NVIDIA’s GTC Conference A series of announcements have begun, highlighting its role in advancing autonomous driving technology.
Chipmakers provide automakers and self-driving car companies with a handful of NVIDIA-branded tools to help power self-driving cars and create digital twin factories. On Tuesday, companies including Torc, Gatik and even General Motors announced plans to use NVIDIA products for auto production and robotics for autonomous driving.
To avoid any confusion in the NVIDIA brand name that gives it hardware and software tools, here is a quick term:
- Driver AGX: NVIDIA’s on-board supercomputer that processes real-time sensor data.
- Drive Orin Soc: A system on chip is a more advanced central computer. Orin processes data from various sensors at Level 4 Autonomous and powers the new AV.
- Drive Thor Soc: Thor is the next step. It is optimized for the transformer architecture so it can accommodate generative AI and is designed to handle everything from autonomous driving operations to cockpit functionality to infotainment.
- Drive: This is the security-centric operating system of NVIDIA’s NVIDIA AV platform that promises security, real-time AI processing, and integration of advanced driving and cockpit capabilities.
- Omniverse: NVIDIA’s Omniverse is an analog platform that allows automakers to develop and operate complex, AI-a-Sable virtual environments to generate synthetic data, test AV software, build digital twins in digital factories, and more.
- universe: Cosmos builds on promoting world model training in physical AI development of AVs and robots developed for world AI.
The most important thing is NVIDIA unveiled today Haloit is defined as AI-driven security systems for AV and future physical AI, such as humanoid animal robots. Halos brings together many of NVIDIA’s lineup of automotive hardware and software security solutions, so it’s considered an umbrella.
Here is a quick overview of GTC’s NVIDIA car announcements on Day 2.
General Motors
General Motors announced it has expanded Partnership with Nvidia In collaborations that touch on all aspects of the automaker’s business, including their factories, robots and self-driving cars. Let’s start with the factory.
General Motors said it will use Omniverse to train AI manufacturing models with Cosmos and help it build its next-generation factory. Omniverse will allow GM to build its digital plant to effectively test new production processes without disrupting existing vehicle production. It will also use Omniverse to train robot platforms for operations such as material processing and transportation.
On the self-driving car side, GM said it will use NVIDIA’s Drive AGX for in-car hardware to enable future advanced driver assistance systems and a Carbine safety experience.
Gettyk
Gatik, a self-driving truck companySupported by Isuzu and Goodyear Ventures, it also joins Nvidia’s automotive ecosystem. The Silicon Valley-based company specializes in automated mid-way logistics through autonomous box trucks, said it will develop and deploy Drive AGX accelerated by Drive Thor to act as its AI brain in its truck fleet. It also runs its AI model on the Driveos system for security, Gatik said.
The startup notes that the partnership will help accelerate the deployment of Level 4 automatic trucks for the company’s customers, including Walmart, Kroger and Tyson Foods.
add
Additionally, an automatic freight software launched, said Tuesday it will use the Cosmos World Foundation model to accelerate the testing and development of its self-driving drive SuperDrive.
The company said Plus’s SuperDrive system is built on NVIDIA’s Drive AGX platform. Plus also said in a statement that it is pioneering “AV 2.0 technology that includes generating AI, visual language models and other basic models.” As can be seen from the vocabulary above, NVIDIA’s AGX platform is more suitable for ADA and low-level autonomy. To get more advanced sensor fusion and the onboard calculations required for higher levels of autonomy, companies often rely on Nvidia’s Orin or Thor Socs.
TechCrunch has been contacted by Plus for clarification.
The startup recently concluded a deal with commercial vehicle manufacturers including Traton Group, Iveco and Hyundai to integrate superdrives into their trucks. Additionally, its technology was tested on public roads in Texas and Sweden, with the goal of a commercial launch in 2027.
Torc
Another autonomous truck company, Torc, announced it is working with NVIDIA to develop scalable physical AI computing systems for its AVs. Virginia-based TORC is a subsidiary of Daimler Trucks and will also work with Flex, which builds an automotive-grade computing platform.
TORC says it is using cocktails with Nvidia chip architecture, including Drive AGX, Drive Orin and Driveos to support future autonomous driving capabilities deployments Commercial Release in 2027.
In October 2024, the company won the first Out-of-drive test On the closed route in Texas.
Volvo
Although Volvo has not partnered with NVIDIA to accelerate its autonomous driving technology, automakers rely on Nvidia’s Blackwell GPU to power aerodynamic simulations.
Instead of using Nvidia’s Omniverse emulator, Volvo is working with ANSYSsoftware simulation company. Powered by eight Blackwell GPUs, ANSYS so-called “fluential” simulation software helped Volvo design its new EX90 electric vehicle to reduce aerodynamic drag and thus improve battery performance.
Ansys says its fluent simulator helps Volvo reduce the total simulation run time from 24 hours to 6.5 hours, allowing multiple design iterations per day, optimized vehicle design, and time to accelerate to the market.
