Nvidia announces “Rubin Ultra” and “Feynman” AI chips for 2027 and 2028
SMRTR summary
Nvidia unveiled new AI-accelerating GPUs, including Vera Rubin, set for release in 2026. The Vera Rubin GPU, featuring a custom CPU and tens of terabytes of memory, promises significant performance improvements for AI tasks, delivering 50 petaflops of FP4 inference performance per chip.
SMRTR provides this summary for quick context. The original article belongs to Daily.dev.
Read the original article