Latham & Watkins Fine-Tunes NVIDIA Nemotron 3 on Private On-Premises GPU Servers

Latham's engineers use the firm's private H200 GPU servers to fine-tune NVIDIA's Nemotron 3 open-weight model for the firm's legal requirements -- among the first major law firms publicly known to do this at scale -- with a technology team of 900, ~100 dedicated to AI, and plans to evaluate next-gen NVIDIA Blackwell/Vera Rubin systems.