reason for regulatory questions.

Nvidia has said Hugging Face will continue supporting a range of chip and AI providers. That commitment could reduce concerns if maintained. However, the DOJ's inquiry shows why ownership matters: control over hosting, testing, and infrastructure can affect which AI products are easiest to develop, distribute, and use.

Nvidia ownership could affect competitors because Hugging Face is a platform where AI models are hosted and tested. If the platform became closely tied to Nvidia products, competing chip or infrastructure providers might find it harder to receive equal visibility, support, or access. The article presents this as a competition concern, not an established outcome.

The concern is especially relevant because Hugging Face also sells AI infrastructure. Nvidia could control both a model distribution point and services that help run AI systems. The article separately says officials have considered whether Nvidia offered more AI chips to customers buying its networking hardware and Groq chips, showing broader scrutiny of how Nvidia serves customers.

Nvidia has promised continued support for a range of chip and AI providers after the buyout. That is the current stated plan. The DOJ's questions indicate that regulators will examine whether practical conditions match that promise and whether customers feel pressured to use Nvidia's wider product family.

AI developers that do not want to rely on Nvidia-owned services can use several established approaches. They can host models on their own servers, contract with cloud providers, or use independent model repositories and testing platforms. They can also select infrastructure that supports chips from different suppliers. These choices can reduce dependence on one company.

For example, a company could keep its models in an independently operated repository, test them on its own systems, and deploy them through a cloud provider that supports multiple types of AI hardware. Developers could also distribute models directly through their own websites or software tools. These arrangements separate model hosting from Nvidia-owned services.

The source article does not name specific alternatives or compare their costs and performance. It does say Nvidia has promised that Hugging Face will support multiple chip and AI providers. The practical choice will depend on whether that support remains broad and whether developers consider independent options reliable and affordable.

AI systems depend on more than hardware. They also need models, testing tools, hosting, and infrastructure. Nvidia is already highly important in AI hardware, and the article says its powerful AI-focused hardware can only be sold to approved entities. That makes its position in the technology stack especially significant.

Buying Hugging Face could connect Nvidia's hardware position with an open-source model platform and an infrastructure business. The DOJ is therefore asking whether Nvidia could use control of software and distribution to strengthen its position against competitors. The concern is about the combined effect of these layers, rather than any single product.

Nvidia's broader scrutiny includes its purchase of SchedMD, export controls affecting its AI hardware, and questions about allocations of chips to customers. Nvidia says Hugging Face will support various providers. The acquisition is expected to complete in the first half of 2027, when its practical effects can be assessed.