SAP turns AI into the reason to move ERP to its cloud

SAP reported the first €10bn quarter in its history on 21 October. Its third-quarter statement, issued from Walldorf that evening, put total revenue at €10.3bn (£8.6bn), up 13%, and non-IFRS operating profit at €3bn (£2.5bn), up 18%. Cloud revenue rose 26% to €6.7bn (£5.6bn) and Cloud ERP Suite revenue rose 29% to €5.9bn (£5bn). Current cloud backlog reached €24.1bn (£20.2bn), up 28%.

The lines moving the other way are the ones most established SAP customers still pay for. Software support revenue, the fees on software customers license and run themselves, fell 5% to €2.4bn (£2bn). Software licenses fell 9% to €147m (£123m) and services fell 4% to €1bn (£856m). SAP’s outlook expects the support decline rate “to accelerate in the coming years as a consequence of an acceleration of customers transforming to the cloud”.

Chief executive Christian Klein put AI at the center of the explanation. “Our Autonomous Enterprise strategy is not just resonating, it is being adopted,” he said in the statement. “Our customers are investing in the technology behind it – from the cloud and data foundation to our Business AI.” Speaking on the earnings call that evening, he went further: “Nine out of 10 of our 50 largest deals included AI and Business Data Cloud.” Answering an analyst later, he said “the majority of our deals are closed with AI”.

The quarter describes a company that no longer sells AI as an extra for the ERP a customer already runs. SAP sells it as the reason to move that ERP to its cloud. For a midsize firm on SAP, adopting the vendor’s AI and migrating the core system are now sold as one purchase, made on a pricing model SAP’s own finance chief says is changing.

AI as the business case

Klein made the link in the release itself. Customers, he said, “see that Joule Work makes the business case for their transformation more attractive than ever”. SAP’s outlook uses the same word for the move it expects to accelerate: customers “transforming to the cloud”. The AI is presented as what pays for the migration, rather than something bought once the migration is done.

For firms new to SAP, the route is SAP GROW. The vendor’s product page calls it “SAP’s entry point to cloud ERP, built on SAP S/4HANA Cloud Public Edition”, designed for “companies new to SAP”, and says it “helps midsize companies modernize with cloud ERP, embedded AI, and industry-ready best practices”. Joule, the page adds, “is available across all three areas”. The Q3 statement names the companies that chose GROW in the quarter, among them Bitzer, BLACKROLL, Rittal and Krannich Solar Group. On the call Klein said: “On the GROW side, AI is also a key differentiator.” He cited Greenbrier Maxion, a South American railway manufacturer that plans to use Joule for real-time operational visibility.

For the installed base, the calendar was set long before the AI pitch. An SAP spokesperson told The Register in October 2023 that the company would provide mainstream maintenance for core SAP Business Suite 7 applications “until the end of 2027, followed by optional extended maintenance until the end of 2030”.

Services spent on adoption

Chief financial officer Dominik Asam explained the fall in services revenue on the call. SAP, he said, is using its services business more deliberately to promote adoption of its AI products and “build critical mass”. He added that the shift is likely to keep pressure on billable hours and services revenue next year.

SAP is putting its own consultants’ time into getting AI running inside customers, and accepting less billable work as the price. That moves part of the cost of adoption from the customer’s implementation budget to the vendor. It also gives SAP a direct commercial interest in how much AI a customer ends up using, because of the way it intends to charge for it.

From tokens to outcomes

Asked about credits and consumption, Asam said SAP wants to “sell outcomes to customers as opposed to tokens”, and described consumption-based pricing as something it wants to “gradually ramp into the mix”. The credits SAP gives customers are amortized, he said, “spread over the life of the contract”, so he expects no cliff in reported revenue.

Klein addressed the data question in his prepared remarks. “We never upload customer data or process IP into large language models,” he said, “not for training, not for model improvement, not for any purpose.” He added that “SAP offers access to more than 70 models, both frontier and open weight”. Under that arrangement, the models on offer and the rules for customer data are set by SAP inside its platform.

SAP expects the support revenue it earns on licensed software to shrink faster in the coming years. The AI SAP is promoting, funding with its own services and planning to price by outcome sits on the other side of the migration. SAP now sells its AI and the move of the ERP to its cloud as one decision, and Asam’s pricing plan would tie more of the cost to how much of it a business uses.