Nvidia Partners with Toyota for Self-Driving Technology

Nvidia CEO Jensen Huang showcased the company's latest innovations at the 2025 Consumer Electronics Show, including the announcement of a significant partnership with Toyota, the largest car manufacturer in the world. During his keynote address, Huang revealed that Toyota will incorporate Nvidia's advanced self-driving chips and software into various car models.

Huang proclaimed, "The AV revolution has arrived," signaling a new era for autonomous vehicles (AVs). He emphasized the partnership's potential to create next-generation AVs, forecasting that self-driving cars will become a major trillion-dollar market in the realm of robotics.

Nvidia's New AI Technologies

In conjunction with the announcement, Nvidia introduced a suite of AI technologies called "Cosmos." This comprehensive package includes cutting-edge generative world foundation models designed specifically for physical movement in environments like vehicles and robots. Huang explained that Cosmos is paired with Nvidia's physics simulation tool, Omniverse, which enables the creation of simulations that can eventually help train robots and cars. "Take thousands of drives and turn them into billions of miles," he noted, capturing the synergy between these technologies.

Huang described Cosmos as "the world's first world foundation model," mentioning that it has been trained on a staggering 20 million hours of video. The goal is to teach AI how to decipher the physical world around it. He drew a parallel between Cosmos and the popular Llama large language model by Meta Platforms, suggesting that he hopes Cosmos will have a similar transformative impact on robotics and AI.

Enhancements in Robotics Training

The integration of Cosmos with Omniverse can also be utilized for training robots in controlled environments. For instance, a robot could practice tasks in a simulated warehouse before executing them in the real world. Huang stated that the Cosmos framework would be open-sourced on GitHub, allowing developers to access and contribute to the technology.

Nvidia has enjoyed a long-standing relationship with Toyota, having previously collaborated on AI model training for self-driving technologies. This latest announcement represents a significant expansion of that partnership. Ali Kani, Nvidia's head of automotive products, confirmed that Toyota will employ the company's powerful "AGX" onboard AI computer, known as the Thor AGX, which is touted to be twenty times more potent than the previous Orin model.

Robotics and AI Software Updates

Huang also unveiled upcoming functionalities for robots, including the use of Apple's Vision Pro headset for demonstration tasks. The headset captures a user's movements, which are then processed through Cosmos and Omniverse to generate synthetic training data for robots. Huang noted the potential for a forthcoming "ChatGPT moment" in robotics, signifying a breakthrough in AI capabilities.

Furthermore, Nvidia presented a new range of AI software updates, including Llama Nemotron, a suite built on Meta's Llama. Huang pointed out that Llama has driven various organizations to explore AI technologies, and Nvidia's versions will refine Llama for enterprise applications.

Exploration of Agentic AI

During his keynote, Huang discussed the rising significance of "agentic AI" - a concept wherein large language models can interact with external programs to accomplish tasks. Justin Boitano, Nvidia's president of enterprise AI software products, elaborated on this development, noting that numerous startups are reimagining the programming model to allow developers to create AI with specific roles and goals simply through prompts.

Huang mentioned that agentic AI, combined with advancements in self-driving vehicles and robots, represents three major types of robotic applications being developed. He also revealed the latest version of the combined CPU and GPU chip, GB10, which powers Nvidia's new DIGITS desktop computer.

Other Key Announcements

Nvidia's keynote featured other announcements, such as the GEFORCE "Blackwell" gaming GPU, which is set to launch at a significantly lower price point, making it more accessible. Additionally, the company unveiled Project DIGITS, a compact personal computer optimized for AI development, which incorporates the new GB10 chip.

Nvidia, Toyota, AI