What Changed in July 2025
SAP's July 2025 licensing restructuring was the most significant change to its cloud commercial model in several years. The centrepiece was the retirement of the RISE with SAP Premium Plus tier — a packaging that had previously included Joule, advanced analytics, and sustainability tools within a single subscription price. SAP replaced this with a modular model where each capability is licensed separately, often on a consumption basis.
This shift was not unexpected — SAP had been moving toward unbundling since the introduction of the Business Technology Platform (BTP) as a horizontal architecture — but the speed and scope of the July 2025 changes caught many enterprise procurement teams off guard. Customers who had included Joule access and Datasphere in their RISE cost models found those assumptions invalidated overnight.
The practical consequence: organisations that had baselined their total cost of ownership (TCO) for cloud migration on an all-inclusive RISE with SAP Premium Plus subscription were facing significant budget overruns once they discovered that AI assistants, data warehousing, and generative AI features were now billed separately.
Joule: What Is It and What Does It Now Cost?
SAP Joule is SAP's generative AI copilot, embedded across S/4HANA Cloud, SuccessFactors, Ariba, and other business applications. Before July 2025, certain Joule capabilities were included in qualifying RISE subscriptions. Post-July 2025, Joule is an add-on — with a small built-in allowance that covers basic navigation and task completion, but advanced capabilities require a paid AI Unit allocation.
SAP prices AI Units at €7 per unit at list, with a minimum purchase of 100 units per year (€700). However, this minimum is misleading for enterprise deployments. A meaningful Joule rollout across finance, HR, and procurement functions will typically consume thousands of AI Units monthly. Organisations deploying Joule across 5,000+ users should plan for AI Unit costs in the range of €50,000–€200,000 per year depending on usage intensity — before any negotiated discount.
SAP offers Joule in two primary commercial forms: a Joule Base tier that covers standard conversational AI features at a lower unit consumption rate, and a Joule Premium tier that enables agentic AI workflows, autonomous task execution, and deep cross-application integration. The Premium tier consumes AI Units at a significantly higher rate. Buyers should map their actual use cases to the appropriate tier before committing to volume.
Understanding AI Units: The New Consumption Currency
AI Units are SAP's universal consumption credit for accessing generative AI services. They work similarly to BTP credits — pre-purchased in batches and drawn down as AI features are consumed. The key difference from BTP's CPEA/BTPEA model is that AI Units are specifically ring-fenced for AI workloads and cannot be applied across the full BTP service catalogue.
Several factors determine AI Unit consumption rates:
- Complexity of AI calls: Simple query completions consume fewer units than multi-step agentic workflows that orchestrate data retrieval, reasoning, and action execution across applications.
- Model used: SAP uses a mix of internally fine-tuned models and partner models (including integrations via the SAP AI Core service). Access to larger or more powerful models carries a higher consumption multiplier.
- Volume of users: While AI Units are not strictly user-based, larger concurrent user populations will exhaust an annual allocation faster. SAP's volume tiers range from 8 units per user per month (lower tiers) down to 1 unit per user per month (highest volume tier).
- Application integration depth: Using Joule embedded within a single application consumes fewer units than cross-application orchestration scenarios.
Enterprises should instrument their AI usage before committing to an annual AI Unit volume. SAP provides utilisation dashboards within the BTP cockpit, but these are only available post-deployment. For initial sizing, Redress Compliance recommends modelling three scenarios — conservative (document navigation and basic query), moderate (workflow automation in one business domain), and aggressive (cross-application agentic AI) — and negotiating flexibility to true up between tiers.
SAP Datasphere: Separate Licensing After July 2025
SAP Datasphere — the rebranded SAP Data Warehouse Cloud — is SAP's enterprise data fabric layer, designed to federate and govern data across SAP and third-party sources in support of analytics and AI scenarios. Prior to the July 2025 restructuring, some Datasphere capacity was bundled within RISE with SAP Premium Plus or available as a discounted add-on. That arrangement no longer applies.
Datasphere is now licensed on a capacity basis, measured in SAP Datasphere Capacity Units (DCUs). Pricing is consumption-driven: you pre-purchase DCU blocks, and consumption is driven by data volumes processed, number of active connections, and compute intensity of data federation tasks. Organisations that built AI and analytics workflows assuming Datasphere was included in RISE need to add DCU costs to their licensing baseline.
The S/4HANA migration context is critical here. When an organisation moves from SAP ECC to S/4HANA Cloud, the licence baseline changes — not just for core ERP, but for the surrounding data and analytics layer. Datasphere DCU requirements for a live S/4HANA Cloud environment are materially higher than for a legacy ECC environment with limited cloud integration. Failure to account for this in migration cost modelling is a common and expensive mistake.
Need to re-model your SAP AI licensing costs after July 2025?
Redress Compliance has mapped the commercial changes for 40+ organisations. We work buyer-side only.CPEA vs BTPEA: What Buyers Must Understand
SAP's BTP credit framework has two main constructs: the Cloud Platform Enterprise Agreement (CPEA) and the newer BTP Enterprise Agreement (BTPEA). Both are pre-purchase credit models, but they serve different purposes after the July 2025 changes.
CPEA is SAP's original credit pooling mechanism for BTP services. Under CPEA, enterprises commit upfront — typically annually — to a credit block that can be applied across a broad range of BTP services. The flexibility of CPEA is its main advantage: unused credits can be redirected between service types as business needs evolve. However, CPEA credits cannot be applied to certain new services introduced since 2024, including some AI Core workloads and the latest SAP Build capabilities.
BTPEA was introduced in 2024 as a refinement designed specifically for BTP-centric workloads. BTPEA governs access to next-generation BTP services — including certain SAP Business AI services — that are excluded from the CPEA catalogue. For enterprises building AI applications on SAP BTP rather than consuming embedded AI in standard applications, BTPEA is the relevant agreement structure. BTPEA credits are not interchangeable with CPEA credits, and SAP does not offer conversion between the two pools.
The practical implication: if your AI strategy combines embedded Joule use (AI Units) with custom AI application development on BTP (BTPEA), you may need to negotiate two separate credit pools alongside your core RISE or S/4HANA subscription. This creates complexity that SAP account teams often under-explain during renewal cycles.
Negotiation Strategies for SAP AI Licensing
The July 2025 restructuring created genuine negotiating leverage for enterprises, because SAP still needs large RISE customers to commit to multi-year AI Unit volumes to stabilise its AI revenue line. Buyers who understand the commercial dynamics can use that need to their advantage.
Bundle AI Units into your RISE renewal or S/4HANA upgrade deal. The most effective moment to negotiate AI Unit pricing is when you are simultaneously renewing or expanding your core S/4HANA or RISE commitment. SAP account teams have material authority to include AI Unit allocations — often €30,000–€100,000 in value — as part of a larger strategic commitment. Do not negotiate AI Units as a standalone purchase.
Challenge the consumption model before you sign. AI Unit consumption rates are not fixed — they are determined by SAP's internal product policies, which can change. Contracts should include a clause that protects your committed unit volume against mid-contract changes to consumption multipliers. Without this protection, SAP can effectively increase your functional cost by increasing the unit consumption rate of your most-used AI features.
Negotiate BTPEA credits as a block rather than service-by-service. If you anticipate building custom AI applications on BTP, negotiate a BTPEA credit block rather than pricing individual services. Block pricing yields a significantly lower effective rate than purchasing individual services piecemeal, and the pooled flexibility allows you to redirect credits as your AI roadmap evolves.
Use the S/4HANA migration baseline argument. When SAP presents AI licensing costs, point out that the migration from ECC to S/4HANA itself generates new AI Unit consumption requirements — because SAP's embedded AI features are more deeply integrated in S/4HANA than in legacy ECC. SAP should provide migration-related AI Unit credits as part of the migration incentive package, not charge for them as a new capability.
Push for a pilot-then-commit structure. SAP has shown willingness — particularly with larger accounts — to allow an AI Unit pilot period of 6–12 months at a subsidised rate before requiring full commercial commitment. This is a valuable mechanism for right-sizing your annual AI Unit requirement against real usage data rather than SAP's estimates.
The Indirect Access Dimension of AI
One dimension of SAP AI licensing that receives insufficient attention is the indirect access implication of AI-driven integrations. When Joule or other AI agents connect to SAP data through external systems, APIs, or orchestration layers, those interactions may trigger SAP's Digital Documents Licence Count (DDLC) metric — the same metric SAP uses in indirect access audits.
SAP's DDLC metric counts digital documents — including purchase orders, sales orders, invoices, and material movements — created or modified by non-human users, including AI agents and automated processes. As organisations deploy AI workflows that read and write SAP data at scale, the volume of DDLC-triggering transactions increases significantly. Without proper licence coverage for machine-to-machine interactions, an AI deployment could inadvertently create a six- or seven-figure indirect access claim.
Redress Compliance strongly recommends that any enterprise deploying Joule or third-party AI agents that interact with SAP ERP data conduct a DDLC impact assessment before go-live. This is particularly important for AI workflows in procurement automation, where the volume of purchase order and invoice documents generated can be substantial.
What to Do Now
If your organisation is currently operating under a RISE with SAP agreement signed before July 2025, you should take the following steps immediately. First, request a full commercial amendment schedule from SAP that documents exactly which capabilities previously covered by your subscription are now billed separately. SAP account teams do not always proactively disclose the full scope of unbundled services.
Second, model your AI Unit consumption requirements against your SAP AI roadmap. Do not allow SAP to size your allocation for you — their estimates typically assume near-maximum usage, while most enterprises start conservatively. Build in flexibility to increase your allocation at predictable incremental pricing rather than at SAP list.
Third, engage independent SAP licensing counsel before signing any AI Unit or BTPEA amendment. The commercial complexity introduced by the July 2025 restructuring is substantial, and the long-term cost implications of signing a poorly structured AI licensing agreement are significant. Redress Compliance operates exclusively on the buyer side and has represented enterprises in over 80 indirect access disputes and hundreds of SAP licensing negotiations.
SAP Licensing Intelligence — Direct to Your Inbox
Receive expert analysis on SAP commercial changes, audit developments, and negotiation intelligence. No vendor spin. Buyer-side only.