Databricks Builds a Single Computer System for Business Records and Data Analysis

Databricks announced a new technology called LTAP — which stands for Lake Transactional and Analytical Processing — on June 16, 2026. The company says this is the first system that can handle two very different kinds of work at the same time: storing business records and analyzing data, without needing separate computers for each job.

The company also released Lakehouse//RT, which adds real-time analytics, and a new tool called the Unity AI agent gateway.

Why This Matters

Think of it this way. Most companies today run two separate computers for different jobs. One handles the transaction work — the moment-to-moment business of storing a customer order, updating a bank balance, or recording a sale. The other system analyzes that data afterward, answering questions like "how many orders did we get this month" or "which products sell best". These two systems never talk to each other directly. Companies spend a lot of engineering time copying data back and forth between them, keeping them in sync, and managing the gaps where the two systems disagree about what the truth is.

For decades, this split has been one of the hardest problems in how companies organize their data. The lakehouse — a newer idea that Databricks pioneered — tried to fix some of this by storing data in a way that could be used for both purposes. But it never fully solved the problem of handling the fast, up-to-the-second record-keeping that a business needs.

Databricks says LTAP and its new product Lakebase solve this. The company claims you can now run both types of work on one system, all at the same time. Lakebase, the actual product, was built from scratch to handle this, rather than trying to patch an older system to do both jobs.

What This Means for AI

The Lakehouse//RT addition brings real-time data into the picture — meaning the data you get right now, not data from an hour ago. This matters for AI systems that need to read a piece of information, make a decision, and write back an answer immediately. Those kinds of systems require the data to be fresh and accurate.

The Unity AI agent gateway adds another layer: it controls how different AI systems access the data, and makes sure they follow the company's rules about who can see what.

What Comes Next

The real question is whether Lakebase can actually handle the speed and accuracy that real business applications need when hundreds of people are using it at the same time. This is a technical claim that will be tested in the real world — not just in a demonstration on stage. That said, Databricks has a track record. Earlier technologies it built — Delta Lake and Photon — started as promising ideas and turned into systems companies actually use in production. The company put itself on a schedule now.

Data engineers will want to test how fast Lakebase is when many people are writing records at the same time, and whether it keeps data consistent in the way it promises. For teams building AI systems, the more immediate draw is that Lakehouse//RT and the Unity gateway speak directly to the speed and control that production AI systems need.

Databricks has not said how much this will cost or when it will be generally available for customers to use.