At the annual CES (Consumer Technology Association) event in Las Vegas, Nvidia founder and CEO Jensen Huang unveiled a supercomputer based on its next-generation AI processor, Rubin. The processor combines with five other chips with an Israeli connection – its new core processor Vera, the NVLink 6 Switch, the ConnectX-9 SuperNIC, the BlueField-4 DPU, and the Spectrum-6 Ethernet Switch, developed at the company’s center in Yokne’am.
The Rubin graphics processing unit represents the next stage of Nvidia chips, after the Blackwell Ultra (B300), and it was revealed last year along with the generation to come, Feynman. Now, however, Huang has revealed the architecture of the data centers in which the chip will be installed, and explained its dependence on the Israeli development center: in order to make the server efficient and fast for training AI models and as the basis for AI agents. Each Rubin GPU will be integrated into a new supercomputer together with a new core processor and four communications processors developed in Irael.
Huang heaped praise on Nvidia’s communications activity based on Israeli company Mellanox, saying that thanks to it Nvidia was the biggest communications company in the world, and that its communications chips had revolutionized the way processors communicated with one another, saving billions of dollars in the construction of new server farms.
Huang particularly mentioned Nvidia’s support for open inference models, saying “2025 was the year of the open models”, as part of its strategy of selling products such as chips, servers and software alongside free models. Among others, Nvidia mentioned support for the new video model from Israeli company Lightricks, LTX-2.
The Rubin chip is in production, but will not be available commercially until the second half of this year.
The “Vera Rubin” combination is after the American astronomer of that name whose research into the rotation of galaxies supported the hypothesis of the existence of dark matter.
Some of the development of the Vera core processor took place in Israel. There are many experts on core processor development in Israel thanks to the fact that Intl’s flagship products in this area were developed in the country.
Huang also presented the next stage in Nvidia’s collaboration with Mercedes-Benz. The two companies have developed a completely autonomous vehicle based on Nvidia’s dedicated language model for automobiles, Alpamayo. The autonomous Mercedes-Benz cars will be launched in the US in the first quarter of this year, and later in the year in Europe and Asia.
Published by Globes, Israel business news – en.globes.co.il – on January 6, 2026.
© Copyright of Globes Publisher Itonut (1983) Ltd., 2026.
-1024x683.jpg "SCOTUS: Florida Can’t Sue Blue States Over Illegal Alien Drivers")
