More Than 90% of the Fortune Global 500 Keep Their Books in SAP—Now Databricks Wants to Read Them Directly

Databricks and SAP announced this week at SAP TechEd that they are expanding their partnership. The core is one thing: SAP Application Data Foundation. Its role is straightforward: make SAP application data easier to bring into SAP Databricks, including both cloud and on-premises SAP data.

First, some context. Where enterprise AI is stuck today is not model capability. Models are already strong enough. What's missing is business context—i.e., how your company actually operates. Most of that context sits in SAP systems. More than 90% of Fortune Global 500 companies use SAP to manage critical business processes. How orders flow, how inventory turns, how finance closes the books—it's all in there.

The problem is that in the past, using this data for AI usually required building custom extraction pipelines. Data would be copied out, business semantics would be lost along the way, and governance would be fragmented everywhere. Data engineers probably know this pain: pipelines are hard to write and even harder to maintain, and the business side still thinks you're slow.

SAP Application Data Foundation is intended to solve exactly this part. SAP data products can be shared with zero-copy through SAP BDC Connect, so there is no need to pull additional pipelines, copy data, or create silos. Data engineers can model while preserving SAP business semantics, so analytics and AI reflect how the business actually runs, rather than a copy that has been cleaned and transformed several times over.

At the same time, several key capabilities on the Databricks platform are also planned to come to SAP Databricks, embedded in SAP Business Data Cloud:

• Genie: an AI coworker that gives teams governed access to data and AI, with enterprise context built in. The planned bidirectional integration between Genie and Joule can bring answers and actions into the SAP experience.
• Lakeflow: a unified data engineering solution for ingestion, transformation, and orchestration, combining SAP data with IoT, CRM, and third-party data sources.
• Unity Gateway: multi-AI governance and cost control across agents, tools, models, and MCP, enabling teams to scale AI with guardrails and spending controls.
• Lakebase: a serverless Postgres database for AI applications and agents, used to build operational data applications on top of SAP and enterprise data.

A case study shows what this can look like. SAF-Holland, a manufacturer of chassis components for trailers, trucks, and buses, used SAP BDC and SAP Databricks to unify more than 100 million customer, delivery, and material records, and then used machine learning to predict the reliability of customer delivery schedules. According to the company, analytics are generated five times faster than with traditional reporting, and complex SAP data that used to take weeks to turn into planning and supply chain insights now takes minutes.

David Gund, Senior Manager of Global Digital Intelligence at SAF-Holland, said: “Opening the SAP data ecosystem to Databricks has unlocked better analytics for our internal teams and ultimately helped us deliver better outcomes for our customers.”

SAP CTO Philipp Herzig commented that bringing Databricks' latest capabilities into SAP Business Data Cloud combines SAP's understanding of business processes and critical applications with Databricks' data platform, helping customers turn enterprise data into the foundation for an autonomous enterprise. Databricks SVP Michael Kiermaier noted that combining Genie and Joule can bring business context to every team and agent, enabling them to get trusted answers and take action.

One point worth noting: this partnership leaves architectural choice to customers. You can consume SAP data directly with SAP Databricks within SAP BDC, or bring governed SAP data—or data products created from SAP data—back into your existing Databricks environment, with no restrictions on model or cloud. For companies that already have assets on Databricks, this option is more realistic than a forced migration.

SAP Databricks and SAP Business Data Cloud Connect for Databricks are currently available. SAP Application Data Foundation and the expanded SAP Databricks capabilities can be learned about through SAP or Databricks customer teams.

Enterprise AI has been talked about for so long, but the truly hard part has never been tuning models. It is letting models know what your company is doing. The pile of books in SAP finally has a path that doesn't require building your own pipelines.

Reference link: https://www.databricks.com/blog/databricks-and-sap-unleash-end-end-data-ai-capabilities-level-enterprise-ai-transformation