On stage at Nvidia’s GTC 2025 convention in San Jose on Tuesday, Nvidia CEO Jensen Huang introduced a slew of latest GPUs coming down the corporate’s product pipeline over the subsequent a number of months.
Perhaps essentially the most vital was Vera Rubin. Vera Rubin, which is ready to be launched within the second half of 2026, will function tens of terabytes of reminiscence and a customized Nvidia-designed CPU referred to as Vera. Vera Rubin delivers substantial efficiency uplifts in comparison with its predecessor, Grace Blackwell, Nvidia claims, notably on AI inferencing and coaching workloads.
When paired with Vera, Rubin can handle as much as 50 petaflops whereas doing inference, greater than double the 20 petaflops for the corporate’s present Blackwell chips, Nvidia says. And Vera is about twice as quick on the CPU utilized in Grace Blackwell.
Rubin will likely be adopted by Rubin Ultra within the second half of 2027, a group of 576 Vera Rubin GPUs, Huang stated.
In the nearer time period — the second half of 2025 — Nvidia will launch Blackwell Ultra. A single Ultra chip will supply the identical 20 petaflops of AI efficiency as Blackwell, however with 288GB of reminiscence — up from 192GB in vanilla Blackwell.
On the far horizon is Feynman. Huang gave few particulars about Feynman’s structure, named after American theoretical physicist Richard Feynman, save that it contains a Vera CPU. It’s additionally unclear when the primary Feynman GPUs will arrive.