Nvidia CEO Jensen Huang announced a collaboration with Toyota at CES 2025 to develop next-gen autonomous vehicles (AVs) using Nvidia’s advanced “Thor” AGX processor, marking a significant step in AV innovation.
The Thor chip, with 20 times the processing power of its predecessor, integrates data from sensors like cameras, radars, and Lidars to enable real-time autonomous navigation and serves as a universal robotics computer for various applications.
Nvidia’s Drive OS achieved ASIL D certification, the highest automotive safety standard, ensuring its CUDA platform is functional-safe for AVs and robotics after 15,000 engineering years of development.
Nvidia introduced Omniverse and Cosmos platforms to generate synthetic data and simulate driving environments, enabling AI models to train on billions of virtual miles and reducing the “sim-to-real gap.”
Nvidia’s partnerships and Thor’s deployment in AVs and driverless trucks, position the company as a key player in the multi-trillion-dollar AV industry, advancing toward Level 5 autonomy.
At the heart of Nvidia’s announcement is the Thor processor, a next-generation robotics computer designed to handle the immense data demands of autonomous vehicles. Huang described Thor as “20 times the processing capability” of its predecessor, the Orin chip, which is currently the industry standard for AVs. Thor integrates data from a “madness amount of sensors,” including high-resolution cameras, radars and Lidars, to predict and navigate driving paths in real-time.
“This is a universal robotics computer,” Huang said, holding up the Thor chip during his presentation. “It’s not just for cars — it’s the brain for robots, whether they’re autonomous vehicles, humanoid robots or industrial manipulators.”
Nvidia ready to provide processors to Toyota’s next-generation autonomous vehicles
The Thor processor is already in full production, marking a significant milestone for Nvidia and its automotive partners. Its deployment in Toyota’s next-generation AVs highlights the chipmaker’s growing influence in the automotive sector, which Huang estimates is already already a $4 billion business for Nvidia, with projections to reach $5 billion by the end of the year.
Nvidia’s Drive OS, a software-defined AI computer, also took center stage at CES 2025. Huang announced that Drive OS has achieved ASIL D certification, the highest standard of functional safety for automobiles under ISO 26262. This certification, the result of 15,000 engineering years of work, ensures that Nvidia’s CUDA platform is now a functional-safe computer for robotics and AVs.
To address the challenges of training AVs in real-world scenarios, Nvidia introduced its Omniverse and Cosmos platforms. These tools generate synthetic data to simulate driving environments, enabling AI models to train on billions of virtual miles.
“We take thousands of drives and turn them into billions of miles,” Huang explained. “This is how we create mountains of training data for autonomous vehicles.”
The combination of Omniverse and Cosmos allows developers to reconstruct digital twins of real-world environments, generate photorealistic driving scenarios, and amplify training data with near-infinite variations. This approach reduces the “sim-to-real gap,” ensuring that AI models are trained on physically accurate and plausible data.
Huang’s keynote painted a vivid picture of the AV industry’s future, predicting it will become the “first multi-trillion-dollar robotics industry.” He highlighted partnerships with major automakers, including Toyota, Volvo, Rivian and Mercedes-Benz, as well as self-driving truck companies like Aurora and Waabi.
Aurora, in particular, announced plans to deploy driverless trucks powered by Nvidia’s Thor chips by 2027, with production samples expected in 2025.
“Delivering one driverless truck will be monumental. Deploying thousands will change the way we live,” said Aurora CEO Chris Urmson.
The push for autonomous vehicles has been decades in the making, with companies like Tesla and Waymo leading the charge. However, the technology has faced significant hurdles, including regulatory challenges, high-profile accidents and the complexity of training AI systems to handle unpredictable real-world scenarios.
Nvidia’s advancements in AI, simulation and hardware are addressing these challenges head-on. By partnering with industry leaders like Toyota and Aurora, Nvidia is positioning itself as a cornerstone of the AV revolution. The Thor processor and its associated platforms represent a critical step toward achieving Level 5 autonomy – where vehicles can operate without human intervention in all conditions.
As Huang noted, “The AV industry is here. This is an incredibly exciting time.”
With Nvidia at the forefront, the dream of fully autonomous vehicles is closer than ever to becoming a reality, promising to reshape transportation, logistics and robotics on a global scale.
A Russian spy was living in a "typical seaside hotel" on the English coast crammed full of electronic surveillance equipment, a court has heard.Orlin Roussev boasted to his controller that he was becoming like the James Bond character "Q" as he prepared his spying "toys" for kidnap and surveillance operations across Europe. He is said to have taken instructions from a handler called Jan Marsalek, who is wanted in connection […]
Post comments (0)