SAP's AI Push Cuts Both Ways: New Governance Tools Arrive as Investors Demand Proof of Payoff

SAP spent the week stacking up announcements — a sovereign cloud pact with NTT, a new AI foundation model for tabular data, governance controls for AI agents, even a planned run of its RPT model aboard the International Space Station. None of it, however, came with a quantified revenue figure. And that gap sits at the heart of the debate now swirling around the stock.

The company's third-quarter report, released on October 21, showed cloud revenue of €6,668 million against total revenue of €10,270 million. Reuters noted it was the first time SAP had cleared the €10 billion mark in a single quarter. Group revenue grew 13%.

Cloud is doing the heavy lifting behind the more confident guidance. SAP now expects both cloud revenue and combined cloud and software revenue to land at the upper end of its existing ranges — a refinement inside the current framework rather than a lift above its prior ceiling. For 2026, that means the top of a €36.3 billion to €36.8 billion band on a currency-adjusted basis.

The Current Cloud Backlog, disclosed in the October 22 publication, stood at €24.1 billion, up 26% at constant currencies. That near-term bookable pipeline is one of the strongest arguments for confidence in the road ahead.

Price Targets Rise, Sector Nerves Fray

Goldman Sachs lifted its SAP price target to €258 from €230 on October 23, keeping a buy rating. The broker pointed to the strength of the Current Cloud Backlog and expected SAP to keep widening its lead over rivals. Other research houses also raised their targets in the wake of the quarterly numbers.

The upbeat view on SAP's business trajectory ran into a gloomier read on software as a whole. On Thursday, disappointing quarterly reports from ServiceNow and Fortinet weighed on SAP shares, with investors fretting that AI is squeezing traditional software business models — and questioning whether the new offerings are yet paying off in measurable terms. The stock had initially climbed after the results before surrendering part of those gains, a choppy reaction that suggests strong cloud numbers alone are not enough to quiet doubts about AI's economic return.

Cheaper Models, Bigger Bills

Comments from finance chief Dominik Asam on Wednesday fit squarely into that discussion. SAP is swapping out some costly AI models for cheaper alternatives inside its own products, a move expected to contribute more than €2 billion in efficiency gains over the next two years. Yet SAP also anticipates AI spending rising as usage grows — so cheaper models do not automatically translate into a lower total cost base.

The product news adds another layer. Friday's unveiling combined a new AI foundation model for spreadsheet-style data with expanded developer tooling and additional governance functions for AI agents — an build-out on both the capability and the control side of the equation. Earlier in the week, SAP and HPE announced a planned deployment of the SAP RPT model on the International Space Station, itself a project rather than a live installation. Wednesday's NTT collaboration, focused on sovereign cloud offerings built on SAP Sovereign Cloud On-Site and NTT infrastructure, operates on a different plane again: it concerns the cloud layer on which such applications would run.

For investors, the product rollouts and the partnership are best assessed separately from the reported business. Both widen the announced possibilities; neither on its own demonstrates commercial success. The cloud figures offer a solid operating foundation, but a more convincing verdict on the AI strategy will require usage, costs and earnings to prove themselves together.