SAP's Cheaper-Model Bet Gets Its First Stress Test in a Global AI Selloff

SAP shares fell 3.4% yesterday as a wave of selling hit technology stocks worldwide, with German media tying the move to an AI-driven retreat on Wall Street and in Asia. Reports pointed to mounting doubts over the returns on hefty AI spending and possible delays in data-center buildouts as the trigger for the sector-wide weakness.

The software patch was mixed rather than uniformly red. Salesforce, Microsoft and Workday all lost ground alongside SAP, while Oracle and ServiceNow managed gains.

For SAP, the pullback sharpens a question investors now have to weigh: can the efficiency gains the company expects from its AI push carry more weight than the market's current skepticism? The economic use of the technology offers a concrete foothold. SAP plans to swap out some expensive AI models inside its own products for cheaper alternatives — a move it told Reuters on October 28 could contribute to more than €2 billion in AI-related efficiency gains over two years.

That forecast is the yardstick. Note the wording: the swap is meant to contribute to those gains, so expectations should be calibrated to that contribution rather than to the full figure. If SAP can show a clear link between cheaper models and economic benefit as the program unfolds, the case for its AI course strengthens. If the payoff lags, the bullish thesis weakens.

A Buyback Ledger That Keeps Filling

Alongside the strategy debate, SAP disclosed on Tuesday that it repurchased 50,000 of its own shares in the October 26–30 window, a mandatory filing showing an aggregate volume of €10,625,306.00.

The company's product news has been arriving at a steady clip. On October 28, SAP and NTT announced a planned collaboration on sovereign cloud and SAP Business AI capabilities. The first step is to develop an offering that links SAP Sovereign Cloud On-Site with NTT-managed infrastructure and operations — NTT running the infrastructure and operations, with SAP's on-site sovereign cloud as part of the joint package. It is, for now, an offering still to be built.

A day earlier, at its TechEd conference, SAP unveiled additions to its developer portfolio spanning Joule Studio, Joule Work and the SAP Business AI Platform. By the company's account, these include tools for developers to create and extend AI agents, plus governance features and capabilities for developing autonomous task execution.

Two Tracks, One Verdict

The constructive case rests on combining AI offerings with cost discipline. Should the cheaper-model rollout land as planned, SAP could shore up its efficiency target — no small thing in a market gripped by worries about where AI investment is heading. The NTT tie-up adds a second leg, provided the announced cooperation turns into concrete offerings while the model swap delivers economic progress. Both tracks have to move from announcement to substance.

The cautious reading starts with the nature of the efficiency claim itself: it is a multi-year expectation, and anyone weighting it heavily is betting that the improvements become verifiable over time. If the economic contribution of cheaper models falls short, or if the NTT collaboration yields less tangible progress than hoped, the positive case takes a hit. Both risks concern execution and belong on the same ledger as the opportunities.

Market backdrop matters too. Should the AI-investment concerns Reuters described persist, SAP's valuation could stay under that cloud, leaving a bullish company view especially dependent on hard evidence. On that basis, a more reserved stance means measuring the promised benefit against firmer company disclosures first.

The next concrete checkpoint is January 28, 2027, when SAP publishes fourth-quarter and full-year 2026 results. Investors will be watching whether the company can make its AI initiatives economically tangible. Convincing specifics would bolster the positive case; disappointing progress would argue for greater caution. That is the frame for deciding how much trust the expected efficiency gains deserve today.