Nvidia introduces Blackwell Ultra and Vera Rubin AI chips

Nvidia introduced brand-new chips for structure and releasing expert system versions at its yearly GTC meeting onTuesday

CHIEF EXECUTIVE OFFICER Jensen Huang exposed Blackwell Ultra, a household of chips delivering in the 2nd fifty percent of this year, along with Vera Rubin, the business’s next-generation graphics refining device, or GPU, that is anticipated to deliver in 2026.

Nvidia’s sales are up greater than sixfold given that its service was changed by the launch of OpenAI’s ChatGPT in late 2022. That’s due to the fact that its “big GPUs” have a lot of the marketplace for creating sophisticated AI, a procedure called training.

Software programmers and capitalists are carefully viewing the business’s brand-new chips to see if they supply sufficient extra efficiency and effectiveness to persuade the business’s largest end consumers– cloud firms consisting of Microsoft, Google and Amazon— to proceed investing billions of bucks to develop information facilities based around Nvidia chips.

“This last year is where almost the entire world got involved. The computational requirement, the scaling law of AI, is more resilient, and in fact, is hyper-accelerated,” Huang stated.

Tuesday’s statements are likewise an examination of Nvidia’s brand-new yearly launch tempo. The business is making every effort to introduce brand-new chip family members on an every-year basis. Before the AI boom, Nvidia launched brand-new chip designs every various other year.

The GTC meeting in San Jose, California, is likewise a program of toughness forNvidia

The occasion, Nvidia’s 2nd in-person meeting given that the pandemic, is anticipated to have 25,000 participants and thousands of firms talking about the methods they make use of the business’s equipment for AI. That consists of Waymo, Microsoft and Ford, to name a few. General Motors likewise introduced that it will certainly make use of Nvidia’s solution for its next-generation cars.

The chip style after Rubin will certainly be called after physicist Richard Feynman, Nvidia stated on Tuesday, proceeding its custom of calling chip family members after researchers. Nvidia’s Feynman chips are anticipated to be readily available in 2028, according to a slide shown by Huang.

Nvidia will certainly likewise display its various other product or services at the occasion.

For instance, Nvidia introduced brand-new laptop computers and desktop computers utilizing its chips, consisting of 2 AI-focused Computers called DGX Spark and DGX Station that will certainly have the ability to run big AI versions such as Llama or DeepSeek. The business likewise introduced updates to its networking components for linking hundreds or countless GPUs with each other so they function as one, along with a software called Dynamo that aids individuals obtain one of the most out of their chips.

Nvidia anticipates to begin delivering systems on its next-generation GPU household in the 2nd fifty percent of 2026.

The system has 2 primary parts: a CPU, called Vera, and a brand-new GPU style, calledRubin It’s called after astronomer Vera Rubin.

Vera is Nvidia’s first custom CPU design, the company said, and it’s based on a core design they’ve named Olympus. 

Previously when it needed CPUs, Nvidia used an off-the-shelf design from Arm. Companies that have developed custom Arm core designs, such as Qualcomm and Apple, say that they can be more tailored and unlock better performance.

The custom Vera design will be twice as fast as the CPU used in last year’s Grace Blackwell chips, the company said. 

When paired with Vera, Rubin can manage 50 petaflops while doing inference, more than double the 20 petaflops for the company’s current Blackwell chips. Rubin can also support as much as 288 gigabytes of fast memory, which is one of the core specs that AI developers watch.

Nvidia is also making a change to what it calls a GPU. Rubin is actually two GPUs, Nvidia said. 

The Blackwell GPU, which is currently on the market, is actually two separate chips that were assembled together and made to work as one chip.

Starting with Rubin, Nvidia will say that when it combines two or more dies to make a single chip, it will refer to them as separate GPUs. In the second half of 2027, Nvidia plans to release a “Rubin Next” chip that combines four dies to make a single chip, doubling the speed of Rubin, and it will refer to that as four GPUs.

Nvidia said that will come in a rack called Vera Rubin NVL144. Previous versions of Nvidia’s rack were called NVL72.

Nvidia also announced new versions of its Blackwell family of chips that it calls Blackwell Ultra.

That chip will be able to produce more tokens per second, which means that the chip can generate more content in the same amount of time as its predecessor, the company said in a briefing.

Nvidia says that means that cloud providers can use Blackwell Ultra to offer a premium AI service for time-sensitive applications, allowing them to make as much as 50 times the revenue from the new chips as the Hopper generation, which shipped in 2023.

Blackwell Ultra will come in a version with two paired to an Nvidia Arm CPU, called GB300, and a version with just the GPU, called B300. It will also come in versions with eight GPUs in a single server blade and a rack version with 72 Blackwell chips.

The top four cloud companies have deployed three times the number of Blackwell chips as Hopper chips, Nvidia said.

China’s DeepSeek R1 model may have scared Nvidia investors when it was released in January, but Nvidia has embraced the software. The chipmaker will use the model to benchmark several of its new products.

Many AI observers said that DeepSeek’s model, which reportedly required fewer chips than models made in the U.S., threatened Nvidia’s business.

But Huang said earlier this year that DeepSeek was actually a good sign for Nvidia. That’s because DeepSeek uses a process called “reasoning,” which requires more computing power to provide users better answers. 

The new Blackwell Ultra chips are better for reasoning models, Nvidia said. 

It’s developed its chips to more efficiently do inference, so when new reasoning models require more computing power at the time of deployment, Nvidia’s chips will be able to handle it.

“In the last 2 to 3 years, a major breakthrough happened, a fundamental advance in artificial intelligence happened. We call it agentic AI,” Huang said. “It can reason about how to answer or how to solve a problem.”

WATCH: Nvidia kicks off its GTC Conference: The Committee debate how to trade it