Nvidia, GM announce deal for AI, factories and next-gen vehicles

General Motors and Nvidia have agreed to a strategic collaboration that includes the automaker using several products and artificial intelligence services from the tech giant for its next-generation vehicles, advanced driver-assistance systems and factories.

The companies on Tuesday announced that the new initiatives include building custom artificial intelligence systems using Nvidia compute platforms, including “Omniverse with Cosmos,” for optimizing GM’s factory planning and robotics.

The Detroit automaker also said it will use “Nvidia Drive AGX” for “in-vehicle hardware for future advanced driver-assistance systems and in-cabin enhanced safety driving experiences.”

GM declined to disclose a cost for the new tools with Nvidia. The tech company has been attempting to diversify its automotive business, which has notably included substantial work in data centers and GPUs.

“The era of physical AI is here, and together with GM, we’re transforming transportation, from vehicles to the factories where they’re made,” Jensen Huang, Nvidia founder and CEO, said in a release. “We are thrilled to partner with GM to build AI systems tailored to their vision, craft and know-how.”

GM has been using Nvidia graphics processing units, or GPUs, for training AI models across various areas of its business, including simulation and validation. The new business expands to in-vehicle hardware, automotive plant design and operations, the companies said.

The automaker also had been testing Nvidia’s Omniverse since at least 2022. Some of GM’s testing was in designing a “digital twin,” or replica, of its new design center and processes to assist virtual vehicle development. It also acted as a single digital environment for employees to work and collaborate in, according to a video last year featuring GM for Nvidia’s GTC developer conference in 2023.

Nvidia anticipated it would strike a deal with GM mid-last year for Omniverse, according to an internal company email viewed by CNBC. At that time, two sources with GM signaled the automaker wasn’t sure Nvidia’s software and GPUs were worth the high cost compared with other companies.

It wasn’t immediately clear what sealed the deal for GM. But since that time, both companies have experienced increased competition from China and uncertain regulatory changes such as tariffs. GM’s stock is off roughly 8% during in 2025, while Nvidia is off about 12% this year.

“AI not only optimizes manufacturing processes and accelerates virtual testing but also helps us build smarter vehicles while empowering our workforce to focus on craftsmanship,” GM CEO Mary Barra said in Nvidia’s release. “By merging technology with human ingenuity, we unlock new levels of innovation in vehicle manufacturing and beyond.”

The companies announced the new initiatives in connection with Nvidia’s GTC AI conference this week in California.

Nvidia describes Omniverse as a platform for “developing and deploying physically based industrial digitalization applications.” It’s essentially connecting a physical environment with a digital, or software, world to optimize processes using a “digital twin” of a physical environment such as a GM design facility or plant.

Users of Nvidia’s Omniverse have included BMW, Amazon Robotics and Samsung, Rev Lebaredian, Nvidia vice president of Omniverse and simulation technology, said during a media briefing a year ago. He said the company was licensing Omniverse for $4,500 per GPU, per year.

It’s unclear how many GPUs GM will need. But given the amount of robotics, sensors and other systems needed to operate a modern assembly plant, it would likely be quite a bit.

More than 20 other automakers have used Nvidia’s “system on a chip” hardware in the central computing units of their smart vehicles, including Mercedes Benz, Volvo, Audi, Volkswagen and BYD, according to an industry equity research note from Jeffries in November 2023.

In recent years, Nvidia has seen soaring demand for its GPUs, which are used for everything from bitcoin mining to AI inference and training.