Open Editor's Digest for free
Roula Khalaf, Editor of the FT, selects her favorite stories for this weekly newsletter.
Nvidia is on the verge of revolutionizing robots with artificial intelligence, chief executive Jensen Huang said on Monday, as he outlined his vision for the next stage of the company's phenomenal growth and predicted a “billion dollar” opportunity.
Huang announced a series of new products and collaborations in the area of ”body AI”, including an AI model for humanoid robots and a major collaboration with Toyota to use Nvidia's self-driving car technology, during his keynote speech at the Consumer Electronics Show in Las Vegas. Vegas.
Nvidia has crossed a $3tn market capitalization following demand for its AI chips to become one of the world's most valuable companies. Huang, meanwhile, has become a household name, more than 30 years after he founded Nvidia as a video game graphics company.
Huge lines had formed outside the Mandalay Bay convention center long before the keynote began, with some people still lining up when Huang appeared on stage in a leather jacket version, joking: “I'm in Las Vegas after that.”
Besides semiconductors, Nvidia has developed software that allows companies to train and use robots, from those used in smart factories and warehouses to self-driving cars and humanoids, pushing to expand the use cases of AI running on its chips.
Breaking down the technological challenges involved in sending robots to scale will pave the way to “the biggest technology industry the world has ever seen”, Huang said.
Nvidia said the robotics field has reached a point place to give technologyas AI accelerates and fine-tunes the process of simulating the virtual world and generating the large amount of data needed to train robots. In the next two years, the market for humanoid robots alone is expected to reach 38bn dollars, according to the company.
On Monday, Nvidia announced a suite of basic AI models on its new Cosmos platform, which developers can use for free to generate data and build their own models.
Nvidia said the underlying models, which have been trained on 20mn hours of video data, are essential to the development of technology such as the modeling of large languages that support applications such as OpenAI's ChatGPT. It integrates with Nvidia's Omniverse platform, which is used to run real-world simulations.
“What (those models) do for language, we can now do to understand the physical world,” Rev. Lebaredian, Nvidia's vice president of Omniverse and simulation technology, told the Financial Times. While data in the physical world is more difficult to collect and process than text, Lebaredian said it is “a necessary part” of the company's work.
“The big takeaway (from Huang's CES speech) is that this moment is going to be special,” he added. “I think this year is a turning point where we're going to see this acceleration of AI and robotics.”
The Omniverse platform and robotics currently represent a small share of the company's revenue. In Nvidia's quarter to the end of October, “technological vision” accounted for 486mn of revenue, while cars and robots totaled 449mn dollars.
This is a wealth of the market as a whole, as the company raked in $30.8bn in revenue from selling chips to data centers that power AI models at the same time.
Nvidia's search for new markets comes as it faces growing pressure from its biggest customers, including Amazon and Microsoft, who are rushing to build their own AI data centers.
Analysts at Bank of America said Nvidia's decision to replicate “physical AI” was the “next logical step”. The challenge will be “to make the products reliable enough, cheap enough and widespread enough to generate credible business models”, they added.
At CES, Nvidia also unveiled a collection of basic models of humanoid robots, called “GR00T Blueprint”, which it said will “pay” the development of robots, as well as new tools for developing and testing factory vehicles and maintenance robots to train autonomous vehicles.
Toyota has announced that it will build its next generation of autonomous vehicles on Nvidia hardware and software, known as Drive AGX. Self-driving car group Aurora and auto parts maker Continental will use Nvidia hardware and software to power thousands of self-driving trucks under their long-term partnership with the chipmaker.
Nvidia said it expects its automotive business to grow to $6bn by the 2026 financial year. Autonomous vehicles “will be the first multi-dollar robotics industry”, Huang told the CES audience.
In contrast, Nvidia said it will release an “AI supercomputer” with its latest and most powerful AI chip, Blackwell, which allows researchers and students to run AI models of billions of parameters locally instead of in the cloud. It will be available in May for a starting price tag of $3,000.