Artificial intelligence news.
Straight from the original announcements.

Databricks: Spatial SQL on Lakehouse//RT up to 38X faster than a cloud warehouse

October 15, 2026, 19:31By the AI News Bites Editorial TeamSource: Databricks
Databricks: Spatial SQL on Lakehouse//RT up to 38X faster than a cloud warehouse
Image: Databricks

Databricks has brought Spatial SQL and spatial joins to Lakehouse//RT, its real-time data warehouse. According to the company, benchmarks show spatial queries running up to 38X faster than on a traditional cloud data warehouse when 50 users query at the same time.

Key points
  • Spatial SQL and spatial joins now run on Databricks' Lakehouse//RT real-time warehouse, available in Beta.
  • Databricks reports up to 38X faster spatial queries than a traditional cloud data warehouse at 50 concurrent users.
  • With single-query concurrency, the benchmark shows up to 20X faster results and 3X on average.
  • Existing spatial queries from Databricks SQL Serverless run with no code changes and no new licenses, according to the company.
  • The goal is to remove the need for a separate geospatial or real-time serving layer.

What Databricks is releasing

Spatial SQL and spatial joins now run on Lakehouse//RT, the real-time warehouse Databricks introduced at Data + AI Summit 2026. The company describes it as a warehouse built to deliver millisecond query performance at massive concurrency directly on lakehouse data. Spatial queries can therefore be served with low latency and high concurrency from governed Delta and Iceberg tables, without copying data into a separate geospatial system or real-time serving layer.

Earlier this year, Databricks brought Spatial SQL to General Availability, including the GEOMETRY data type, 100+ spatial functions, high-performance spatial joins, AI/BI maps and Delta Sharing. According to the company, the GEOMETRY and GEOGRAPHY data types are part of the open lakehouse, available in Apache Spark and supported in the same underlying parquet files used by Delta and Iceberg v3.

Spatial SQL on Lakehouse//RT is available now in Beta. To start, users create a Lakehouse//RT real-time warehouse from the SQL Warehouses page and run their existing spatial queries.

The benchmark results

Databricks benchmarked Lakehouse//RT against a traditional cloud data warehouse using SpatialBench, a suite of spatial queries. The company reports the following:

  • Lakehouse//RT was 3X faster on average and up to 20X faster across the tested queries with single-query concurrency.
  • At 50 concurrent users, it was up to 38X faster on the queries where the traditional warehouse could keep up.
  • Excluding the largest improvement, the average speed-up is over 3.5X.
  • Every spatial benchmark query was faster on Lakehouse//RT.

Databricks points readers to the SpatialBench documentation to generate the synthetic data and reproduce the results. For query #8, it recommends removing the ST_GeomFromWKB call that converts data stored as well-known binary and using the native GEOMETRY data types instead. According to the company, native types allow query optimizations and file-skipping that give the best performance.

Why separate spatial stacks are the target

Databricks says many organizations with both spatial OLAP workloads and real-time location-based applications still depend on traditional data warehouses. Some also run a separate stack centered on PostGIS for each use case to meet latency requirements. In the company's view, this means more copies of data, more synchronization pipelines and more governance overhead.

Databricks now describes four pillars for geospatial workloads: Lakehouse for long-running spatial OLAP, Lakebase for spatial OLTP, apps and agents, Databricks Apps for purpose-built experiences, and Lakehouse//RT for real-time analytics. Spatial SQL on Lakehouse//RT combines real-time performance with open formats and Unity Catalog governance, according to the company.

What it means for users

The main point for existing customers is portability. Databricks says the same spatial functions and joins that run on Databricks SQL Serverless today run on Lakehouse//RT with no code changes and no new licenses. Users create a Lakehouse//RT warehouse and point their queries at it. All H3 expressions supported by Databricks Lakehouse are also supported, and AI/BI dashboards can build maps with native geometry and geography types.

The company lists example use cases across several industries:

  • Telecom and retail: evaluating subscriber signal health in coverage areas and trade-area overlap for new store sites.
  • Energy: intersecting well locations with land leases and pipeline routes with suitability corridors.
  • Insurance: matching policies to projected storm paths and calculating insured value at risk.
  • Logistics: assigning orders to the nearest service area.
  • Security: serving point-in-polygon lookups to dashboards and agents at high concurrency.

The figures come from Databricks' own benchmarks, and the feature is still in Beta. For Lakehouse//RT in general, the company says preview customers have seen latency as low as 10ms and concurrency as high as 12,000 queries per second.

Related news

Eclipse Aerospace brings two decades of flight data into Databricks
Products and toolsNovember 12, 2026, 20:20

Eclipse Aerospace brings two decades of flight data into Databricks

Eclipse Aerospace has modernized the ingestion of nearly 20 years of proprietary flight data with Databricks, giving service centers and engineers faster access to…

Products and tools
Products and toolsNovember 12, 2026, 19:26

Databricks page says Genie Migrate is generally available

Databricks says Genie Migrate is now generally available for migration to the Databricks Lakehouse. The company page carrying the announcement is currently unavailable…

Databricks says Lakehouse//RT cuts the cost of real-time analytics
Products and toolsNovember 12, 2026, 17:53

Databricks says Lakehouse//RT cuts the cost of real-time analytics

Databricks says Lakehouse//RT is designed to absorb the rising query volumes generated by AI agents without a proportional increase in infrastructure spending. Customers…