(Bloomberg) — Shares of Nvidia Corp. fell Tuesday after Chief Executive Officer Jensen Huang’s wide-ranging product presentation failed to catapult the artificial intelligence chipmaker to new heights.
Most Reads on Bloomberg
New York stocks fell 6.2% to $140.14, the biggest one-day decline in four months. Nvidia’s recent announcement offered some optimism about the company’s long-term prospects, but it didn’t offer as much near-term upside as some investors had been seeking. “Nvidia’s announcement today is an important but long-term announcement,” Stifel Financial Corp. said in a report.
Subscribe to the Bloomberg Daybreak podcast on Apple, Spotify, or wherever you listen.
Huang took the stage at an arena in Las Vegas to kick off the CES trade show on Monday, showcasing the new lineup and laying out his vision for how AI will spread throughout the economy. The company wants its products to be at the center of a future technological world with 1 billion humanoid robots, 10 million automated factories, and 1.5 billion self-driving cars and trucks.
As companies rush to deploy new AI computing equipment, interest in Nvidia products and Huang’s predictions has exploded. The CEO spent more than 90 minutes explaining Nvidia’s products and strategy to an audience of hundreds, including Toyota Motor Corp., whose shares rose more than 3%. and partnerships with MediaTek Inc.
Before Tuesday’s retreat, Nvidia’s stock had more than tripled in the past 12 months. Asian suppliers including Taiwan Semiconductor Manufacturing Co. and Hon Hai Precision Industry Co. also surged on optimism about Nvidia’s prospects.
Chief Financial Officer Colette Kress said at a separate event that the AI transition will continue to drive growth over the next decade. She made the announcement Tuesday at the JPMorgan Chase & Co. conference held concurrently with CES. “He will be with us for the next 10 years and beyond,” he said in the chat. “We still have a lot of growth opportunities in the future.”
At Huang’s presentation Monday, he also delivered the news to his traditional audience: gamers. Nvidia is releasing an update for its GeForce GPUs (short for graphics processing units), built with the same Blackwell design it uses in its AI accelerators, Huang said.
The new GeForce 50 series cards will leverage Blackwell’s capabilities to deliver a more realistic experience for computer gamers, the company said. Traditional graphics chips build images by calculating the shade of each pixel in a photo, but new technologies will rely more on AI to predict what the next frame will look like.
“GeForce made AI accessible to the masses, and now AI is back with GeForce,” Huang said in his presentation.
The flagship RTX 5090 model is scheduled to launch later this month for $1,999, with less powerful cards to follow later. Nvidia says the $549 RTX 5070 will debut in February with better performance than the previous lineup’s top model, the RTX 4090.
As recently as 2022, gaming was Nvidia’s biggest source of revenue. Now the chipmaker’s data center operations are much larger. The company plans to donate more than $100 billion this year as its accelerator chips are highly regarded by the world’s largest technology companies. The next step is to diversify Nvidia’s revenues by rolling out hardware and software to more enterprises and government agencies.
Huang announced that Toyota, the world’s largest automaker, is now a customer of Nvidia’s self-driving AI products and will use its Drive chips and software. Toyota shares in Tokyo extended gains following the announcement.
NVIDIA says extending AI into more of the physical world will transform an industry worth $50 trillion. But this move will also bring challenges. Robots and cars require software that can handle the complexities of real life in a safe way. The company created NVIDIA Cosmos to help make robots smarter and produce fully self-driving cars, Huang said.
Cosmos technology can generate video from inputs such as text. Those videos then become the basis for virtual training, helping reduce reliance on costly and time-consuming physical experiments. Generated video can be searched and refined to repeatedly test critical but infrequent events, such as collisions with emergency vehicles.
Nvidia is also working with Uber Technologies Inc. to develop self-driving technology. The millions of trips Uber handles every day will provide massive data for training AI models.
Nvidia said mass-market automakers will move toward using one computer and operating system across their entire model lineup, rather than splitting systems by vehicle class. The company believes this transition will pave the way for broader availability of the chip designer’s comprehensive offerings. To accelerate this, Nvidia has received product certification from government transportation safety agencies.
Nvidia is now also offering a desktop PC called Project Digits. The company is equipping its tiny $3,000 device with a single Grace Blackwell Superchip (a combination of a central processor and a graphics chip), with large memory and fast connectivity. The idea is to provide developers with hardware that can run large-scale AI models that are difficult for current laptops to handle.
The new system, developed in collaboration with Taiwan’s MediaTek, runs a version of the Linux operating system and is not designed for everyday use. Instead, it is intended to help AI developers work locally when connecting to the cloud or using traditional computers is not practical or possible.
Nvidia chose MediaTek to help make Digits’ main chip because of the company’s skills in making low-power semiconductors. The Taiwanese company will also offer products containing this technology to other customers, Huang said.
Asked if Project Digits is a sign that Nvidia wants to enter the PC market more broadly, Huang said the machines are intended for AI developers and students. But he also hinted that Nvidia has greater interest in this space.
“Obviously we have a plan,” he said during a briefing with financial analysts. “I guess I’ll have to wait to tell you about that.”
