Every modern car is a rolling computer, and the major carmakers are moving to more aggressively leverage artificial intelligence as they strive to build vehicles that can drive autonomously and provide passengers with a digital-first experience. NVIDIA and General Motors announced strategic partnerships March 18, when America’s largest automaker gave notice of its plans to do just that.
Speaking at NVIDIA GTC 2025, the conference for developers, analysts, AI experts and others taking place in San Jose, Calif., GM said it would expand a long-standing partnership with NVIDIA around manufacturing and autonomous vehicle development. David Richardson, GM’s senior vice president of software and services engineering, said that GM would begin using Omniverse, NVIDIA’s solution that helps companies build three-dimensional environments, including digital twins, and Drive AGX, its AV development platform.
“We’ve been working with NVIDIA for a while,” Richardson said. “What’s happening now is we’re taking that to the next level.”
Click the banner below to sign up for our newsletter and follow the latest business and tech trends.
GM Will Use Digital Twins To Improve Manufacturing
With respect to Omniverse, he said the company was attracted to the opportunity to “set up full digital twins in our factory lines,” so that it can “look for optimizations before we build factories.” Digital twins are fully virtual versions of physical environments, such as a factory, updated in real time with information from sensors and other technology.
“We have a huge manufacturing footprint,” he said. “So, you have to think about things like how to set production lines efficiently, how to batch fill a vehicle so you get the most throughput you can. I’m not a mechanical engineer, but there’s stuff like the wear and tear on parts, conveyor speed, you name it, there are a lot of components that you have to get right. And for us, it’s not just efficiency and therefore the bottom line of the company but it’s also safety.”
In addition to using Omniverse to build digital twins, GM has used it in other ways. For example, Richardson said, on its latest Corvette ZR1, GM “did a bunch of work with Omniverse to model out the performance of that engine, a bunch of calibrations, before we took it to Germany to set the land speed record at 233 miles per hour,” which it did last October. “That was important to us because we only had three hours on that track.”
GM will also work with NVIDIA on AI-powered robotics to help automate certain factory processes. Most welding is already done by robots in GM factories, Richardson said, an example of how the company is trying to build more automation into manufacturing process so that “the human workforce can focus on craft and quality.”
Click the banner below to learn how organizations are deploying artificial intelligence.
Level 3 AV Technology Coming to GM Cars
In December, GM announced a shift in its strategy concerning autonomous vehicles. It said it would combine the teams working on Cruise LLC, a driverless taxi company that had previously operated as a mostly independent subsidiary, with the team that worked on Super Cruise, a “hands-off, eyes-on driving feature” that is offered on many GM models.
Part of the problem with Cruise LLC, Richardson said, is that it required the company to run a taxi business when it wanted to focus squarely on developing the AV technology. Now, its two AV technical teams, previously siloed, are working together on that single objective. “Our first target is a Level 3 solution, and that’s what we’re working on now,” he said.
A Level 3 autonomous vehicle can handle all driving tasks under specific conditions, but the driver must be ready to take over when prompted by the system. It’s two steps below the most advanced system, as defined by the Society of Automotive Engineers, in which no driver is needed.
To get there, GM will leverage Omniverse and other technologies to train the AI in lifelike digital road environments, and Drive AGX, an in-vehicle platform that NVIDIA says “integrates a full sensor architecture, high-performance AI compute, and a robust software stack to accelerate autonomous vehicle development and deployment.”
“With NVIDIA, obviously there’s a cloud-based training aspect for that autonomy, but then there’s the on-vehicle hardware part of the platform. I think it’s really a good thing because NVIDIA’s ecosystem is really world-class, and it allows us to tap into the talent that’s out there, existing platforms and toolchains, and the same stuff that runs in the cloud can run in the car.”
Richardson said GM hopes to offer Level 3 autonomous driving “as quickly as possible,” but declined to define a more specific time frame, nor would he say in which models the technology would be offered.
Keep this page bookmarked for articles from the event, and follow us on the social platform X @BizTechMagazine and the official conference feed, @NVIDIAGTC.
