Databricks and NVIDIA are expanding their partnership to bring NVIDIA GPUs, the new Vera CPU, and agentic AI tooling into the Databricks platform. Key highlights include: AI Runtime (AIR) for GPU-accelerated training and fine-tuning on governed enterprise data; Model Serving powered by NVIDIA hardware and Triton Inference Server for production inference; the NVIDIA Vera CPU designed specifically for agentic workloads, RL, and CPU-based analytics to eliminate bottlenecks in agent orchestration; NVIDIA Agent Toolkit hosted on Databricks Apps for building agentic workflows; and domain-specific NVIDIA libraries (BioNeMo, Parabricks, Isaac Sim, Omniverse, etc.) integrated directly into the Databricks platform for industries like healthcare, genomics, and robotics.
