Databricks makes AI Runtime generally available for GPU model training

Databricks says its AI Runtime is now generally available, giving developers serverless Nvidia GPUs for training and fine-tuning open models. The release supports Nvidia NeMo, NGC containers and Nemotron cookbooks, alongside Docker environments and private container registries. Databricks manages the underlying clusters and networking for multi-node training. Unity Catalog controls access to training data, while MLflow tracks experiments and GPU utilization. Teams can deploy evaluated models through Databricks Model Serving; developers still review and refine code suggested by Genie Code before execution.