Cognition First to Deploy NVIDIA Vera Rubin via CoreWeave

Core summary

One sentence to understand this update

Cognition is announced as the first customer to run NVIDIA Vera Rubin GPUs, powered by CoreWeave, achieving 4.8x more token throughput than GB200 on SWE-2 inference.

Impact & opportunity

What this could mean

This adoption showcases the bleeding edge of AI inference hardware, signaling massive performance gains for complex AI models and setting a new benchmark for computational capabilities.