How SAP is going all in on AI

SAP is effectively asking the market to underwrite a monetization reset before the proof points exist, which usually creates a valuation air pocket: investors de-rate on uncertainty first, then re-rate only after attach rates and retention stabilize. The near-term loser is the installed-base economics model itself, because any move toward consumption pricing raises customer budget scrutiny and could delay renewals, especially for large enterprises with procurement discipline. The likely second-order winner is the hyperscaler and infrastructure stack underneath SAP’s AI workloads, since higher inference usage shifts value away from application licenses and toward compute, orchestration, and model distribution.

The biggest risk is not headline margin compression, it is a mismatch between usage growth and revenue recognition. If AI engagement rises faster than customers accept variable pricing, SAP could see a period where usage is being encouraged but monetization lags, which is the same transition risk that has punished other software names when usage-based economics were introduced too early. That gap can last 2-4 quarters before management proves conversion, and in that window the stock tends to trade off guidance credibility rather than long-term TAM.

The contrarian read is that the market may be overestimating how disruptive this is to core demand. For enterprise software, pricing model changes often look more dangerous than they are because the real moat is embedded workflow and switching costs, not the invoice structure. If SAP can package AI as a premium layer rather than a discounting mechanism, the longer-term outcome could be higher ARPU and better gross retention, but the catalyst likely needs to come from specific KPI disclosure at earnings or Sapphire, not general AI messaging.