Nvidia has upped the ante by releasing its new line of GPUs. It has come as a part of their launch of the new GPU architecture codenamed Rubin.
It was announced at the Computex 2024, as Rubin became the successor to the Blackwell series. The new series is named after Vera Florence Cooper Robin, the pioneer astronomer who proved the presence of dark matter.
The new GPU is set to be released in 2026, along with Rubin’s Ultra GPUs. That comes along with the new Vera, which is slated as a powerful competitor to Intel processors.
On top of that, get ready to see the combo Vera Rubin into a super chip, combining GPU and CPU.
When is Rubin releasing?
The first versions, named B100 and B200, will be available in data centers later this year. These GPUs will improve data processing and computing power.
NVIDIA also plans to release a more powerful version of the Blackwell GPUs. This version will have more memory layers, with 12 layers at each of the 8 sites on the chip, compared to the current 8 layers.
This upgraded version of the Blackwell GPU is expected to be available in 2025.
The Tech
NVIDIA’s upcoming Rubin R100 GPUs are expected to use a 4x reticle design, compared to Blackwell’s 3.3x. They will be made using advanced TSMC packaging technology on the N3 process node.
TSMC has plans for even larger chips with a 5.5x reticle size by 2026, which would allow for more memory bandwidth.
Additionally, the company is considering a new design with a reticle size greater than 8x in a larger 120x120mm package.
However, for now, the Rubin GPUs are expected to realistically have a 4x reticle size.
