Nvidia is making a deeper move into the artificial intelligence race, expanding its role beyond the powerful chips that have made it one of the biggest beneficiaries of the AI boom.

The company is developing a new family of AI models called Nemotron 4, according to reporting by The Information cited by Reuters. The project is intended to compete with leading open models and could include a model with at least 1 trillion parameters, although Nvidia has not confirmed that figure or announced a release date.

The development is significant because Nvidia's dominance has largely been built on supplying the computing hardware needed to train and operate AI systems. Its latest move suggests the company wants a stronger position in the software layer as well.

Nvidia Steps Further Into the Model Race

Nvidia is already developing and releasing open AI models through its Nemotron family. The company describes Nemotron as a collection of open models designed for applications including AI agents and other complex workloads. Its models, training resources and supporting tools are available to developers through Nvidia's platforms.

The company also recently released Nemotron 3.5 Lightning, a smaller open model designed for high-volume workloads, alongside NeMo Switchyard, a tool that helps developers route requests between different AI models.

Nemotron 4 would represent a much larger step. According to Reuters, Nvidia is targeting leading open models with the new family, while people familiar with the project told The Information that the largest version could reach at least 1 trillion parameters. The company has not confirmed when the model will be ready.

Open AI Becomes a Bigger Battleground

Nvidia's push comes as open and open-weight AI models gain momentum around the world.

Chinese companies have released increasingly capable and relatively inexpensive models, putting pressure on American technology companies to offer alternatives. Reuters reported that companies including Nvidia and Meta are responding with greater investment in open models as businesses look for AI systems that can be customized and deployed more affordably.

Open models can give developers greater control over how AI systems are adapted and deployed. They can also reduce dependence on a single provider, an increasingly important consideration for businesses building AI into their operations.

But the approach has also sparked debate over security. Open models can be modified and used for a wide range of purposes, making questions about cybersecurity and potential misuse more complicated.

The issue has even reached Washington. Republican Senator Jim Banks recently urged the Trump administration to support U.S. open-weight AI development while reducing reliance on Chinese models, arguing that the technology has national security implications.

A Strategy That Could Strengthen Nvidia’s Core Business

Nvidia's interest in AI models is also closely connected to its hardware business.

Powerful AI systems require enormous amounts of computing capacity, and Nvidia remains one of the industry's most important suppliers of the GPUs used to train and run them. Developing competitive models could therefore create another reason for developers and businesses to build their AI operations around Nvidia's technology.

The company has increasingly presented itself as more than a chip manufacturer. Its software, developer tools and AI models are becoming part of a broader ecosystem designed to keep customers within Nvidia's computing platform.

Nemotron 4 could push that strategy further

The project is still under development, so its eventual capabilities, size and release schedule remain uncertain. It is also too early to know whether it will challenge the strongest models from other AI companies.

What is already clear, however, is that Nvidia is no longer content to simply provide the machinery powering the AI revolution. With its growing investment in models and software, the company is positioning itself to influence what happens on top of the chips as well as inside them.