Nvidia revealed its latest Vera Rubin artificial intelligence platform at the Consumer Electronics Show in Las Vegas, positioning the new product as central to maintaining its lead in the chips that power the artificial intelligence boom. Chief executive Jensen Huang’s keynote was billed as an early highlight of the global tech gathering, underscoring how closely the industry is watching Nvidia’s moves as the world’s most valuable company seeks to extend its dominance in data center hardware.
The California-based firm said Vera Rubin, first announced in September, is designed to secure its grip on the artificial intelligence chip business, where it currently holds an estimated 80 percent of the global market for artificial intelligence data center chips. The new architecture, named after American astronomer Vera Rubin, is described by Nvidia as a profound shift from its previous Blackwell generation, which launched in late 2024, and Rubin-based products would be available from partners in the second half of 2026. Dion Harris, Nvidia’s director of data center and high-performance computing, explained that the platform comprises “six chips that make one artificial intelligence supercomputer” to serve the most demanding models while aiming to drive down the cost of delivering advanced intelligence.
Nvidia is rolling out Rubin against a backdrop of intensifying competition on several fronts, including traditional rivals AMD and Intel, which are pushing aggressively into artificial intelligence accelerators. The company also faces pressure from its own largest customers, as Google, Amazon, and Microsoft increasingly develop in-house chips to lessen their reliance on Nvidia, with Google’s latest artificial intelligence model, Gemini 3, trained without Nvidia technology. China is racing to develop domestic alternatives in response to United States export restrictions on Nvidia products, while concerns grow about the rising energy demands of artificial intelligence; Nvidia says the Rubin product will run five times more efficiently than its previous offerings. Since ChatGPT’s release in 2022, Nvidia has been refreshing its lineup at a rapid pace, raising questions over how long model builders can afford to continuously upgrade to the latest hardware generations.
