SAP cloud revenue beats estimates, but profit miss and acquisition costs trim 2026 outlook

Cloud revenue and backlog accelerate

SAP reported second-quarter cloud revenue of €6.28 billion, a 24% increase at constant currencies, exceeding analyst expectations. The growth rate eased from 27% in the first quarter, but the current cloud backlog, which captures contracted revenue for the next twelve months, accelerated to 26% constant-currency growth and reached €22.93 billion. That was stronger than at the start of the year and signalled resilient future cloud revenue. Total revenue rose 9% to €9.88 billion. Software licence revenue continued its structural decline, falling 32% to €131 million as customers shift from upfront licences to subscriptions.

Profit misses and outlook trimmed

Non-IFRS operating profit increased 7% to €2.74 billion, missing consensus forecasts. The shortfall was partly attributed to investments in artificial intelligence and the ongoing corporate restructuring. Net profit climbed 25% to €2.21 billion. Management lowered the 2026 non-IFRS operating profit target to €11.8–12.2 billion, from €11.9–12.3 billion, and trimmed the adjusted EBIT growth outlook to 13–17%, down from 14–18%. The revisions were driven by the dilutive effect of the recently closed acquisitions of Dremio and Prior Labs. Cloud and software revenue guidance was left unchanged.

AI must move past chatbots, CFO says

CFO Dominik Asam used the earnings call to argue that enterprise AI's real returns lie beyond the "low-hanging fruits" of coding assistants and chatbots. He said the "lion's share" of AI token consumption today goes to applications where hallucinations carry limited risk. But applying AI to finance, supply chain or other core business processes is harder because errors can compound across multiple steps.

If you have some hallucinations in the process, the errors will actually compound statistically over many steps.

Asam stressed that companies need clean, governed data before AI can operate reliably, and that the most advanced model is not always the right one. "The idea that AI will solve all these problems if they are messy, legacy data silos is not true," he said, adding that such an approach comes with "extremely high token costs." Companies should use the cheapest reliable tool for each task, whether that is simple software, an open-source model or an expensive frontier model. CEO Christian Klein has been promoting an "Autonomous Enterprise" vision, with SAP-built AI agents handling business processes, a strategy unveiled at the Sapphire customer conference in April. The push is a direct response to earlier investor fears that generic AI tools could eat into SAP's core business, a concern that had been summed up by the industry mantra "KI frisst Software" (AI eats software).

Stock jumps as cloud fears ease

SAP's American depositary receipts rose as much as 8% in after-hours trading. The strong backlog print helped calm nerves that had been frayed since early 2026, when a one-percentage-point miss in cloud backlog growth triggered a 17% single-day sell-off. The after-hours surge partially reversed the damage from late January, when SAP shares tumbled after the cloud backlog grew only 25% instead of the forecast 26%. That miss had fuelled a narrative that generic AI tools could displace SAP's software. Klein's April counter-offensive at the Sapphire conference, where he laid out the Autonomous Enterprise roadmap, was designed to reclaim the AI narrative. The approaching 2027 end of mainstream maintenance for legacy ECC systems is also pushing customers toward cloud upgrades, providing a structural tailwind for the backlog and reinforcing the revenue pipeline.