The ERP comfort zone
Andreas BORN reads the valantic SAP Study 2026: private cloud at 64 percent, on-premises collapsing, public edition barely moving. Why the D-A-CH market chose process conservatism with a cloud subscription, what the Business Data Cloud numbers quietly decide, and what to do about it on Monday morning.
Based on the valantic SAP Study 2026 (techconsult, 409 companies, D-A-CH, May 2026). All figures below are from that study. The interpretation is mine.
The German-speaking SAP market has quietly made its decision, and it is not the one SAP wanted.
The numbers first. Roughly 70 percent of D-A-CH companies are now either mid-migration to S/4HANA (42 percent) or already live (28 percent). Only 5 percent have no plans at all. The transformation is no longer a strategic debate. It is an operational reality.
The interesting part is not whether companies are moving. It is where they are moving to.
The destination SAP did not choose
SAP Cloud ERP Private now runs at 64 percent of surveyed deployments, up six points in a year. Traditional on-premises has collapsed from 20 to 11 percent. And the public cloud edition, the model SAP itself positions as the strategic destination, grew by three points to 25 percent.
Read that again. On-premises lost nine points in twelve months, and the public edition captured three of them. The other six went private.
Combine that with the migration strategy data and a pattern emerges. Brownfield gained eight points. Hybrid approaches still lead at 47 percent. Greenfield, the clean-slate model that public cloud effectively requires, sits at 18 percent and is the preferred route in two out of three public cloud projects.
The study authors call the result the “ERP comfort zone”: private cloud plus brownfield or hybrid migration. Keep your processes, keep your customizations, move the infrastructure, modernize in a controlled way.
This is not cloud resistance. It is process conservatism with a cloud subscription. Companies are not saying no to SAP’s platform. They are saying no to re-justifying twenty years of process logic on a vendor’s timeline. The nine-point collapse of on-premises tells the other half of the story: with ECC maintenance ending, staying put has stopped being an option, a dynamic I unpacked in Why ECC Ends and Cloud Begins.
In Switzerland the pattern is even sharper: 75 percent private cloud, 57 percent hybrid migration, and the public edition at just 15 percent, the lowest in the region.
The migration is not the transformation
The second finding deserves more attention than it will get. For a growing majority, go-live is not the end of the programme. It is the entry ticket to the actual transformation.
Interest in the Business Technology Platform is now led by two topics in a dead heat at 51 percent each: artificial intelligence and the Business Data Cloud. Automation and integration follow. Application development, the original BTP story, trails at 31 percent.
The BDC numbers are the quiet headline of this study. Sixty percent of companies now treat the Business Data Cloud as an integral part of their data management, and among companies still in the Cloud ERP planning phase the figure is 66 percent. Meanwhile, the share running a comparable data architecture outside the SAP portfolio, on Snowflake or hyperscaler-native platforms, fell from 40 to 30 percent in one year.
That is a ten-point shift in data gravity toward the SAP-native stack, decided largely before the ERP migration even completes. Whatever you think of the architecture, the commercial consequence is plain: the data platform decision is being made inside the ERP decision, not after it. Anyone who treats these as separate procurements is already a year behind their own roadmap. I have examined what the BDC actually is, and is not, in Shared Business Context: Unified Data for Enterprise AI, and the Snowflake question specifically in How Zero-Copy Changes the Architecture.
Agents up, chat down
The AI section confirms what anyone watching enterprise adoption has suspected. Sixty-seven percent of companies are using AI or have embedded it in business processes. But the demand profile has changed.
AI agents top the wish list at 42 percent, ahead of AI-supported planning and analysis at 41 percent. The GenAI Hub, for building your own cases, draws 27 percent. And Joule, the conversational interface SAP has marketed hardest, sits at 16 percent and falling, as the study notes that early users have been “cautious in their assessment of its benefits.”
The message from the buy side could not be clearer. Nobody wants to chat with their ERP. They want work executed: tasks performed, processes initiated, exceptions handled, within defined business logic and governance. The interface was never the product. The outcome is the product. This is the shift from copilot to agentic process I described in AI Agents in ERP, and the study’s numbers confirm its second half too: agents fail not for lack of data but for lack of business context.
The barriers are equally consistent: IT security risk is the number one obstacle in all three countries, with regulatory uncertainty close behind, and 45 percent see SAP’s Sovereign Cloud as a way to address exactly these concerns. Digital sovereignty has moved from a procurement footnote to an architecture criterion.
What this means on Monday morning
Three implications for the CIO, CFO or transformation sponsor carrying this decision.
First, the comfort zone is rational, and it is still a decision. If your organisation is defaulting to private cloud plus brownfield, understand what you are buying: continuity, lower disruption, preserved customization. And understand what you are deferring: the process standardization and clean-core discipline that produce most of the long-term value. In my experience, the comfort zone is the right place to land and the wrong place to live. Enter it with a dated plan for leaving it, workstream by workstream, or the 2030s version of this study will describe your landscape the way this one describes ECC. The thinking that belongs before that choice is the subject of Every Transformation Begins Twice.
Second, make the data platform decision consciously, now. The study shows the market folding the BDC decision into the ERP decision. If that is also happening in your programme, it should happen in a governance forum, with the Snowflake and hyperscaler alternatives priced and benchmarked, not by default because the migration partner proposed the reference architecture. The SAP-native path may well win. It should win on evidence.
Third, buy AI outcomes, not AI interfaces. The market has already voted: agents over chat. Before funding any SAP AI initiative, name the process, the business logic boundary, the measurable outcome and the owner. If the proposal cannot survive those four questions, it is an interface looking for a justification. If you want to know where your organisation actually stands before that conversation, start with the Autonomous Enterprise Readiness self-assessment.
The reality check
A fair reading of this study also acknowledges what it is not. It is a vendor-commissioned survey of 409 companies, weighted toward Germany, and the direction of some findings serves its publisher. That is not a reason to dismiss it; it is a reason to hold it to the same standard I argued for in ERP ROI: Evidence over Folklore. The trends, however, align with what DSAG and the implementation market have been reporting for two years, and with what I see in projects: controlled migration, private-cloud gravity, data consolidating toward the SAP stack, and AI expectations maturing from demos to delegation.
SAP wanted the market in the public cloud by now. The market went where its processes could survive. Both parties will spend the rest of this decade negotiating the difference.
That negotiation, not the migration, is the real transformation.
Source: valantic SAP Study 2026, conducted by techconsult GmbH, 409 companies in Germany, Austria and Switzerland, published May 2026. Available at valantic.com. All interpretations and recommendations in this article are the author’s own.