Today, SAP and Databricks announced SAP Business Data Cloud (BDC) with SAP Databricks. Product capabilities include integration of Databricks capabilities in the SAP data platform, delta sharing for SAP data and Insight Applications. These new capabilities put SAP firmly on the map as a leading global enterprise data platform provider. Here is what customers need to know.
SAP Business Data Cloud? What is that?
The short and incomplete answer is that BDC is SAP DataSphere, with Databricks integration, some new product features and a different pricing model. For the long answer, you can visit the product tour or register for the webcast.
What problem is the SAP Databricks partnership going to solve?
I see two main problems with the current SAP platforms (SAP DataSphere and SAP BW/4HANA):
- Poor integration of data engineering and data scientist toolsets in the SAP ecosystem
Data engineers and data scientist love to work with language like Python, Scala and R, using Notebooks and controlling their source code with git-based solutions. Databricks is built on this premise. On SAP data platforms, it is all but impossible to use this. Bringing Databricks to SAP bridges a very significant shortcoming of the current SAP platforms. - Unable to provide scalable solutions at competitive prices
SAP HANA infrastructure is expensive compared to the infrastructure Databricks is using. Databricks runs on cheap file storage with Spark engines, which is offered at competitive prices by all Hyperscalers. SAP HANA runs on high spec database servers. HANA compute is expensive by nature and that is not a problem SAP can easily resolve. SAP Business Data Cloud will continue to run on SAP HANA, but it will support Databricks ‘Delta Sharing’ capabilities. This means datasets residing outside SAP HANA can seamlessly be integrated in analytical models and reported on. In theory, this would open the door to also offloading SAP data to the Delta Lake, reducing the HANA footprint and cost of the SAP Business Cloud license. Whether SAP will allow this remains to be seen. Probably not without paying a fee for this.
What’s in it for Databricks and for Databricks customers?
That is two questions with very different answers. For Databricks, the partnership means that SAP will sell Databricks as a white label product with their SAP Business Data Cloud. This will result in increased revenue for Databricks, with very little effort required in the sales process.
For Databricks customers, it looks like initially not much will change. Their biggest challenge will still be to get data out of SAP onto the Databricks platform. If customers have a Datasphere license and they would move over to SAP BDC then in theory they could benefit from delta sharing, integrating HANA delta files directly into Databricks. I doubt SAP will allow this. Databricks have been working on an SAP connector for a while. Unconfirmed rumours are that this connector is to be released in Q1 2026. Whether Databricks will still go ahead with this is still unclear, as is the mechanism used for getting data from SAP into Databricks. For now, if you’re not on SAP BDC and you want SAP data in Databricks, you will still need to rely on 3rd party solution or use a sub-optimal solution based on OData.
Actually, there are perhaps some further benefits for Databricks user: Have you noticed how well Databricks AI assistant ‘knows’ SAP? It’s a great help with generating metadata and tags for Unity Catalog so business users can easily find the SAP data they are looking for in their analytics applications.
New feature: Insight Applications & data packages
Insight Applications & data packages are end-to-end solutions, offering data transformation and modelling, reporting and AI driven insights for a specific business function. For those familiar with ‘Business Content’ think of this as business content on steroids. Or, think of it as Google Cortex, which does exactly the same.
SAP has promised that the Insight Applications & data packages will provide integrated insights across all SAP products, not just S/4HANA. For now though, the scope is limited to just ERP. If SAP lives up to its promise, then I can see this will be a great way to deliver value to the business very early on in a migration/implementation process. It is not yet clear to me how you can extend the content packages with customised tables or how you integrate non-SAP data in standard models. I guess we’ll have to wait a bit longer, perhaps until we can actually get hands-on with this.
Final thoughts
I believe SAP and Databricks are onto something great here. SAP DataSphere took a while to mature but in the last year or so the platform has become quite good. There were two major problems left: the poor integration with popular programming languages and the ability to scale at competitive prices. With the Databricks partnership, these problems should now be resolved, and SAP can once again become a leader in the data platform space.
Customer who are still on SAP BW have no reason to drag their feet any longer. SAP BDC is the future and you’re missing out on all innovation if you stick to SAP BW.
It will be interesting to see Snowflake’s response to today’s announcement. For a long time, Snowflake was almost the de-facto choice for customers moving away from SAP data platforms. Now, the ability to integrate data from SAP within Snowflake is lagging compared to Google (Cortex) and Databricks. Snowflake have been promoting SNP Glue as a preferred tool for SAP data integration. Perhaps Snowflake responds with a strategic acquisition as well?
I do hope SAP will take a sensible approach to customers who have good reasons to not follow with SAP BDC. There are other great data platforms out there – Snowflake, Databricks, Google Big Query – and customers have their own reasons why they might prefer one of those. Delta Sharing for SAP data could be incredibly beneficial for all these platforms. Customers will be happy to pay a fee for this service. Surely SAP could come up with a commercial model for this to work, and truly live up to their statement they put the customer first?