Databricks Unlocks Agent Bricks, Genie Code and Lakebase for Free

A year ago, Databricks launched its Free Edition with a simple bet: make the platform the best place to learn modern data and AI. That bet is paying off. The company has given over $10M in free credits to more than 500,000 people learning, experimenting, and building on Free Edition. Now, Databricks is doubling down by unlocking five of its most powerful enterprise-grade tools at no cost.

The five new additions are Genie Code, serverless GPUs, Lakebase, Agent Bricks, and Lakeflow Designer. With these additions, Free Edition now includes every core Databricks practitioner feature, giving users a complete toolkit for building end-to-end data and AI projects. This is not a watered-down sandbox. It is the same stack that enterprises pay for, now available from signup.

The five tools, unpacked

Each addition targets a specific gap that previously forced free users to hit a wall or switch to a paid tier.

• Genie Code -- Genie Code goes beyond natural language querying, generating, running, and iterating on code autonomously. Ask it to analyze a dataset, clean a pipeline, or build a visualization, and it will write the code, execute it, interpret the results, and refine its approach based on what it finds. This is not autocomplete. On real-world data science tasks, Databricks found Genie Code more than doubled the success rate of leading coding agents, jumping from 32.1% to 77.1%. In fact, Genie Code is now writing more code on the platform than humans do.
• Serverless GPUs -- Free Edition now includes access to GPUs, subject to availability. This matters because training neural networks, fine-tuning models, and running inference all require GPU compute that has historically been locked behind paid tiers or separate cloud accounts.
• Lakebase -- Lakebase brings a fully managed Postgres-compatible database to Free Edition, purpose-built for data apps and AI agents. With Lakebase, you can build transactional applications that sit directly on your lakehouse, with no separate infrastructure, no data silos, and no reconciliation headaches. Databricks says Lakebase now handles 12 million database launches a day at the enterprise level, and that same infrastructure is now available to free users.
• Agent Bricks -- Agent Bricks is Databricks' framework for building production-ready AI agents. It provides pre-built, composable components including tools, memory, orchestration, and evaluation, letting you move from idea to working agent in a fraction of the time. Agent Bricks has seen 100,000+ agents built since launch and processes more than one quadrillion tokens per year.
• Lakeflow Designer -- Lakeflow Designer makes it easier to build data pipelines visually. Instead of writing Spark code from scratch, you can design data flows and connect pipeline steps in a drag-and-drop interface, making data engineering concepts tangible for learners and faster for practitioners.

The bigger picture: a developer flywheel at scale

This is not a charitable move. It is a calculated one. Developers adopt open-source tools, become familiar with the Databricks ecosystem, and then advocate for the commercial platform within their organizations. This developer-led go-to-market strategy is increasingly important in enterprise software. Giving away the full practitioner toolkit accelerates that flywheel dramatically.

The timing also aligns with Databricks' broader momentum. The company crossed a $5.4 billion revenue run-rate, delivering over 65% year-over-year growth, and is completing investments in excess of $7 billion, including approximately $5B of equity financing at a $134 billion valuation. Databricks now ranks as the fourth most valuable private company globally, behind only ByteDance, SpaceX, and OpenAI. Growing a half-million-strong free user base is a direct pipeline into that enterprise revenue engine.

The announcement also came off the back of Data + AI Summit 2026, held in San Francisco with approximately 30,000 attendees, where Databricks centered the event on agents, context, and governance. The Free Edition expansion was one of several moves positioning Databricks as the default platform for the agentic era, not just for enterprises, but for anyone learning to build with data and AI.

Who wins, and what it means for the competition

The clearest winners are individual builders and students who previously had to stitch together Google Colab for notebooks, a free Postgres instance, and separate GPU credits just to approximate what Databricks enterprise customers get out of the box. That fragmented setup is now replaced by a single, governed, production-grade environment.

For Databricks' competitors, the pressure is real. Gartner named Databricks a Leader in the Magic Quadrant for Data Science and Machine Learning Platforms, ranked highest in both Ability to Execute and Completeness of Vision, for its fifth consecutive year. Snowflake, the most direct rival, has been pushing into AI with Cortex and natural-language features, but machine learning capabilities lag behind Databricks significantly. Giving away the ML and agent stack for free raises the bar for what any competitor's free tier needs to offer.

The second-order effect is on the talent market. Builders have used Free Edition to learn data engineering, train machine learning models, analyze real datasets, and build AI agents. As those users graduate into professional roles, they carry platform familiarity with them, and that familiarity shapes procurement decisions. This is the same playbook that made GitHub and VS Code dominant: win the developer before the enterprise budget conversation even starts.

What is now possible

The combination of Lakebase and Agent Bricks in a free environment is particularly significant. Previously, building a production-grade AI agent that could read and write to a transactional database, governed by Unity Catalog, required an enterprise contract. Now it requires a signup.

As organizations adopt agentic workflows powered by tools like Genie and Agent Bricks, Lakebase provides the operational database layer agents rely on to manage state and application workflows. This foundation also enables a new wave of agentic data engineering and agentic data science with tools like Genie Code. The full loop, from ingestion to pipeline to agent to transactional database, is now available to experiment with at zero cost.

There is one caveat worth noting: GPU access in Free Edition is subject to availability. For compute-intensive workloads like fine-tuning large models, free users may encounter queuing or limits during peak demand. But for most learning and prototyping use cases, the access is real and meaningful.

Getting started

Free Edition is available now. Sign up, open Genie, and run your first analysis in seconds. The Free Edition signup page requires no credit card. For a deeper look at each new feature, the official announcement blog walks through what each tool does and how they fit together. If you want to go straight to the agent-building layer, the Lakebase product page and the Genie Code launch announcement are the best starting points. The full practitioner stack is free. The only question is what you build with it.