(Webinar) Databricks AI Gateway + Open Model Serving

​Most platform teams we speak with in the region are running the same experiment at once: several LLM providers, a growing number of GenAI apps and coding agents, and no single place to see who is calling what, what it costs, or what data went out the door.

​Join us for a 60-minute session on how to put every model call Databricks-hosted, open source, or your existing third-party provider behind one governed endpoint, and how open model serving changes the economics once you do.

​What we will cover:

• ​The four ways AI sprawl breaks platform teams: model sprawl, no cost visibility, compliance exposure, and no failover.

• ​What you can turn on today with AI Gateway on your serving endpoints: permissions and rate limits, payload logging and usage tracking into Unity Catalog, guardrails, and automatic fallbacks.

• ​Open model serving in practice with open model catalog including GLM5.2, Kimi3, GPT OSS, Llama, Qwen and Gemma, and how to size dedicated capacity for a production app.

• ​A live demo: enable governance on an endpoint, watch usage and payload data land in Unity Catalog, then swap a frontier model for an open one without touching your governance setup.

• ​Where this is heading with Unity AI Gateway one control plane across models, agents and MCP servers and the migration path, which for existing endpoints is a URL change, not a rebuild.