NVIDIA has just confirmed its next-gen Rubin GPU architecture while announcing Rubin & Blackwell GPUs along with the latest Vera CPU.
As a surprise announcement, NVIDIA's CEO, Jensen Huang, revealed their next GPU architecture codenamed Rubin which is named after American astronomer, Vera Rubin, who made significant contributions to the understanding of dark matter in the universe while also pioneering work on galaxy rotation rate. Although NVIDIA just revealed its Blackwell platform, it looks like NVIDIA is accelerating its roadmap, offering a new GPU product each year as we reported recently.
But let's start with Blackwell first, while the first iteration of Blackwell GPUs (B100/B200) will come to data centers later this year, NVIDIA also plans to release a supercharged version that would feature 12Hi memory stacks across 8 sites versus the 8Hi memory stacks across 8 sites on existing products. This chip is expected to launch in 2025.
Then soon after Blackwell, NVIDIA will release its next-gen Rubin GPUs. The NVIDIA Rubin R100 GPUs will be part of the R-series lineup & are expected to be mass-produced in the fourth quarter of 2025 while systems such as DGX and HGX solutions are expected to be mass-produced in the first half of 2026. According to NVIDIA, Rubin GPUs and the respective platform will be available by 2026 followed by an Ultra version in 2027. NVIDIA also confirms that Rubin GPUs will utilize HBM4 memory.
It is expected that NVIDIA's Rubin R100 GPUs will use a 4x reticle design (versus 3.3x of Blackwell) and will be made using the TSMC CoWoS-L packaging technology on the N3 process node. TSMC recently laid out plans for up to 5.5x reticle size chips by 2026 which would feature a 100x100mm substrate and allow for
Read more on wccftech.com