In This Article:
Update: March 18, 2:37 p.m. ET: Huang announced the Vera Rubin GPU, which has twice the performance of its Grace chips with more memory and bandwidth. Rubin is expected in the second half of next year.
Huang said the following Rubin Ultra is slated for the second half of 2027.
Update: March 18, 2:35 p.m. ET: Huang says Blackwell Ultra NVL72 is coming in the second half of the year. Earlier in his keynote, Huang said Blackwell is in full production.
Blackwell Ultra has one-and-a-half times more memory and more bandwidth.
Update: March 18, 2:16 p.m. ET: Huang announced Nvidia Dynamo — “essentially the operating system of an AI factory” — for the future of AI agents.
All eyes will be on Nvidia (NVDA) chief executive Jensen Huang today as he shares more about the chipmaker’s next artificial intelligence chips.
In February, Huang said he would discuss the chipmaker’s Blackwell Ultra AI chip, next-generation Vera Rubin platform, and plans for following products at the annual GPU Technology Conference, also known as the GTC.
Huang said the company has “some really exciting things to share” at the developer conference about enterprise and agentic AI, reasoning models, and robotics.
The chipmaker introduced its highly anticipated Blackwell AI platform at last year’s GTC, which has “successfully ramped up” large-scale production, and made “billions of dollars in sales in its first quarter,” Huang said on the company’s fiscal fourth quarter earnings call.
Analysts at Bank of America (BAC) said in a note last week that they “expect Nvidia to present attractive albeit well-expected updates on Blackwell Ultra,” with a focus on inferencing for reasoning models, which major firms such as OpenAI and Google are racing to develop.
The analysts are also anticipating more information on Nvidia’s next-generation networking technology, and what the chipmaker’s long-term opportunities are in autonomous cars, physical AI such as robotics, and quantum computing.
Kevin Cook, senior stock strategist at Zacks Investment Research, said Nvidia is likely to talk more about its Project DIGITS personal AI supercomputer at the GTC, in comments shared with Quartz.
“I’ve been saying for years that Nvidia GPU systems are ‘like the iPhone cycle, but better,’” Cook said. “What I mean is that enterprises will pay for the latest, better, and faster from Jensen, but they can still use their old hardware, too, because it’s all integrated and updated by CUDA software.”
Cook said he thinks DIGITS will give smaller developers the ability to build their own physical AI models and agentic AI applications.