Our Honest Take on SAP’s AI Pivot: A Strategic Smoke Screen for a €2 Billion Shortfall
Verdict at a glance
- The Impressive: SAP is finally acknowledging reality by softening its "cloud-only" stance for AI, potentially allowing legacy "Private Edition" users access to the Joule platform.
- The Disappointing: The shift to AI appears less like a technological breakthrough and more like a tactical distraction from a massive €2 billion failure to meet cloud migration targets.
- Who it’s for: Large enterprises stuck on legacy ECC systems who need "bite-sized" innovation without the trauma of a full ERP "rip-and-replace."
- Price/Performance Verdict: Highly uncertain. SAP is moving away from per-user pricing toward "AI consumption" models—a move that could either save costs or lead to unpredictable, skyrocketing "API-style" bills.
What’s actually new
Strip away the marketing gloss, and the recent reorganization at SAP reveals two concrete shifts in strategy.
First, the structural pivot. CEO Christian Klein is moving Thomas Saueressig—previously the head of product engineering—into a newly created "Customer Value Group." This isn't just a title change; it signals that SAP’s biggest challenge is no longer engineering the cloud, but proving to its massive, reluctant installed base that there is any financial reason to move there.
Second, the commercial pivot. SAP is openly admitting that the "per-seat" licensing model is a dead end in the age of automation. If SAP’s AI tools successfully automate tasks, companies will naturally employ fewer people. Under the old model, SAP would be punished with lower revenue for making their customers more efficient. The "new way of charging for AI consumption" mentioned in reports is an essential, albeit late, attempt to decouple SAP's revenue from human headcount.
The hype check
SAP is framing this as an "innovation" focus, but the numbers tell a story of stagnation.
The Claim: SAP is leading a "cloud and AI-powered" transformation. The Reality: SAP is currently trailing its own 2025 revenue targets by approximately €2 billion. In 2022, the company promised investors that on-prem support revenue would drop to €8.5 billion as customers migrated. Instead, it sat at €10.5 billion for 2025—a meager 7% drop.
When SAP talks about a new focus on AI, they are using the tech industry's favorite shiny object to draw eyes away from the fact that their "RISE with SAP" program has failed to move the needle for their most complex, high-value customers. The "innovation" here isn't just the AI itself; it's the creative accounting and sales restructuring required to cover a 24% shortfall in their migration strategy.
Real-world implications
For the CTO of a global manufacturing firm, this shift is a double-edged sword.
- Lowered Barriers to Entry: The shift toward "brownfield" migrations and allowing AI tools like Joule to work on Private Edition models means you might not have to endure a decade-long ERP overhaul just to get modern features.
- Licensing Complexity: Expect a "Wild West" period of negotiation. As SAP moves away from per-user licensing, procurement teams will face a steep learning curve in calculating the ROI of "consumption-based" AI. If your automated workflows scale, your SAP bill might scale just as fast, regardless of your headcount.
- The 2033 Pressure Cooker: While SAP has softened its stance on where you run your software, the 2033 support deadline for ECC (with a migration plan) remains a looming cliff. The AI pivot is designed to make that transition feel like an "upgrade" rather than a forced "migration."
Limitations they’re not talking about
The source material highlights a glaring omission in SAP’s narrative: The Data Gap.
AI is only as good as the data it consumes. SAP’s legacy ECC customers are often running highly customized, fragmented, and "dirty" data environments. You cannot simply "upsell" an AI license to a company whose core data structure is a 20-year-old mess and expect "innovation."
Furthermore, SAP has ceased publishing specific figures on ECC migration since rebranding its products. This lack of transparency suggests that the "10,000 customers" Gartner expects to remain on legacy systems by 2030 are the largest, most complex organizations. For these giants, SAP’s new AI unit (staffed by "hundreds of people") is a drop in the bucket compared to the architectural debt they face.
How it stacks up
Compared to Microsoft (Dynamics 365) and Salesforce, SAP is in a defensive crouch. Microsoft and Salesforce have integrated AI (Copilots and Agents) into existing subscription tiers or clear add-ons. SAP, however, is trying to use AI as a "carrot" to fix a "stick" problem—the stalled migration of their core ERP. While Microsoft can offer AI as a feature of the cloud you’re already on, SAP has to use AI as the reason to finally get you into the cloud.
Constructive suggestions
If SAP wants this pivot to be viewed as genuine innovation rather than a desperate revenue grab, we suggest the following:
- Normalize the Data Layer First: Instead of pushing Joule, SAP should prioritize AI-driven data cleansing tools that work on legacy ECC systems to prepare them for the cloud. Make the "path to AI" a toolset, not just a destination.
- Price for Outcomes, Not Just "Consumption": Consumption-based pricing often feels like a tax on usage. SAP should lead the industry by tying AI licensing to verifiable productivity gains or business outcomes, aligning their success with the customer's.
- Radical Transparency: Resume publishing clear, audited migration numbers for ECC and S/4HANA. Trust is low after the €2B target miss; transparency is the only way to rebuild it with the C-suite.
Our verdict
Wait and Watch.
If you are one of the 10,000+ companies still on ECC, do not let the AI hype rush you into a "RISE" contract you aren't ready for. The fact that SAP is softening its stance on Private Edition access to Joule suggests that if you wait, the terms for legacy-adjacent innovation will likely become even more favorable.
However, if you are already on S/4HANA Private Edition, the upcoming "Customer Value Group" initiatives might offer a lower-friction way to pilot AI without a total system overhaul.
FAQ
Should we switch from our current on-prem SAP to the cloud just for the AI features?
No. Based on the current shortfall and the complexity of the "€2 billion gap," the migration risk still outweighs the "innovation" reward for most. Wait until SAP proves that their "AI consumption" pricing is stable and that Joule provides measurable ROI on "brownfield" data.
Is the 2030/2033 deadline for ECC support real, or will it move again?
SAP is under immense pressure to maintain its maintenance revenue (which was €2B higher than expected because people stayed on ECC). While the deadlines are officially set, the "softening" of their AI stance suggests they are prepared to support legacy environments longer than they publicly admit, provided they can find new ways (like AI licenses) to monetize those users.
How will the new "AI consumption" pricing affect our budget?
Expect volatility. Moving away from per-user licensing means your SAP costs will no longer be tied to your HR headcount. If your company heavily adopts AI automation, your SAP costs could increase even as your human staff decreases. Ensure any new contract includes "caps" or "circuit breakers" on consumption billing.
Sources
All technical specifications, pricing, and benchmark data in this article are sourced directly from official announcements. Competitor comparisons use publicly available data at time of publication. We update our coverage as new information becomes available.