SAP's revenue model is deceptively simple at the top level — sell software subscriptions to large companies — but the mechanics underneath reveal why this business is so durable and so difficult to replicate. Cloud subscriptions are now the dominant growth engine. Traditional license revenue — the old model of selling perpetual software rights — has shrunk to near-irrelevance in new bookings. The 86% predictable revenue figure (combining cloud subscriptions and software support) is the metric that explains the valuation. Revenue model: SAP earns revenue from cloud subscriptions, software support, licenses, services, and enterprise applications across finance, supply chain, HR, procurement, and analytics. The acquisitions of LeanIX and WalkMe are tactical acknowledgments that the moat only holds if the migration journey feels manageable. The old model (sell a perpetual license, collect 22% annual maintenance) generated fantastic margins but lumpy, decelerating revenue. The new model (cloud subscriptions with expansion over time) sacrifices near-term revenue per customer but creates a compounding base. For a firm that sells back-office software to other businesses — no consumer brand, no viral growth, no hardware margins — that valuation reflects something profound about where durable economic value actually accumulates. Cloud subscriptions initially generate less revenue per customer than the old license-plus-maintenance model. The S/4HANA migration cycle creates a multi-decade runway of consulting, licensing, and cloud subscription revenue as 30,000+ enterprise customers transition from legacy SAP ECC systems to the cloud-native platform. SAP's growth story comes down to one massive bet: converting 400,000+ customers from on-premise ERP to cloud subscriptions before competitors can poach them during the transition chaos. That deadline isn't just a product lifecycle decision — it's the largest coordinated enterprise software migration in history, creating a multi-year pipeline of consulting demand, subscription conversion, and platform expansion. The AI layer through Joule adds a wrinkle: customers who convert early get twenty years of transactional data feeding contextual intelligence that late movers won't have.