Skip to main content
  1. Home
  2. Cars
  3. Computing
  4. News

Nvidia offers a peek under the hood of its Drive PX 2 platform for self-driving cars

Add as a preferred source on Google

At the Hot Chips conference in Cupertino, California, Nvidia offered more details about its hardware platform for self-driving cars.

It used the conference to unveil Parker, the processor used in the Drive PX 2 platform unveiled at CES earlier this year. Two Parker processors are used in the Drive PX 2, which is already being used by 80 carmakers, Tier 1 suppliers, and university research centers for development of autonomous cars, according to a Nvidia blog post detailing the processor.

Recommended Videos

The Parker processor is part of a scalable architecture, allowing carmakers to use just one processor as part of a stand-alone system, or upgrade to the full Drive PX 2 setup. The system is capable of hosting multiple functions, including infotainment, digital instrument clusters, and driver assistance functions. Self-driving cars with Parker processors can also receive updates from the cloud, letting engineers deploy tweaks to the algorithms that control them more easily.

Read more: Ford promises self-driving car by 2021

The processor uses a new 256-core Pascal GPU, which Nvidia says provides the performance needed to run “advanced deep learning inference algorithms” for self-driving cars, as well as the graphics capabilities for in-vehicle displays such as digital instrument clusters and center-stack screens. Nvidia claims the complete Drive PX 2 system with two Parker processors can run up to 24 trillion deep learning operations per second.

Nvidia already has plans with Volvo to deploy the Drive PX 2 platform in the Swedish carmaker’s autonomous test vehicles. The Nvidia hardware will be used in 100 automated XC90 SUVs, which will be tested on designated public roads in Gothenburg, Sweden, as part of Volvo’s ongoing “Drive Me” autonomous-car development program.

Drive PX 2 will also be used in the autonomous race cars being developed for Roborace. The series is an offshoot of the Formula E race series for electric cars, and will be the first based around autonomous vehicles. Roborace released renderings of a radical robotic race car earlier this year, and has been testing a prototype over the past few months. It hopes to race the finished car in conjunction with the 2016-17 Formula E season.

Stephen Edelstein
Stephen is a freelance automotive journalist covering all things cars. He likes anything with four wheels, from classic cars…
iOS 26.4 adds ChatGPT to you car’s infotainment screen
Apple's iOS 26.4 brings ChatGPT, Gemini, and Claude to your car's screen, adds calming ambient music widgets, and previews the in-car video future that drivers have been waiting for.
CarPlay shown in March 2025.

Apple rolled out iOS 26.4 recently, and while your iPhone got several upgrades, CarPlay quietly had one of its best days in years. The latest iPhone updates bring two meaningful features that can change the way you use CarPlay on your car’s infotainment screen. 

Would you use ChatGPT while driving?

Read more
Sony and Honda’s electric car dream with Afeela series is officially dead 
Sony Honda Mobility has shelved the Afeela 1 and its follow-up, and the EV market has another high-profile casualty.
Machine, Wheel, Adult

Sony and Honda’s shared dream of launching an electric car has just come to an end. The joint venture between the two brands — Sony Honda Mobility — has just announced that plans for the upcoming Afeela 1 electric car have been shelved. Additionally, the follow-up model has been nixed from the roadmap. 

But why did the Afeela go?

Read more
This AI checks if your driving habits signal crash risk
Researchers say eye tracking, heart rate, and personality data can flag risk early.
Person, Wristwatch, Car

A new AI model is taking aim at a question most drivers don’t ask soon enough. How likely are you to crash before you even start the engine?

The system looks at how you behave behind the wheel, pulling in signals like eye movement, heart rate, and personality traits to flag warning patterns early. Instead of waiting for real-world mistakes, it relies on simulated driving tests to surface behaviors linked to dangerous outcomes.

Read more