Nvidia's $3,000 mini PC steals the show at CES

Nvidia CEO Jensen Huang was hailed like a rock star this week at CES in Las Vegas, after artificial intelligence Boom This made the chipmaker the second most valuable company in the world.

In his nearly two-hour keynote address Monday at the start of the annual conference, Huang packed a 12,000-seat auditorium and drew comparisons to the way Steve Jobs unveils products at the expo. apple Events.

Huang concluded his speech with an Apple-like trick: the surprise product reveal. He presents one of Nvidia's server racks and, using some theatrical magic, lifts up a much smaller version, which looks like a small cube of a computer.

“This is an artificial intelligence supercomputer,” Huang said while wearing a crocodile-skin jacket. “It powers the entire Nvidia AI stack. All of Nvidia's software runs on this.”

The computer is called Project Digits and runs a relative of the Grace Blackwell graphics processing units (GPUs) that currently power more advanced AI server farms, Huang said. The GPU is paired with armGrace-based CPU. Nvidia worked with the Chinese semiconductor company mediatek To create a system on a chip called GB10.

CES, formerly known as the Consumer Electronics Show, is the place to launch flashy, futuristic consumer gadgets. At this year's expo, which began on Tuesday and concluded on Friday, several companies announced the integration of artificial intelligence with appliances, laptops and even grills. Other major announcements included a Lenovo laptop that has a foldable screen that can expand vertically. There were also new robots, including a Roomba rival with a robotic arm.

Unlike Nvidia's traditional gaming GPUs, Project Digits isn't aimed at consumers. Instead, it targets machine learning researchers, small companies, and universities that want to develop advanced AI but don't have billions of dollars to build massive data centers or buy enough cloud credits.

“There's a big gap for data scientists and machine learning researchers who are actively working, who are actively building something,” Huang said. “You probably don't need a giant cluster. You're just developing early versions of the model, and you're constantly iterating. You can do it in the cloud, but it costs a lot of money.”

Nvidia said the supercomputer will cost about $3,000 when it becomes available in May, and will be available from the company itself as well as some of its manufacturing partners. Project Digits is a placeholder name, suggesting it may change by the time the computer goes on sale, Huang said.

“If you have a good name for it, reach out to us,” Huang said.

Data center sales accounted for 88% of Nvidia's $35 billion global revenue last quarter.

Wall Street is focusing on Nvidia's ability to diversify its business so that it is less dependent on a few customers buying huge AI systems.

“It was a little scary to see Nvidia come out with something so good for so cheap,” Melius Research analyst Ben Retzes wrote in a note this week. He said Nvidia may have “stole the show” due to Project Digits as well as other announcements including graphics cards for gaming, new robot chips and a deal with… Toyota.

Project Digits, which runs Linux and the same Nvidia software used in the company's GPU server clusters, represents a huge increase in capabilities for researchers and universities, said David Bader, director of the Data Science Institute at New Jersey Institute of Technology.

Bader, who has worked on research projects with Nvidia in the past, said the computer appears capable of handling enough data and information to train the largest and most advanced models. He told CNBC Anthropic: Google, Amazon Others “would pay $100 million to build a training supercomputer” to have a system with that kind of capability.

For $3,000, users could soon get a product they can plug into a standard electrical outlet in their homes or offices, Badr said. It's especially exciting for academics, who often leave for private industry in order to have access to larger, more powerful computers, he said.

“Any student who is able to get one of these systems that costs roughly the same as a high-end laptop or gaming laptop will be able to do the same research and build the same models,” Bader said.

The computer could be Nvidia's first step into the $50 billion PC and laptop chip market, Ritzes said.

“It's not too hard to imagine that it would be easy to do everything themselves and let the system run Windows one day,” Ritzes wrote. “But I think they don't want to step on too many toes.”

Huang did not rule out this possibility when asked by Wall Street analysts on Tuesday.

He said MediaTek may be able to sell the GB10 chip to other PC manufacturers in the market. He made sure to leave some mystery in the air.

“Obviously we have plans,” Huang said.

He watches: Nvidia declines due to CES expectations