--- title: "Databricks Data + AI Summit 2026: Key Announcements" url: "https://atlan.com/know/ai-agent/databricks/databricks-data-ai-summit-2026-announcements/" description: "A full recap of Databricks Data + AI Summit 2026: Genie One, Genie Ontology, Agent Bricks, Unity Catalog Metrics, Unity AI Gateway, Lakebase, and LTAP." author: "Emily Winks" author_role: "Data Governance Expert" published: "2026-06-19" updated: "2026-06-19T00:00:00.000Z" --- --- Databricks Data + AI Summit 2026 made one argument louder than any single product launch: agents are only as good as the context they are grounded in. Across four days at Moscone, the headline releases (Genie One, Genie [Ontology](https://atlan.com/know/what-is-ontology-in-ai/), Agent Bricks, Unity Catalog Metrics, Unity AI Gateway) all pointed at the same idea Databricks CEO Ali Ghodsi put on the keynote stage. The agentic data stack is real and shipping, and the open question for every enterprise is how to carry that grounding across a real estate that spans far more than the lakehouse. This hub recaps every major announcement and where the [context layer](https://atlan.com/know/what-is-context-layer/) fits. --- ## Quick facts | Attribute | Detail | |---|---| | What it is | Databricks Data + AI Summit 2026 (DAIS 2026), the company's flagship data and AI conference | | When and where | June 15-18, 2026, Moscone Center, San Francisco; ~30,000 in-person attendees | | Theme | "Apps and agents that work": agentic AI from lab to enterprise; AI cost in focus | | Headline launches | Genie One (GA), Genie Ontology, Agent Bricks, Unity Catalog Metrics, Unity AI Gateway, LTAP, Lakehouse RT | | Who it is for | Data leaders, data architects, and platform teams running AI on Databricks | | Cross-cutting message | Agent quality is a context problem; governance is table stakes for agents | | How Atlan complements it | Atlan is the context layer that extends governed, live context across the whole estate and serves it back to Databricks Genie | --- ## The story of the Summit: agents move from lab to enterprise The framing for Data + AI Summit 2026 was set in the opening keynote. CEO Ali Ghodsi told the room that AI does not have an intelligence problem, it has a context problem, and illustrated it plainly: if a CFO cannot get AI to explain why margins changed, or a sales leader cannot get it to surface the next upsell, that is a context gap, not a model gap. Across the week, "apps and agents that work" was the recurring phrase, with the subtext that [agentic AI](https://atlan.com/know/what-is-agentic-ai/) is moving out of the lab and into production at companies like PepsiCo, Mastercard, and AstraZeneca. Two other themes ran through every keynote. The first was cost: Ghodsi was candid that agentic consumption "is going to get extremely expensive," and CTO Matei Zaharia echoed the need for new layers above the model and the agent harness to control it. The second was governance as table stakes for agents, where models, agents, tools, and MCP services all need to be governed the way data already is. Each product below maps to one of these three ideas: better agents, governed runtime, and the data infrastructure underneath. Is your data estate ready for the agentic stack? The Summit made agent grounding the headline. Check how ready your context foundation is for AI agents in a five-minute assessment. Assess Your Readiness --- ## Announcements at a glance This table indexes the major Data + AI Summit 2026 releases by group, with what each one is and its announced status. Each capability is covered in more depth in the sections that follow. | Capability | What it is | Status | |---|---|---| | Genie One | Agentic coworker for every team, across structured and unstructured data | GA | | Genie Ontology | Live, self-improving context layer that grounds Genie answers | Launched at Summit | | Genie Agents | Reusable agents built from Genie conversations, governed by Unity Catalog | GA | | Genie Code | Agentic data and ML workspace; imports Tableau and Power BI workbooks | GA, expanded | | Agent Bricks | Agent-building platform; supports Claude Code SDK, LangGraph, CrewAI, OpenAI Agent SDKs | Expanded | | Omnigent | Open-source meta-harness for combining and controlling coding agents | Managed beta | | Unity Catalog Metrics | Governed, reusable business KPIs queryable from SQL, BI, APIs, and agents | Available | | Unity AI Gateway | Runtime governance for models, agents, tools, MCPs; spend caps and guardrails | Announced | | Business Glossary | Authoritative definitions of business concepts in Unity Catalog | Preview soon | | Domains | Business-aligned organization of data and AI assets | Public Preview | | Catalog Federation | Zero-copy federated governance across accounts, regions, and clouds | Announced | | Automatic Identity Management | Identity sync (Entra GA, Okta preview) | GA / preview | | Lakebase | Serverless Postgres on open object storage; branching, search, DR | GA, expanded | | LTAP | Lake Transactional/Analytical Processing; unifies OLTP and OLAP on one copy | Coming soon (Lakebase) | | Lakehouse RT | Real-time analytics on governed tables via the new Reyden engine | Announced | --- ## Agents: Genie One, Genie Ontology, Agent Bricks, and Omnigent The agent group was the heart of the event. **[Genie One](https://atlan.com/know/ai-agent/databricks/genie-one/)**, now generally available, is positioned as an agentic coworker that helps business teams in marketing, finance, and sales automate and orchestrate work across structured and unstructured data. It goes beyond answering questions to produce documents, reports, and artifacts, and it ships across web, iOS, and Android with scheduling, alerts, and MCP tool integration. **Genie Agents** (GA) turn Genie conversations into reusable workflows, and **Genie ZeroOps** runs autonomous background monitoring. Underneath Genie One sits **[Genie Ontology](https://atlan.com/know/ai-agent/databricks/genie-ontology/)**, which Databricks describes as a live, self-improving context layer that automatically extracts and continuously updates business knowledge from Databricks and connected workplace apps such as Google Drive, Jira, Slack, Confluence, and SharePoint. The promise is more accurate answers, faster, at lower token cost, by grounding responses in governed data rather than document embeddings. This is the announcement with the most direct vocabulary overlap with the context layer, and it makes the cross-ecosystem question concrete, which we cover below. **Agent Bricks**, the developer platform, was expanded to support the Claude Code SDK, LangGraph, Agno, CrewAI, and OpenAI Agent SDKs, with horizontal autoscaling through Databricks Apps; Databricks cited 100K+ agents built on the [Agent Bricks](https://atlan.com/know/ai-agent/databricks/agent-bricks/) platform. **Genie Code** gained a full-page command center, production data and ML engineering upgrades, scheduled tasks, and the ability to import Tableau and Power BI workbooks into AI/BI dashboards connected to Unity Catalog metric views. Finally, **Omnigent** is a new open-source (Apache 2.0) meta-harness, a "harness of harnesses" in Zaharia's words, for combining and controlling coding agents across frameworks with cost budgets and shared sessions; a managed version runs on Databricks in beta, governed by Unity AI Gateway. | Agent release | What it does | Status | |---|---|---| | Genie One | Agentic coworker producing reports and artifacts across data types | GA | | Genie Ontology | Live context layer grounding Genie in governed business knowledge | Launched at Summit | | Agent Bricks | Multi-framework agent build platform (Claude, LangGraph, CrewAI, OpenAI) | Expanded | | Genie Code | Agentic data and ML workspace; BI workbook import | GA, expanded | | Omnigent | Open-source meta-harness for coding agents | Managed beta | --- ## Unity Catalog and governance: Metrics, AI Gateway, Federation The governance group reframed Unity Catalog as the grounding and the guardrail for agents. **[Unity Catalog Metrics](https://atlan.com/know/ai-agent/databricks/unity-catalog-metrics/)** lets teams define KPIs such as revenue, churn, active users, and margin once as governed, reusable objects, then query them consistently from SQL, BI tools, APIs, and agents. It adds richer semantic modeling with multi-fact relationships, level-of-detail expressions, parameterization, and materialization. Alongside it, **Business Glossary** (preview soon) and **Domains** (public preview) give people and agents a shared, governed source of meaning. **[Unity AI Gateway](https://atlan.com/know/ai-agent/databricks/unity-ai-gateway/)** extends governance from data and AI assets to the runtime interactions between models, agents, MCP services, skills, and enterprise tools. It provides visibility into AI spend across tools and models, enforces hard spend caps, routes workloads intelligently to balance quality and cost, and applies guardrails for PII and prompt injection. This is the direct answer to the cost theme Ghodsi raised. **Catalog Federation** extends Unity Catalog governance across accounts, regions, and clouds with a zero-copy federated path that brings compute to the data under native governance, and **Automatic Identity Management** syncs identity (Entra GA, Okta preview). See the Context Layer in action Watch how a governed context layer delivers consistent business meaning to every agent, across Databricks and the rest of your stack. Watch Context Layer Live --- ## Data infrastructure: Lakebase, LTAP, Lakehouse RT, and Free Edition The third group is the data foundation rebuilt for how agents read, loop, and write. **Lakebase**, serverless Postgres on open object storage, is now GA and reported at 12 million database launches per day, with cross-cloud and cross-region disaster recovery, git-style branching and snapshots, autonomous operations, and a Lakebase Search beta for hybrid vector and full-text retrieval. **LTAP** (Lake Transactional/Analytical Processing) builds on Lakebase to unify transactions, analytics, and streaming on a single governance model and storage layer, storing Postgres-native transactional data in Delta and Iceberg format from the point of write, which removes the ETL pipelines that have historically connected operational and analytical systems. **Lakehouse RT** is a real-time analytics product powered by a new compute engine called Reyden, delivering sub-100ms latency at high query volume directly on governed Delta Lake and Iceberg tables, collapsing what used to require a separate real-time serving tier. **Databricks Free Edition** (500K+ users) added five products including Genie Code, serverless GPUs, Lakebase, Agent Bricks, and Lakeflow Designer, lowering the on-ramp for builders. | Infrastructure release | What it does | Status | |---|---|---| | Lakebase | Serverless Postgres with branching, search, and DR | GA, expanded | | LTAP | Unifies OLTP and OLAP on one governed copy | Coming soon (Lakebase) | | Lakehouse RT (Reyden) | Sub-100ms real-time analytics on governed tables | Announced | | Free Edition additions | Genie Code, serverless GPUs, Lakebase, Agent Bricks, Lakeflow Designer | Available | --- ## The cross-cutting themes: agents, cost, and context as grounding Read together, the Summit announcements describe one strategy: build a stack where agents are first-class, governed at runtime, and grounded in business meaning. Three themes connect every group. First, agentic AI is moving from lab to enterprise, which is why Genie One went GA and Agent Bricks opened to more frameworks. Second, AI cost is now an executive concern, which is why Unity AI Gateway leads with spend caps and smart routing and why LTAP and Lakehouse RT collapse redundant infrastructure tiers. Third, and most important for this hub, context is the grounding for agent accuracy. Genie Ontology, Unity Catalog Metrics, Business Glossary, and Domains all exist to give agents trustworthy meaning, not just data. As Moor Insights & Strategy observed, agents read, loop, and write differently from software built for people, and the underlying data layer has to change to keep up. That is the right read. It also surfaces the practical question every enterprise leaves the Summit with: governed context inside Databricks is excellent, and enterprises run far more than Databricks. --- ## How the context layer extends the Databricks agentic stack The through-line of Data + AI Summit 2026 was Ghodsi's own line: agent quality is a context problem. Atlan agrees completely, and this is exactly where the context layer is additive to everything Databricks announced. Genie Ontology and Unity Catalog are excellent for grounding agents inside Databricks. The reality for most enterprises is that the context an agent needs also lives in Snowflake, BigQuery, dbt, Tableau, Power BI, Salesforce, and SAP. Atlan is the context layer for AI: the governed infrastructure that delivers enterprise knowledge to every model, every agent, and every team from a single source of truth. The relationship is better together. Atlan layers on top of your existing data stack, and many enterprises run Atlan alongside Databricks Unity Catalog, Snowflake Horizon, or Microsoft Purview, pulling context from all of them into a unified context layer rather than rebuilding from scratch. Atlan ingests Databricks context, enriches it with cross-system lineage and certified semantics, and serves it back to Databricks Genie. Four products carry this: - **[Enterprise Data Graph](https://atlan.com/know/enterprise-data-graph/):** 80+ connectors and column-level lineage form a living graph of assets and relationships across the whole estate, the context that extends beyond the lakehouse. - **[Context Agents](https://atlan.com/know/context-agents/):** AI teammates that auto-generate descriptions, link terms, infer metrics, and propose ontologies. Atlan AI Labs reports 690K+ descriptions generated, 87% rated on par or better than human writing, across 50+ enterprise customers (April 2026). - **[Context Engineering](https://atlan.com/know/what-is-context-engineering/) Studio:** bootstrap, test, and ship context as code, with CI-integrated evals before production, so the definitions Genie reasons over are validated, not assumed. - **Context Lakehouse:** an Iceberg-native, open-format context store that activates via MCP, SQL, and open APIs, keeping context portable rather than locked to any vendor's schema. The payoff is measurable. Atlan AI Labs measured a 5x accuracy improvement in agents grounded in Atlan's context layer, and the same research notes that 83% of AI pilots never reach production because the gap is context, not the model. [AI agents](https://atlan.com/know/ai-agent/what-is-an-ai-agent/) get enterprise context through Atlan's MCP server, SQL interface, and open APIs, the same way Genie consumes governed signals. A [semantic layer](https://atlan.com/know/semantic-layer/) defines what metrics mean, primarily for BI. A context layer is broader: it includes semantic definitions plus lineage, policy rules, decision traces, and runtime delivery for AI agents, and it consumes and enriches the semantic definitions a [semantic layer](https://atlan.com/know/context-layer-vs-semantic-layer/) provides. For Databricks teams specifically, the [context layer for Databricks](https://atlan.com/know/context-layer-for-databricks/) guide and the [Genie Ontology and Atlan context layer](https://atlan.com/know/ai-agent/databricks/genie-ontology-and-atlan-context-layer/) page show how the two fit together, and [adding Atlan context to a Genie space](https://atlan.com/know/ai-agent/databricks/add-context-to-databricks-genie-space/) shows the practical first step. Inside Atlan AI Labs and the 5x accuracy factor See how context engineering drove a 5x accuracy improvement in real customer systems, with experiments, results, and a repeatable playbook for getting agents to production. Download E-Book --- ## Why the Summit was really about context Data + AI Summit 2026 will be remembered as the moment the agentic data stack stopped being a roadmap and became a product line. Genie One shipping GA, Unity Catalog turning metrics and glossary into governed objects, Unity AI Gateway governing the runtime, and LTAP collapsing the pipeline between operational and analytical data all point the same direction: agents that work in production, governed and grounded. The unifying insight, stated from the keynote stage, is that the bottleneck is context. Genie Ontology is Databricks' answer inside the lakehouse, and it is a strong one. The enterprise reality is that context lives everywhere, across warehouses, BI tools, CRMs, and operational systems. That is the role of the [enterprise context layer for AI agents](https://atlan.com/know/context-agents/): to unify governed, live context across the whole estate and serve it back to Genie and every other agent. Not Databricks or Atlan, but Atlan making the Databricks agentic stack work across the entire ecosystem. Book a Demo --- ## FAQs about Databricks Data + AI Summit 2026 1. **When and where was Databricks Data + AI Summit 2026?** Databricks Data + AI Summit 2026 ran June 15-18, 2026 at the Moscone Center in San Francisco, with around 30,000 in-person attendees and tens of thousands more joining virtually. The theme was "apps and agents that work." 2. **What is Genie Ontology?** Genie Ontology is a live context layer that powers Genie. Databricks describes it as a self-improving layer that automatically extracts and continuously updates business knowledge from Databricks and connected workplace apps, delivering more accurate answers, faster, at lower token cost. It grounds answers in governed data rather than document embeddings. 3. **What is the difference between Genie One and Genie Ontology?** Genie One is the agentic coworker, the interface that business teams interact with to automate and orchestrate work. Genie Ontology is the context layer underneath it that supplies the business knowledge Genie One reasons over. Genie One is the agent; Genie Ontology is the context it draws from. 4. **What is Unity Catalog Metrics?** Unity Catalog Metrics lets teams define business KPIs such as revenue, churn, active users, and margin once as governed, reusable objects, then query them consistently from SQL, BI tools, APIs, and agents. It supports multi-fact relationships, level-of-detail expressions, parameterization, and materialization. 5. **What is Unity AI Gateway?** Unity AI Gateway is Databricks' runtime governance layer for enterprise AI. Built on Unity Catalog, it brings models, agents, tools, and MCP services under one governance layer, providing visibility into AI spend, enforcing hard spend caps, smart routing across models, and guardrails for PII and prompt injection. 6. **What is LTAP and Lakehouse RT?** LTAP (Lake Transactional/Analytical Processing) unifies transactions, analytics, and streaming on a single governance model and storage layer, storing Postgres-native transactional data in Delta and Iceberg format from the point of write. Lakehouse RT is a real-time analytics product powered by the new Reyden engine, delivering sub-100ms latency at high query volume directly on governed tables. 7. **How does Atlan work with the Databricks agentic stack?** Atlan is the context layer for AI. It layers on top of Databricks Unity Catalog and Genie Ontology, and extends governed, live context across the rest of the estate, including Snowflake, BigQuery, dbt, Tableau, Power BI, Salesforce, and SAP. It serves that unified context back to Databricks Genie via an MCP server, SQL interface, and open APIs, making the Databricks agentic stack work across the whole ecosystem. 8. **What was the main theme of Databricks Summit 2026?** The Summit centered on agents that actually work in production, with governance as table stakes and context as the grounding for agent accuracy. In his keynote, CEO Ali Ghodsi framed it directly: AI does not have an intelligence problem, it has a context problem. AI cost was a recurring executive theme across the event. --- ## Sources 1. [Databricks Launches Genie One: All-New Agentic Coworker for Every Team, Databricks Newsroom](https://www.databricks.com/company/newsroom/press-releases/databricks-launches-genie-one-all-new-agentic-coworker-every-team) 2. [What's new with Unity Catalog at Data + AI Summit 2026, Databricks Blog](https://www.databricks.com/blog/whats-new-unity-catalog-data-ai-summit-2026) 3. [AI governance at Data + AI Summit 2026: What's new with Unity AI Gateway, Databricks Blog](https://www.databricks.com/blog/ai-governance-data-ai-summit-2026-whats-new-unity-ai-gateway) 4. [Agent Bricks: Data + AI Summit 2026, Databricks Blog](https://www.databricks.com/blog/agent-bricks-dais-2026) 5. [What's new in Genie Code at Data + AI Summit 2026, Databricks Blog](https://www.databricks.com/blog/whats-new-genie-code-data-ai-summit-2026) 6. [Databricks Launches LTAP: The First Lake Transactional/Analytical Processing Architecture, Databricks Newsroom](https://www.databricks.com/company/newsroom/press-releases/databricks-launches-ltap-first-lake-transactionalanalytical) 7. [Ali Ghodsi's Keynote: AI Doesn't Have an Intelligence Problem, It Has a Context Problem, BigDATAwire](https://www.hpcwire.com/bigdatawire/2026/06/16/ali-ghodsis-keynote-ai-doesnt-have-an-intelligence-problem-it-has-a-context-problem/) 8. [Key Takeaways From Day Two of Databricks Data + AI Summit, SiliconANGLE](https://siliconangle.com/2026/06/17/key-takeaways-day-two-databricks-data-ai-summit/) 9. [Databricks Bets on Owning the Agentic Data Stack at Data + AI Summit 2026, Moor Insights & Strategy](https://moorinsightsstrategy.com/field-notes/databricks-bets-on-owning-the-agentic-data-stack-at-data-ai-summit-2026/) 10. [Everything Databricks Announced at the DAIS Data + AI Summit 2026, Qubika](https://qubika.com/blog/everything-databricks-announced-dais-data-ai-summit-2026/) 11. [Databricks Summit 2026 Day 2: Agentic AI and Catalog Federation Move From Lab to Enterprise, TechTimes](https://www.techtimes.com/articles/318450/20260616/databricks-summit-2026-day-2-agentic-ai-catalog-federation-move-lab-enterprise.htm)