Nvidia Reveals New Robot Training Tech, Toyota Deal, and Advanced Gaming Chips

During his keynote address at CES 2025, Nvidia CEO Jensen Huang unveiled a series of groundbreaking advancements aimed at propelling the company’s influence in the AI, robotics, automotive, and gaming industries. The CEO’s speech focused on the potential of artificial intelligence (AI) to transform a wide array of sectors, particularly through the introduction of new models and chips that promise to revolutionize how robots, self-driving cars, and video games are developed.

A centerpiece of Huang’s presentation was Nvidia’s introduction of Cosmos “foundation” models, a set of AI-driven technologies designed to generate photo-realistic video that can be used for training robots and autonomous vehicles. Unlike traditional methods, which rely on gathering real-world data through fleets of cars or repetitive human tasks, Cosmos offers a much more cost-effective solution. By creating synthetic training data, these models can simulate realistic environments that adhere to the laws of physics, allowing robots and self-driving cars to “see” the world and learn to navigate it in a manner akin to how large language models have advanced natural language processing. Users can provide a text description, and Cosmos will generate videos of scenes that follow realistic physical interactions, offering a revolutionary method of creating data without the high costs of real-world collection.

Huang, however, tempered expectations, acknowledging that the Cosmos models will require extensive additional data before they reach a transformative level comparable to the “ChatGPT moment.” Despite this, the announcement marks a major leap forward in the convergence of AI and robotics. Cosmos will be made available under an open license, similar to Meta Platforms’ Llama3 language models, with Huang expressing hope that Cosmos could have a similar impact on robotics and industrial AI as Llama3 has had on enterprise AI.

In addition to the Cosmos models, Huang unveiled Nvidia’s next-generation gaming chips, powered by the company’s new ‘Blackwell’ AI technology. These chips are designed to enhance video game graphics to cinematic levels, especially in the realm of “shaders,” which add realistic imperfections such as fingerprints or dust to objects like a ceramic teapot. With a focus on improving the realism of human faces—an area where players are particularly discerning—the new RTX 50 series chips aim to provide a more immersive gaming experience. The chips, priced between $549 and $1,999, are expected to redefine the gaming landscape, with the mid-range models offering performance on par with the company’s previous flagship RTX 4090, which retailed for $1,600. The top models will be released on January 30, with lower-tier variants arriving in February.

Additionally, Huang highlighted Nvidia’s continued expansion in the automotive sector, announcing that Toyota Motor would utilize Nvidia’s Orin chips and automotive operating system for advanced driver assistance systems (ADAS) across several vehicle models. While specific model details were not disclosed, this partnership reinforces Nvidia’s growing dominance in the automotive industry. The company expects to generate $5 billion in automotive hardware and software revenue by February 2026, up from an anticipated $4 billion this year.

CES 2025, which runs from January 7 to 10 in Las Vegas, has become a key platform for unveiling cutting-edge products, from automotive innovations to AI-powered solutions. Nvidia’s presence at the conference underscores its position at the forefront of technological advancement, with the company’s stock reaching a record high of $149.43, pushing its market valuation to a staggering $3.66 trillion—making it the second-most valuable company in the world, trailing only Apple. With the continued growth of AI, robotics, and gaming, Nvidia is poised to remain a dominant force in the tech industry for years to come.