Two SAP Stories Landed in the Same Week
SAP had two very different stories to tell in the first two weeks of September, and both are worth an enterprise buyer's attention for opposite reasons. The first is an execution story: a nine-month project that separated Irish dairy and nutrition group Tirlan from its former parent Glanbia's shared SAP environment without a single day of production downtime across operations in more than 95 countries. The second is a confidence story, and it is running in the wrong direction. Over roughly the same window, UBS, Santander, and AlphaValue all turned more cautious on SAP stock, citing the pace of its agentic AI rollout specifically, not its underlying ERP business.
Read together, these two stories say something sharper than either does alone. SAP can still execute enormously complex, deadline-driven migration work at a scale few competitors can match. What it has not yet convinced the market it can do is monetize the AI layer it has spent two years building on top of that core fast enough to justify current expectations. For CIOs deciding how much weight to put on SAP's AI roadmap versus its migration and integration credibility, that distinction matters more than either headline on its own.
The Carve-Out: A Real Proof Point
Tirlan's separation from Glanbia was not a routine upgrade. It was a corporate divestment with a fixed, non-negotiable deadline, the kind of project that tends to expose every weakness in a vendor's migration tooling because there is no room to slip the date. SAP's team converted Tirlan's S/4HANA environment in under five hours and completed the HCM conversion, migrating SuccessFactors data, in fifteen minutes. Getting there required analyzing, classifying, and selectively removing more than 30 billion records so that only Tirlan's own data carried over. Noel Liston, Tirlan's IT transformation manager, put it plainly: delivering the separation in nine months while keeping production running required significant planning and the right team on site.
Ray Joyce, head of SAP UKI Ireland, called corporate separations among the most demanding initiatives a business can undertake, particularly when operations cannot pause, and that framing is accurate rather than promotional. Any enterprise that has lived through a divestment, acquisition integration, or ERP consolidation knows the real risk is never the target architecture, it is the migration window. Tirlan's timeline gives SAP a genuinely strong reference case for exactly that scenario, and it is one of the more concrete, numbers-backed migration proof points to come out of the RISE with SAP ecosystem this year.
Stable Core, Movable Edges
Within days of the Tirlan news, SAP announced a partnership pairing SAP Commerce Cloud with Vercel, the company behind Next.js, built around the same underlying principle applied to a different problem. The core commerce engine, catalog, pricing, inventory, and order processing, stays put and stable. The customer-facing storefront becomes a separate, independently deployable layer that commerce teams can redesign, test, and ship without touching backend order flows. Preview deployments let teams validate changes against live SAP Commerce Cloud services before anything reaches customers, and prebuilt Next.js templates cover product discovery, cart, and checkout without a custom frontend rebuild.
The pitch, in SAP's own words, is that teams can change the storefront without changing everything behind it at the same time. That is a real architectural shift for organizations used to treating commerce platforms as monoliths where a frontend redesign meant a backend regression testing cycle. One industry survey cited alongside the announcement found that 78 percent of organizations prioritize minimal disruption during S/4HANA-adjacent migrations, which is exactly the design constraint both this partnership and the Tirlan carve-out were built around. It is a coherent architectural narrative. It is just not the same narrative Wall Street is currently reacting to.
What Is Actually Driving the Downgrades
SAP shares sit around 185 euros, down roughly 12 percent since January and 24 percent below their 52-week high of 242 euros. That alone would not justify three separate rating downgrades in a matter of weeks. What is driving the analyst moves is more specific: SAP has reportedly deployed only about 17 agentic AI use cases against an internal target of roughly 200 per year, a gap that UBS flagged directly when it cut the stock from Buy to Neutral. Santander downgraded from Outperform to Market Perform on September 1, and AlphaValue and Baader Europe went further, assigning an outright Sell rating by the end of August.
SAP is not sitting still in response. The company repurchased 676,583 shares on the Xetra exchange between August 24 and 28 under its 2026 buyback program, and board member Thomas Saueressig made a personal share purchase around the same time, a move typically read as insider confidence. Neither gesture addresses the underlying question analysts are asking, which is not whether SAP's core ERP business is healthy, it clearly is, but whether the company can convert its AI roadmap into deployed, revenue-generating agent use cases at the pace it has publicly promised.
The Backlog Tells a Different Story Than the Stock Price
Here is what makes this moment genuinely interesting rather than just another vendor stock story: SAP's cloud backlog grew 27 percent to 22.9 billion euros in the same period its shares fell. That divergence is the clearest signal available that customers are still signing multi-year cloud commitments even as public market investors grow skeptical of the AI monetization timeline layered on top of those commitments. The two audiences are evaluating fundamentally different things. Customers are buying migration reliability, integration depth, and platform stability, the things Tirlan's project actually demonstrated. Investors are pricing in a specific agentic AI adoption curve that has not yet materialized at the pace SAP guided toward.
That gap between backlog growth and stock performance is worth watching closely over the next two quarters, because it will resolve one of two ways. Either SAP's agentic AI deployment pace accelerates toward its stated target and the stock catches up to what the backlog already implies, or the AI narrative gets quietly descoped while the core ERP and cloud migration business continues compounding regardless. Enterprise buyers do not need to guess which outcome is more likely before signing a RISE with SAP contract. They need to separate the migration commitment from the AI roadmap commitment in their own contract negotiations.
What This Means for Teams Running SAP
If your organization is evaluating an S/4HANA migration, a carve-out, or a commerce replatforming, the Tirlan case gives you a legitimate, numbers-backed reference to push your SAP account team on: what would a comparable timeline and data-scrubbing approach look like for your specific separation or migration scenario. That conversation is grounded in a delivered project, not a roadmap slide, and it is worth demanding the same level of specificity SAP provided publicly about Tirlan's nine-month timeline and record counts.
On the AI side, we would treat SAP's agentic roadmap as a separate negotiation entirely, one where you ask for deployed customer references at the same evidentiary standard as the migration story, not adoption percentages or aspirational agent counts. The 17-against-200 gap analysts are pricing in is exactly the kind of detail that should show up in your own vendor risk assessment before you commit budget to SAP's AI agent licensing tier. Buy the migration credibility. Underwrite the AI roadmap separately, and revisit it every two quarters until the deployment numbers close the gap analysts are currently pointing at.
