SAP and Databricks Expand SAP Databricks Access to SAP Application Data

SAP announced an expansion of its partnership with Databricks at SAP TechEd Berlin. The centerpiece is SAP Application Data Foundation, which makes data from SAP applications available to the SAP Business Data Cloud (SAP BDC) engine a customer chooses, including SAP Databricks.

The goal is to let data teams build AI on SAP data without the custom pipelines that have typically copied it out of SAP systems. Databricks plans to bring more of its tools to SAP Databricks, which embeds the Databricks platform in SAP BDC.

Direct Access to SAP Data Without Custom Pipelines

Reaching SAP application data has typically meant building and maintaining custom extraction pipelines, Databricks notes. Those pipelines copy the data, lose business context along the way, and fragment governance.

SAP Application Data Foundation makes the data available inside SAP BDC instead, which SAP says lowers total cost of ownership and shortens time to value. SAP data products can then be shared through SAP BDC Connect for zero-copy access, without additional pipelines, copies, or silos.

The access reaches SAP data in cloud and on-premises systems, Databricks states. Customers can use SAP data inside SAP BDC with SAP Databricks, or bring governed SAP data into an existing Databricks estate, with any model on any cloud.

A Demand Forecasting Example Shows the Workflow

A manufacturer in SAP’s example uses historical sales, inventory, and supply chain data to build demand-forecasting models. Data professionals enrich the data from SAP BDC with third-party data and create custom data products in SAP Databricks. Engineers build on SAP’s business semantics, which SAP says grounds forecasts in logic already embedded in SAP systems.

Databricks offers an on-premises version. A manufacturer running finance and supply chain on SAP ECC, while planning its move to the cloud, can bring years of order history into SAP Databricks and keep that work after the move. Databricks lists other uses customers want, including detecting anomalies in the financial close and optimizing procurement decisions.

Databricks Plans to Add More Tools to SAP Databricks

The expanded partnership plans to bring Lakeflow for data engineering across SAP and third-party sources, Unity Gateway for governing AI and controlling its cost, and Lakebase, a serverless Postgres database for AI apps and agents, to SAP Databricks inside SAP BDC. SAP also lists Agent Bricks among the planned additions.

Databricks plans a two-way integration between its Genie assistant and Joule, which would bring Genie’s answers and actions into SAP experiences.

Available today are SAP Databricks, which the two companies introduced last year, and SAP BDC Connect for Databricks, Databricks says. It reports that SAF-Holland, which uses SAP BDC and SAP Databricks, generates analyses five times faster than with traditional reporting.

What This Means for SAPinsiders

SAP data becomes a shared product. SAP positions data products and zero-copy sharing as the way into SAP data, in place of custom extraction pipelines that Databricks says copy data, lose business context, and fragment governance. SAP Application Data Foundation feeds those products to the engine a customer chooses.

The platform choice stays open. Databricks says SAP data can be used inside SAP BDC with SAP Databricks or brought into an existing Databricks estate. The decision about where AI work runs remains with SAP customers that already invest in Databricks.

AI work can start on on-premises SAP data. Databricks describes SAP Application Data Foundation as reaching SAP data in cloud and on-premises systems, and its example is an ECC manufacturer building forecasts while it plans its move to the cloud. AI work and migration timing become separate decisions for ERP buyers in that position.