Databricks

Attribute-based access control (ABAC) row filtering and column masking policies, governed tags, and data classification capabilities in Unity Catalog are now generally available for fine-grained data governance.

Build AI agents grounded in enterprise data with optimization for quality and cost using synthetic data, custom evaluation, and automated tuning.

Build production-quality AI agents with built-in evaluation using AI-assisted assessments and human feedback integration for ensuring output quality.

Production-ready tracing for any AI agent using OpenTelemetry and Unity Catalog on Databricks, enabling observability and monitoring across agent deployments.

A platform solution for building, deploying, and governing AI agents on the Databricks Data Intelligence Platform.

A highly performant vector database with real-time syncing of source data for RAG and semantic search applications.

General availability of Apache Iceberg v3 on Databricks with Open Sharing and Unified Governance via Unity Catalog.

Security feature allowing customers to manage their own encryption keys for Lakebase Postgres, providing greater control over data security.

Scalable monitoring that detects anomalies, tracks freshness, and delivers consistent quality signals across AI assets.

The industry's first certification for reliable AI agent systems, aimed at context engineers building on Databricks.

Unified platform integrating data engineering, AI/ML, analytics, and governance to build end-to-end data and AI applications.

Natural-language analyst agent on top of Lakehouse data; conversational SQL and chart generation for business users.

Next-generation conversational AI interface that unlocks business user access to insights across the entire data estate without requiring SQL expertise.

Databricks Genie applied to financial services use cases, democratizing data access for business users through conversational AI on financial data.

Collaborative notebooks enabling real-time team collaboration and streamlined data science workflows for AI and ML development.

Postgres-compatible database product optimized for data apps and AI agents, featuring 5x faster write performance through a new lakebase architecture.

Automatic savings pricing model for Databricks Lakebase that provides cost optimization without manual configuration.

Change Data Feed feature for Databricks Lakebase enabling real-time change tracking for Postgres databases used by data apps and AI agents.

Database branching capability for Databricks Lakebase (Postgres for data apps and AI agents), enabling evolutionary database development workflows.

Open agentic SIEM built for the AI era, providing security intelligence and threat detection capabilities on the Databricks platform.

Enterprise-grade MLOps platform extending open source MLflow with reliability, security and scalability for building and deploying models and GenAI apps.

Unified deployment platform for agents, GenAI models, and classical ML models in production environments.

Fine-tune open source LLMs, pretrain custom LLMs, or build classical ML models for enterprise use cases.

Production-grade model training, fine-tuning, and serving on top of the Lakehouse; Vector Search for retrieval.

An AI-powered business productivity tool providing unified search, chat, dashboards and apps that unlocks business user access to insights across the data estate.

A feature that accelerates LLM inference by caching prompts for open-source models hosted on Databricks, reducing latency and compute costs.

Databricks capability for reliable large language model inference at scale, targeting data science and ML workloads.

A single control plane to apply data governance, access controls, and rate limiting across every LLM and MCP in the enterprise.

End-to-end governance platform for agents and AI models providing access controls, rate limiting, content filtering, and data lineage tracking.

Unified governance layer for data + models; lineage, access controls, audit. Critical for regulated finance use.

Expanded open APIs for Unity Catalog enabling broader interoperability with external tools and platforms for unified data governance.