Nvidia might release its next-generation AI chip R100 for mass production in the last quarter of 2025, with the system/rack solution going into manufacturing in the first half year of 2026.
The speculations about Nvidia’s next-generation AI chip R100 came to light after analyst Ming-Chi Kuo shared it via Medium.
Nvidia is one of the leading technology companies working on graphics processing units (GPUs) and brings extreme focus to AI, HPC, robotics, and more.
The new chip is predicted to use Taiwan Semiconductor Manufacturing Company’s (TSMC’s) N3 node and its CoWoS-L packaging, found in the Blackwell variant B100 chip.
R100 is expected to be composed of four smaller chips, known as reticles, that create one bigger chip. It should be equipped with eight units of HBM4 memory.
Nvidia’s central processing unit (CPU) – GR200 Grace – is expected to be manufactured with TSMC’s N3 process. Now, it uses TSNC’s N5 technology for the latest GH200 and GB200 Grace superchip.
With the release of R-Series chips, the company hopes to work on “improving power consumption in addition to enhancing AI computing power,” notes Kuo.
The speculations about the chip maker’s plans came just a few months after the launch of Nvidia’s B200 Blackwell chip.
Your email address will not be published. Required fields are markedmarked