Loading…
Agent Bricks: Data + AI Summit 2026
Hanlin Tang, Kasey Uhlenhuth, Akhil Gupta, Patrick Wendell
- Source
- Databricks
- Published
- Added to Yomu
Summary
At Data + AI Summit 2026, Databricks announced Agent Bricks as a comprehensive developer platform for building and operating agents, extending a product launched the previous year. The announcement frames the core agent loop as only 1% of the work, with token capacity, deployment, security, evaluation, monitoring, context, and sharing forming the remaining infrastructure burden. Agent Bricks addresses choice, context, and control through support for multiple proprietary, open-source, and custom models, any agent harness, MCP-connected data, Genie Ontology, managed memory, document intelligence, sandboxes, and governed tools. Unity AI Gateway adds catalogs, fine-grained access controls, budgets, traffic routing, contextual policies, monitoring, and registry support for agents, tools, and models. Databricks says more than 100,000 agents have been built and customers including AstraZeneca, 7-Eleven, Fox Corporation, and Block have shipped agents on the platform.
Context
The announcement says developers have become stuck building infrastructure rather than agents because the core agent loop represents only 1% of the work. It identifies three platform challenges: providing model choice, retrieving reliable business context from complex data estates, and controlling security and costs for highly privileged agents.
Approach / What changed
Agent Bricks combines model access, support for multiple agent harnesses, MCP-connected data, Genie Ontology, managed memory, document intelligence, secure sandboxes, governed tools, and Unity AI Gateway. The platform connects agent outputs and traces with enterprise data governance through Unity Catalog, contextual policies, monitoring, budgets, and traffic routing.
Takeaways
- Agent Bricks supports frontier proprietary, open-source, and custom models, including OpenAI, Anthropic, Gemini, Qwen, and newly added Kimi, allowing developers to balance behavior, latency, and cost.
- MCP support in Unity Catalog connects agents to external sources such as Google Drive, JIRA, Slack, and GitHub, while Genie Ontology adds business semantics for search and analysis.
- Unity AI Gateway provides catalogs, fine-grained access controls, budget enforcement, intelligent routing, monitoring, contextual policies, and registry support for agents, tools, and models.