Nvidia wants to be the mastermind of your own self-driving car

Nvidia wants to be the mastermind of your own self-driving car

Chip designer Nvidia on Tuesday unveiled a new processor called Drive Thor that it expects to fuel the autonomous car revolution.

Thor processors should arrive in 2024 for cars hitting the roads in 2025, starting with Chinese auto maker ZeekrThe 001 EV is the 001 EV, said Danny Shapiro, vice president of the automotive business at Nvidia. They depend on Nvidia’s new Hopper GPU To better deal with the artificial intelligence software that is key to self-driving cars.

“You will definitely rise to the level of full autonomy,” Shapiro said, referring to Level 4 or Level 5 self-driving capabilities, where cars can drive themselves without occupants noticing or attending.

Nvidia planned to Chip called Atlan for 2024 but scrapped it in favor of Thor, which handles AI software at 2 quadrillion operations per second – twice the planned speed of Atlan and eight times the speed of the current Orin processor. Thor includes one major Hopper feature: the ability to accelerate a powerful AI technology called transducers. Nvidia also anticipates lower-level Thor variations for less revolutionary driver-assist technologies like lane keeping and automatic emergency braking.

The car processor market is large and growing as automakers demand more and more processors and other semiconductor chips for driver assistance, infotainment, and electronic control units that oversee everything from engine combustion to GPS navigation. All The Porsche Taycan contains 8000 semiconductor components.

Chip designers are taking advantage of the new market. Nvidia has $11 billion in auto chip ordersand its biggest competitor, Qualcomm, has $19 billion in auto orders in the pipeline.

Also new in GTC from Nvidia

Among other Nvidia developments at its GTC event:

  • that it GeForce RTX 4090 graphics cardsIt’ll go on sale in October, powered by its new generation Ada Lovelace GPUs for PCs and workstations, with prices ranging from $899 to $1,599.
  • The Jetson Orin line of robot processors now includes nano models for smaller robots. It draws between 5 and 15 watts of power to improve battery life, costs $199 and up, and begins shipping in January. Nvidia said recently announced companies that use Jetson Orin include Canon, John Deere, Microsoft Azure, Teradyne and TK Elevator.
  • The new Nemo LLM technology is designed to help researchers make more use of large language models, a hot new area in artificial intelligence responsible for rapid advances in processing language, images, and other data. Retraining an LLM consumes huge resources, but Nemo’s technology allows researchers to perform much faster AI training that customizes big AI.

Thor auto auto chip details

With 77 billion transistors, Thor would be a huge, if not the largest processor on the market. But it will allow automakers to replace a batch of smaller chips that are heavier, more expensive and more energy-hungry. In addition to using Hopper GPUs, it borrows CPU cores from Nvidia’s 2023 Grace Processor for traditional computing tasks. It also draws technology from Nvidia’s latest GPU technology for gaming and design, the Ada Lovelace architecture.

Huang said the design will make it easier for automakers to improve their cars’ software through over-the-air updates. Tesla has been a major technology leader in this technology for years.

Huang said Thor will also be used in robots and medical equipment. It will be able to run three operating systems simultaneously – Linux, QNX and Android – for different parts of the vehicle’s computing environment. Nvidia said the segmentation technology ensures that less important work, such as information and entertainment, does not interrupt important safety-related work.

Computer display of Nvidia Thor processor on a complex electronic board

This computer rendering shows Nvidia’s Drive Thor processor integrated into an automotive electronics board with several connectors for cameras, radar, lidar, and other sensors to enable self-driving cars.


With self-driving vehicles, promised for years but still only being tested, these chips become even more important.

“The industry has realized that it is a much more complex task than initially thought,” Shapiro said of self-driving vehicles. “With safety so critical, no one is ready to release these vehicles into the wild until there is more computing involved.”

#Nvidia #mastermind #selfdriving #car

Leave a Comment

Your email address will not be published.