Back to AI intel
重点
趋势
Cognition Becomes First Customer to Run NVIDIA Vera Rubin, Powered by CoreWeave
AI intel briefing
Core summary
One sentence to understand this update
Cognition is the first customer to run NVIDIA Vera Rubin, powered by CoreWeave, achieving approximately 4.8x more token throughput for SWE-2 inference than GB200.
Impact & opportunity
What this could mean
This signals a significant leap in AI inference efficiency with next-gen GPUs; builders should track such hardware advancements for planning future high-performance AI deployments.
Source
View original