General Motors is turning to Nvidia to bring AI into the physical world with an expanded collaboration designed to touch every aspect of the automaker’s business, including factories, robots and self-driving cars.
Nvidia founder and CEO Jensen Huang, who announced the partnership on Tuesday at the company’s GTC conference held in San Jose, said the time for self-driving cars has arrived.
“I look forward to building GM AI in all three areas,” he said on stage. “Manufacturing AI allows them to revolutionize the way they manufacture.
This transaction means that NVIDIA will provide GM with AI infrastructure (essentially GPU) and help automakers build their own AI.
Nvidia has been associated with the automotive and autonomous vehicle industry for decades, supplying GPUs to companies like Tesla, Wayve and Waymo to use them in data centers and their vehicles. Nvidia has also developed an autonomous vehicle platform for automakers. This includes an operating system called DriveOS, which offers real-time AI processing and integration of advanced driving and cockpit functions. Earlier this year, Toyota announced plans to equip the next-generation vehicles with automated driving capabilities powered by Nvidia’s Drive AGX Orin SuperComputer and Drivos, a safety-focused operating system.
“We work with the automotive industry, but the automotive industry wants to work with them,” fans said in their keynote speech. “Build all three computers: a training computer, a simulation computer, a robotics computer (autonomous vehicle computer). All the software stacks above it, models and algorithms just like every other industry I have demonstrated.”
GM will work with NVIDIA to build custom AI systems using some of Tech Giant’s products. GM did not disclose the financial value of the transaction.
GM will use the space-equipped Nvidia Omniverse to train AI manufacturing models to help build next-generation factories and robotics. With Omniverse, GM can build factory digital twins, and even assembly lines, to effectively test new production processes without disrupting existing vehicle production. The initiative includes the training robotics platform that GM already uses for operations such as handling, transporting, and precision welding.
Automakers will also use NVIDIA Drive AGX in in-vehicle hardware for future advanced driver assistance systems and safe in-cabin driving experiences. The automaker recently stopped funding its commercial Robotaki development project with a pivot that shifted resources to a driver assistance system over handoffs known as Super Cruises. GM is in a process that combines it with its own efforts to absorb cruises from its autonomous automobile subsidiary and develop driver assistance functions.
The relationship between GM and Nvidia is not new. The Detroit-based automaker uses NVIDIA GPUs to train AI models for simulation and verification. The expanded transaction includes the use of NVIDIA AI products, and therefore focuses on improving the design and operation of automobile plants.