#DataAISummit 2026 may be over, but the keynotes and select breakout sessions are now available on demand! Catch up on this year's top product announcements, technical sessions, and customer stories: databricks.com/dataaisummit?…

Say hello to Omnigent 👋 Databricks' open-source meta-harness helps developers orchestrate AI coding workflows across multiple coding agents, making it easy to use the right model for each task. In this video, @cole_medin explains why meta-harnesses are becoming an important part of modern AI coding and demonstrates how Omnigent orchestrates coding assistants with shared sessions, guardrails, and human-in-the-loop workflows. youtube.com/watch?v=oGE_Dwz-…

⏳ 90 second crash course on @databricks ’ biggest launches across the Data AI stack

[DEMO] Agentic coding is moving faster than ever. Your database needs to keep up. Lakebase Branching creates isolated, ephemeral database environments that let AI coding agents experiment and iterate without putting production data at risk. See how branching works, why it's fast and cost-efficient, and how to set up an agentic workflow with the Databricks CLI and AGENTS.md. youtube.com/watch?v=jLX5LBzE…

Introducing Genie App Builder for AI-assisted creation with all Databricks context. Genie App Builder is a purpose-built AI app authoring tool with native awareness of your data assets, Unity Catalog semantics and workspace context. It's part of a Databricks suite of products and capabilities that brings governed app vibe coding to enterprises. Learn more: databricks.com/blog/enabling…

My co-founder @rxin personally wrote this really good paper that explains the main idea behind postgres Lakebase as well as LTAP. It almost serves as a primer on how transactional databases are built and how Lakebase and LTAP work. Maybe more importantly, what are the tradeoffs, and what are you giving up by adopting this new approach. Highly recommended reading: databricks.com/blog/lakebase…

#DataAISummit 2026 was packed with product announcements, hands-on training, and technical insights. Missed a session? Catch up on Genie One, LTAP, CustomerLake, and more by watching the keynotes and select breakout sessions on-demand: databricks.com/dataaisummit?…

Full-text search indexes are now available in Beta on Databricks Runtime 18.2. Create a search index with a single SQL statement to accelerate substring and keyword queries, with no query hints or application changes required. Teams are already seeing 100x faster substring search queries against petabyte-scale tables. Learn how it works: databricks.com/blog/accelera…

Databricks ranks #1 on NVIDIA’s SOL-ExecBench kernel leaderboard, in the L1 single operation track, powered by KDA (Kernel Design Agents) 🎉 What’s crazy is: we 100% leveraged AI agents to beat the competition. This is a sneak peek at recursive self-improvement. The core frameworks we used were KDA, Humanize, and Omnigent: Claude writes code, Codex reviews. Together, they enabled agents to run autonomously for as long as possible. The key is setting up the right framework to let the agents cook. This work was driven by @leshenj15 at Databricks, in collaboration with NVIDIA and MIT HAN Lab’s @LigengZhu and @DongyunZou03 . Databricks AI is like a neolab. Join us if you’re cracked!

See the latest updates in Unity AI Gateway: Databricks’ governance solution for AI assets, models, agents, LLM endpoints, MCP servers, Skills, and coding agents. Now you can create a centralized inventory of all your AI assets with the agent registry in Unity AI Gateway. Register external agents, agents built on Databricks, MCPs, and Skills to create a comprehensive system of record. databricks.com/blog/ai-gover…

⚡ What are Declarative Pipelines in < 90 seconds? Let's walk with @michaelarmbrust , Distinguished Engineer @databricks who led the design & development of Declarative Pipelines created Spark SQL, Structured Streaming, Delta Lake, & much more as he breaks it down! 💡 TLDR: instead of telling the system HOW to do every step, you just tell it WHAT you want to exist. Spark figures out the how. The technology itself is open source in Apache Spark, and Databricks offers a managed version on top called Lakeflow Spark Declarative Pipelines It was such a blast getting to chat with Michael! Did you know he was also the 9th employee at Databricks?!

.@AnthropicAI's Claude Sonnet 5 is now available on Databricks, governed by Unity AI Gateway, across AWS, Azure, and GCP! It brings near-Opus intelligence with top performance on Agentic Search and Document AI multimodal extraction benchmarks plus advanced reasoning and thinking modes for complex tasks. The same balance of capability, cost, and speed teams already rely on Sonnet for. Databricks customers can build agents on Sonnet 5 that are specific to their domain. Agents deploy as Databricks Apps with Lakebase-powered memory so agents maintain context across sessions. Build your first long-running agent with Agent Bricks: docs.databricks.com/aws/en/r…

Most data teams run two systems: one database for transactions, another for analytics, with data copied between them. It works, but the pipelines are brittle, the copies drift, and governance gets complicated. At this year's Data AI Summit, we announced LTAP, a new architecture that does away with the second copy entirely. With LTAP, transactions and analytics run on a single copy of your data, stored in open formats that both sides read directly. Analytics always sees the freshest data your application just wrote, with no CDC, no mirroring, and no slowdown to your transactional workload. Databricks co-founder @rxin walks through the architecture that makes this possible, starting at the storage layer of a traditional database and building from Lakebase up to LTAP. databricks.com/blog/lakebase…

I find the Lakebase design for serverless Postgres very elegant, so I spent some time explaining how it works in this blog. The blog starts by explaining how databases really persist data (with a write-ahead-log and data files that are updated async), and how Lakebase separates storage and compute by externalizing those two components. It ends with how the Lakebase architecture naturally leads to LTAP, enabling OLTP and analytical workloads against a single governed copy of data. My goal was to make it readable by anyone curious about how these systems work, not just database and storage experts. That turned out to be a lot more challenging than I first thought. Database storage is one of the most complex areas in computer science (the ARIES paper cited in blog was the hardest paper I personally ever had to read). The first draft had too little detail and I couldn't land the ideas. The second had too much and I'd lost anyone who isn't already a storage expert. This is the third draft, and I'd love feedback on whether the depth feels right. databricks.com/blog/lakebase…

.@Ecolab built its Retail Intelligence application on Databricks and @AnthropicAI's Claude to unify nine siloed data sources into a single intelligence engine. Built as a native Databricks App with Lakebase Postgres and Unity Catalog, the solution cuts compliance report compilation from two weeks to under two minutes while delivering cited, real-time answers from 700-page FDA food safety manuals for frontline retail staff. Read how Ecolab is powering retail intelligence with AI agents: databricks.com/blog/how-ecol…

We are thrilled to share that Databricks has been named a @usnews Best Company to Work For! 🏆 In addition to making the Overall list, Databricks was also recognized as one of the: - Best Companies to Work For in Information Technology - Best Companies to Work For in the West 🧱 Thank you to all Bricksters for fostering a culture where we can innovate, collaborate, and grow together every day. careers.usnews.com/companies…

The next frontier for AI isn't building a single agent. It's running fleets of AI agents across the enterprise. Join Databricks, @OpenAI, and @Stellantis for a live discussion on what it takes to ship agentic apps at scale. Hear the latest on building accurate, governed agents, what developers need to trust AI agents, and how one of the world's largest automakers is industrializing agentic AI across the enterprise. Register now: databricks.com/resources/web…

.@databricks co-founder Arsalan Tavakoli at SaaStr AI 2026 did a deep dive on where B2B is really going. His top point: 👋Any business with a monopoly today won't have one in 12 to 24 months. Three things broke pricing power at the same time: 1. The cost of building software collapsed A new entrant builds on top of data that's already ingested and governed, and ships fast. 2. The low end got good AI plus third-party APIs turns the cheap one-workflow tool into something real. "You get what you pay for" is breaking. 3. Migration cost collapsed Migrations used to die on the vine: save 50%, but the migration costs 5x the savings. Now LLMs read legacy code, convert it, and validate the output. Databricks does enterprise-grade migrations in 30 days. Once switching is cheap, lock-in stops defending your price. A few other things he said that are true on the front lines, not on X: - Everyone is token maxing Spend is going straight up. Almost no one knows the ROI. The companies winning budget tie AI to an outcome. Standing up agents earns zero points. - The real bottleneck on enterprise AI isn't the model. It's context And context is not data. It's every definition someone learned by asking a colleague, buried in emails and transcripts, going stale by the day. - BI is basically dead One car maker just put 70,000 employees on Genie. Not the 5% who write SQL. The 95% who run the business and now get answers in 30 seconds. If you're attacking an incumbent, the wedge is finally there. If you're the incumbent, lock-in is not a strategy anymore youtu.be/Gv6Eq0Q1O_U

At #DataAISummit, @matei_zaharia, Co-founder and CTO at Databricks, announced Omnigent: an open meta-harness that sits as a layer above your agents and makes them interoperable parts of a powerful system. Omnigent provides a common interface to compose multiple agents, enforce advanced policies, and collaborate in real-time across your team. databricks.com/blog/introduc…

Building beautiful dashboards just got easier. New dashboard and workspace themes in AI/BI make it easy to customize fonts, colors, and visualization palettes for a consistent, on-brand experience. Apply a theme to a single dashboard, or set a workspace theme once and have it automatically apply across every dashboard in your workspace. See how it works: databricks.com/blog/design-b…