SAP's €2 Billion AI Efficiency Bet Meets a Buyback in Progress
Published on 10/28/2026 at 19:01 | Editorial boerse-global.de
SAP is pursuing two very different levers at once: squeezing more margin out of its artificial intelligence stack while steadily mopping up its own shares. The company disclosed on Thursday that it repurchased another 42,848 of its own shares, and separately signalled that swapping out its most expensive AI models should save €2 billion over the next two years, according to Reuters.
The buyback notice is the thirteenth interim report under the current programme. Purchases were made between 19 and 23 October via XETRA, bringing the total acquired under the scheme to 9,086,956 shares. It marks a continuation of an existing capital measure rather than the launch of a new one — a distinction investors should keep separate from the operating story, since the transactions touch the share count while the cloud figures speak to business momentum.
Cheaper Models, Not a Smaller Budget
On the AI side, the plan is not a retreat from the technology. Finance chief Dominik Asam expects the savings to come from replacing particularly costly models with cheaper alternatives, provided those alternatives deliver the required performance. Total AI spending is still expected to rise as usage grows, so the €2 billion target describes a more efficient delivery of the same work — not an absolute cut in the budget.
That distinction matters for anyone weighing the stock. The projected relief should not be treated as banked additional profit. Its value depends on whether SAP can deploy the cheaper models reliably and how strongly demand expands at the same time.
Should investors sell immediately? Or is it worth buying SAP?
The decisive metric is margin per completed customer task. Under the billing model described by Reuters, customers pay for tasks or business actions rather than for raw computing capacity consumed. If compute costs per task fall while customer prices hold steady, SAP keeps a larger slice of revenue. That is what separates this from a simple trim of the technology budget: the goal is to hit the same required quality with a less expensive model.
SAP reaches roughly 70 models through its Generative AI Hub within SAP BTP, including offerings from Azure OpenAI, Anthropic Claude and Google Gemini. That breadth creates switching options — though it does not mean every model suits every task equally well.
The Cloud Backdrop
Efficiency gains would land against a fast-growing base. SAP posted cloud revenue of €6.67 billion in the third quarter of 2026, up 26% year on year. For the full year, management expects cloud revenue near the upper end of a range of €25.8 billion to €26.2 billion, with adjusted operating profit of €11.8 billion to €12.2 billion.
Those forecasts are the yardstick. The expected AI savings cannot simply be added on top of the profit target — no such contribution has been confirmed. In the favourable case, cheaper models help SAP carry rising usage at healthy margins. The third-quarter cloud figure, however, was flattered by one-off effects, so the latest growth rate should not be extrapolated unchecked. Weaker growth would only sharpen the case for disciplined cost control.
Where the Plan Could Slip
The serious counterargument sits with customer requirements. According to Reuters, public authorities and regulated enterprises pay closer attention to where the models they use come from, and European customers increasingly ask for local solutions. The cheapest provider on paper is therefore not automatically a suitable choice. Such constraints could narrow the room for model switching, forcing SAP to juggle cost, quality and accepted provenance at the same time.
Add the explicitly expected increase in overall AI spending, and successful per-task savings do not guarantee a smaller total bill. If usage climbs steeply, the funding requirement could keep rising — leaving investors to judge whether the extra revenue covers that outlay.
SAP at a turning point? This analysis reveals what investors need to know now.
Two Signals, One Calendar
The buyback and the AI efficiency drive sit alongside a cloud order book that has drawn scrutiny of its own. On 21 October, SAP reported a current cloud backlog of €24.1 billion for the third quarter of 2026, up 28%. Its warning of a slight slowdown in the fourth quarter then weighed on the shares, compounded by profit-taking after strong results and worries about the near-term monetisation of AI.
Product news from SAP TechEd — including orchestration for multiple AI agents, governance functions and expanded data integration — adds to the operating picture, but announcements are not the same as realised additional revenue.
The next firm checkpoint is 28 January 2027, when SAP publishes preliminary results for the fourth quarter and for the 2026 financial year. That release will test growth, profit and spending in one go. For investors, the question is less how many models SAP can access than whether it can demonstrably run them more economically — while the buyback rolls on in the background.
Ad
SAP Stock: New Analysis - 28 October
Fresh SAP information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.
