By Max A. Cherney and Stephen Nellis
(Reuters) -AI to better train robots and cars, as well as new gaming chips dominated Nvidia (NASDAQ:) CEO Jensen Huang’s keynote speech at the CES 2025 conference on Monday as the world’s second most valuable firm expounded upon its potential to expand its business.
Nvidia introduced what it calls Cosmos ‘foundation’ models that generate photo-realistic video which can be used to train robots and self-driving cars at a much lower cost than using conventional data.
By creating what is known in the tech industry as “synthetic” training data, the models can help robots and cars understand the physical world similar to the way that large language models helped chatbots generate responses in natural language.
Users will be able to give Cosmos a text description that can be used to generate video of a world that obeys the laws of physics.
This promises to be much cheaper than gathering data as it is done today. To train self-driving cars, for example, companies have fleets of vehicles that roam streets to gather data, and humanoid robots are often trained by having real humans repeat tasks over and over.
Huang, however, cautioned that the Cosmos models will need much more data before hitting their ‘ChatGPT moment’.
Cosmos will be made available on an “open license,” similar to Meta Platforms (NASDAQ:) Llama3 language models that have become widely used in the tech industry.
“We really hope (Cosmos) will do for the world of robotics and industrial AI what Lama3 has done for enterprise AI,” Huang said.
The new gaming chips use Nvidia’s ‘Blackwell’ AI technology to give video games movie-like graphics, especially in a field known as ‘shaders,’ which can help images like a ceramic teapot look more realistic by adding imperfections and fingerprint smudges to its surface.
The new chips also have AI technology to help game developers generate more accurate human faces, an area where players are apt to notice even slightly unrealistic features. The chips, which Nvidia calls its RTX 50 series, will range in price from $549 to $1,999, with top models arriving on Jan. 30 and lower-tier models coming in February.
Nvidia said its mid-grade $549 gaming chips will match the company’s previous flagship chip, the RTX 4090 which sold for $1,600.
Nvidia also said Toyota Motor (NYSE:) will use its Orin chips and automotive operating system to power advanced driver assistance in several models. It did not give details about the models.
Huang expects automotive hardware and software revenue of $5 billion in fiscal 2026, up from an expected $4 billion this year.
Huang also showed off a desktop computer called Project DIGITS. The computer will feature the same chip at the heart of the company’s data center offerings, but paired with a central processor built with help from Taiwan’s MediaTek.
The chips will come in a smaller package that can be used by individual software developers to test their AI systems quickly.
The initial Project DIGITS system won’t exactly be consumer friendly – it will run an Nvidia operating system based on Linux, which is mostly used by computer programmers rather than consumers, and will cost $3,000.
CES, formerly known as the Consumer Electronics Show, runs Jan. 7-10 in Las Vegas.
Nvidia’s stock closed at a record high of $149.43 on Monday, bringing its valuation to $3.66 trillion and making it the world’s second-most valuable listed company behind Apple (NASDAQ:).
