Databricks @databricks

We are honored to be recognized by @OpenAI with the Data Intelligence Platform Award. This recognition reflects the work our teams have done together to bring frontier models into the enterprise, grounded in the governed data and context essential for solving complex business problems. We are excited to continue partnering with OpenAI to push the boundaries of enterprise intelligence. We are looking forward to hosting @gdb, Co-founder and President of OpenAI, at Data + AI Summit. He will join Databricks CEO @alighodsi to discuss why enterprise context is the foundation for capable AI and what the future of agentic data work looks like.

AI governance is becoming a data governance challenge, and organizations need a consistent way to govern agents, models, MCP servers, and data together. Unity Catalog and Unity AI Gateway extend governance across AI systems with identity-aware access controls, runtime policies, guardrails, auditability, observability, and cost intelligence across every agent interaction. Built on open standards, the approach supports consistent governance across Databricks and third-party AI ecosystems without slowing AI adoption. Learn more: databricks.com/blog/governing…

Delta Sharing became one of the most popular ways to exchange data thanks to its open cross-platform nature. We’ve now expanded it to also support any Iceberg client and to share AI assets like agent skills and unstructured data. It needed a new name, so welcome OpenSharing!

The Linux Foundation has launched the OpenSharing Project, an open protocol for sharing AI assets and data across organizations, platforms, and cloud environments. Contributed by @databricks, OpenSharing extends Delta Sharing to support the agentic AI era, enabling interoperable exchange of AI models, agent skills, and unstructured data through a common standard. Learn more: bit.ly/4v4gHg8

Announcing OpenSharing, a new open standard for sharing data and AI assets across platforms and organizations. Five years ago, we pioneered open data sharing with Delta Sharing, the most widely adopted open protocol for secure zero-copy data sharing - used by thousands of partners and customers. Today, we're excited to announce its evolution to OpenSharing: The first open, vendor-neutral protocol for sharing AI assets, including agent skills, AI models, and unstructured data - in collaboration with partners including @OpenAI, @SAP, @LSEGplc, @stripe, @TheTradeDesk, @Minio, @AmadeusITGroup, @Atlassian, and more. Enterprises can now use OpenSharing to share AI Assets, on-premises assets, as well as share structured and unstructured data in all formats. Hosted by The @linuxfoundation, OpenSharing creates an open foundation for data and AI collaboration in the agentic era. databricks.com/company/newsro…

💳 Card fraud happens in seconds, but many detection systems still rely on slow batch processing or complex streaming architectures. This open-source Solution Accelerator shows how teams can build an end-to-end fraud detection workflow with Spark Real-Time Mode and Lakebase, including low-latency ML scoring, explainable fraud signals, and live monitoring on a unified platform. With it, organizations can achieve sub-second fraud intervention while reducing operational complexity and avoiding the need for separate streaming infrastructure. databricks.com/blog/how-build…

.@AnthropicAI's Claude Fable 5 is now available on Databricks, governed by Unity AI Gateway, across AWS, Azure, and GCP! Fable 5 sets a new state of the art on OfficeQA Pro at 57.9% and is designed for work that was previously too complex or too long-running to hand to an AI. Databricks customers can build domain-specific agents on Fable 5 and deploy them as Databricks Apps with Lakebase-powered memory, enabling agents to maintain context across sessions. Build your first long-running agent with Agent Bricks: databricks.com/blog/claude-fa…

The Big Book of Data Warehousing and BI looks at why lakehouse architecture is foundational to unifying data, enabling real-time analytics, and preparing for AI across the enterprise. Explore practical guidance on: - Migrating from traditional data warehouses - Modeling data in a lakehouse - Managing cost and performance at scale - Governing data while supporting modern BI and analytics workflows databricks.com/resources/eboo…

keyboard_arrow_left Previous keyboard_arrow_right Next

In just one week, #DataAISummit kicks off in San Francisco! Join data and AI leaders for four days of technical sessions, keynotes, hands-on training, networking, and product announcements shaping what’s next across the industry. If you haven’t registered for the virtual or in-person experience yet, secure your spot today: databricks.com/dataaisummit?u…

AI workloads create new cost management challenges, from runaway retry loops to uncontrolled agent experimentation. AI Spend Controls in Unity AI Gateway introduce proactive budget alerts across users, workspaces, use cases, and entire accounts so organizations can monitor and manage AI spend before it becomes a business risk. Now organizations can apply AI spend controls consistently across coding agents, production agents, overnight batch jobs, models, providers, and AI workloads running throughout the business. databricks.com/blog/introduci…

There’s a ton of interest in custom model tuning as agents reach production and scale up. Here is how we made Databricks Knowledge Assistant 3x faster using our new Instructed Retriever model trained end-to-end to do parallel test-time compute. It’s rolling out to customers now!

Databricks AI Research @DbrxMosaicAI

6 days ago

Most agentic search systems get better by thinking longer: more tool calls, more reason-act loops, each step waiting on the last. Quality goes up, but so does latency. Instructed-Retriever-1 takes a different route. Instead of scaling test-time compute sequentially, it scales it

5 9 71 27K 65

A 9,000-mile road trip. 35 days. 25 cities. One data platform. As part of #RoadToSummit, v4c.ai founder Vijay Rao is making his way to #DataAISummit with every layer of the journey running on the Databricks Lakehouse, including vehicle telemetry, trip analytics, social engagement, and a live AI-powered chatbot. Next stop: #DataAISummit. Follow along for live updates: road2summit.ai

keyboard_arrow_left Previous keyboard_arrow_right Next

The Databricks Architecture Center is your hub for practical ways to build pipelines, dashboards, and AI models. Find designs for common data and AI projects across use cases: - Intelligent Data Warehousing on Databricks - Reference Architecture for Credit Loss Forecasting - Telecom Next Best Action Reference Architecture - and much more! Chart a faster path from idea to implementation, with patterns teams can build on: databricks.com/resources/arch…

keyboard_arrow_left Previous keyboard_arrow_right Next

Lakebase brings operational databases into the same governance layer as analytics. This walkthrough explores what happens when Backstage, Spotify’s open-source internal developer portal, runs with Unity Catalog governance across ephemeral Postgres branches. What changed: • Access controls, audit trails, row-level security, and masking policies applied automatically across branches • Branch creation, deletion, and cost attribution traced with a single SQL query • Developers safely created isolated database branches without exposing sensitive production data databricks.com/blog/backstage…

🛠️ Building apps should be fast. Managing infrastructure shouldn’t slow developers down. Watch @andrelandgraf from Databricks and Kevin Niparko from @cursor_ai walk through how to build and deploy a production-ready app on Databricks using built-in governance and serverless Postgres. You’ll see how teams can: ✓ Prototype and iterate in real time ✓ Branch, scale, and manage transactional data ✓ Govern data securely with SSO and Auth ✓ Deploy apps where their data already lives Watch the on-demand webinar ↓ databricks.com/resources/webi…

Unity Catalog continues to expand open lakehouse interoperability. External access to UC managed Delta tables is now in Beta, allowing engines like Apache Spark, Apache Flink, and DuckDB to create, read, and write managed tables with transactional safety and centralized governance. Credential vending is also now GA, adding M2M OAuth support and automatic credential refresh so external engines can securely access UC-governed data without long-lived static credentials. databricks.com/blog/expanded-…

"We don't really need AI to get smarter. What we need is for it to have context." Databricks co-founder and CEO @alighodsi joined @CarolineHydeTV and @EdLudlow at the Bloomberg Tech event in San Francisco to make case that the real breakthrough will come not from smarter models but from giving AI the data context it needs to perform. That is the problem Databricks Genie is built to solve. Watch the full interview: youtube.com/watch?v=CipJDv…

Most agentic search systems get better by thinking longer: more tool calls, more reason-act loops, each step waiting on the last. Quality goes up, but so does latency. Instructed-Retriever-1 takes a different route. Instead of scaling test-time compute sequentially, it scales it in parallel. One retrieval-specialized model fans the work out: it generates multiple query and filter formulations to widen recall, then reranks the merged evidence with a multi-pivot reranker to sharpen precision. Both stages run at once, so searching more broadly no longer means searching more slowly. The result inside Knowledge Assistant: search time drops more than 3x and answer time 2x, with time to first token around two seconds, and no drop in quality (it matches Claude Sonnet 4.5 retrieval quality on KARLBench). For the people using it, that means far less waiting between question and answer, the freedom to ask more follow-ups, and more of the knowledge base actually surfaced. Rolling out to all customers now, with no reconfiguration. Read how we did it: databricks.com/blog/3x-faster…

[DEMO] Bring Genie to where your teams already work 🧞 Genie can now connect directly with Microsoft Teams, M365 Copilot, and SharePoint through a native integration with Copilot Studio. Users can ask questions of their data in natural language and get insights without leaving their workflow. Learn how to connect Databricks → add Genie to an agent → and publish it so teams can access answers in the tools they already use. youtu.be/2zLzCptIr9A

Introducing the lab sponsors for the Databricks Grounded Reasoning Cup at #DataAISummit 2026: @AnthropicAI, @OpenAI, and @GoogleDeepMind. Each lab is partnering with leading academic teams to build agents that tackle grounded reasoning over complex government data using the latest models and tooling. Meet the teams, see the agents they’re bringing to the competition, and join us as they push the boundaries of enterprise AI reasoning live on stage! 🏆

keyboard_arrow_left Previous keyboard_arrow_right Next