NVIDIA RTX 6000: R100 Rubin GPUs will go into mass production in late 2025

An American scientist-astronomer Vera Rubin will become the "patron" of the graphic architecture of the hypothetical NVIDIA RTX 6000 video cards. Although the company has not officially announced this name yet, it has been known for some time after the leak by Kopite7kimi. The company recently introduced its most powerful chip Blackwell B100, video cards with which will appear this year.

Astronomer Vera Rubin at work at the Lowell Observatory in Flagstaff, AZ, 1965. Rubin transformed modern physics and astronomy with her observations showing that galaxies and stars are immersed in the gravitational grip of vast clouds of dark matter. pic.twitter.com/C632eshHZP

— World of Science (@scientists_feed) April 20, 2022

Analyst Ming-Chi Kuo reports that the GPU R100 will enter mass production by the end of 2025. The production of finished server solutions will not begin until the first quarter of 2026. This concerns production timelines, not deliveries.

NVIDIA is still finalizing the specifications for the new chip, including the size of the crystal and the packaging. Three options are currently being considered, including the use of CoWoS-L packaging, which is already used by the Blackwell graphic processors. It is reported that Rubin will use the TSMC N3 technology process.

One of the things NVIDIA is focusing on in the Rubin architecture is power consumption. High energy costs can deter customers from acquiring powerful computational devices. The Blackwell B200 graphics processor can consume up to 1000W. The analyst confirms previous statements that Rubin uses stacks of high-speed memory 4x HBM4. There are no details yet on superchips based on the new architecture, similar to GH200 or GB200.