Contents
SAP's July 2025 commercial restructure was the most significant change to SAP's AI and data licensing since the introduction of BTP credits in 2020. The restructure did three things simultaneously: it introduced the AI Units consumption model as the universal pricing currency for all Joule and embedded AI features; it restructured the Business AI subscription tiers; and it reorganised SAP Datasphere, SAP Analytics Cloud, and related data products under a unified data licensing framework.
The problem is that SAP communicated these changes primarily through product update notes, pricing page revisions, and account team briefings — not through proactive outreach to customers whose contracts were affected. Many enterprises discovered the commercial implications of July 2025 only when their account team began the renewal conversation and quoted prices that bore no resemblance to their previous contract structure.
This guide provides a complete, buyer-side analysis of what changed, what it means for your licence position, and what commercial actions your organisation should take before the next renewal. For broader context on SAP's AI commercial direction, see our guide on SAP's shift to use-based AI pricing.
Want an Independent View of Your SAP Position?
Our advisors are former SAP insiders working exclusively for enterprise buyers. A free 30-minute discovery call will tell you whether independent advisory would materially change your commercial outcome.
Book a Free Consultation → Download Free SAP Audit Guide →What Changed in July 2025
SAP's July 2025 restructure affected four areas of the commercial framework:
The cumulative effect of these changes is that most enterprises with SAP AI or data product investments are now operating under a fundamentally different commercial framework than the one they contracted against — often without having been explicitly informed of the change.
The AI Units Model: What Replaced What
Before July 2025, SAP AI licensing was fragmented across multiple commercial vehicles. Joule access was controlled by product entitlement flags within S/4HANA and SuccessFactors subscriptions. AI Foundation services on BTP were priced as BTP service credits, with each service having its own credit consumption rate. Document information extraction, machine translation, and anomaly detection each had separate service descriptions and billing mechanics.
The July 2025 AI Units model replaced this fragmentation with a single consumption metric — but the consolidation benefited SAP commercially more than it benefited buyers:
Joule access: feature flag in product subscription. No incremental AI charge for standard copilot interactions. BTP AI services: separately metered BTP credits at per-service rates.
Joule access: AI Unit consumption applies to all interactions including standard copilot. BTP AI services: consolidated under AI Units with SAP-controlled rates. No more per-service differentiation visible to buyer.
AI consumption was predictable at the service level — you could model BTP credit consumption for each AI service independently with published rates per call.
AI Unit consumption depends on SAP's black-box conversion model. Published "typical" rates are illustrative only. Actual consumption depends on model version, input complexity, and service configuration — all SAP-controlled variables.
The consolidation premium: Enterprise buyers who moved from BTP service credits to AI Units under the new model typically experienced effective cost increases of 15–35% for equivalent AI workloads, when measured by actual business outcomes delivered per monetary unit of AI spend. SAP's account teams framed this as "simplification" — which it is, from SAP's billing perspective. From the buyer's perspective, it is a repricing that was not disclosed as such.
Business AI Package Restructure
SAP's Business AI offering was reorganised into two subscription tiers in July 2025: Business AI Base and Business AI Premium. Both include AI Unit allocations and access to Joule capabilities across the SAP product portfolio, but with significant differences in scope, consumption limits, and which premium features require additional AI Unit expenditure.
Business AI Base
The Base tier covers core Joule copilot functionality across embedded SAP products — conversational interfaces, document summarisation, process recommendations, and standard generative AI responses. The AI Unit allocation included in Base is sized for moderate conversational use cases. Agentic workflows, custom model training, and advanced analytics AI capabilities are not covered under Base and require either an upgrade to Premium or additional AI Unit purchase.
The critical commercial issue with Business AI Base: the AI Unit allocation is calculated based on SAP's internal model of "typical" usage patterns, which assumes primarily conversational copilot interactions at low agentic complexity. Organisations that have deployed Joule Agents or custom BTP AI workflows will exhaust Base allocations faster than SAP's model predicts — often within the first quarter of a new contract year.
Business AI Premium
The Premium tier includes the full Joule Agent capability set, higher AI Unit allocations, access to advanced AI Foundation services on BTP, and early access to new AI features as SAP releases them. SAP's list price for Premium represents approximately 2.5–3x the Base tier cost. However, for enterprises with significant agentic deployment plans, Premium provides better effective unit economics than purchasing Base plus top-up AI Units at non-bundled rates.
An important nuance that SAP's commercial team does not proactively disclose: the AI Unit rates for Joule Agent operations in Premium are still not publicly specified in the subscription documentation. You are purchasing a bucket of units at a bundled rate without knowing the per-operation cost of the specific agent workflows you intend to deploy. This ambiguity is intentional — it preserves SAP's ability to characterise high consumption as usage-driven growth at renewal.
Navigating the July 2025 Changes at Your Next Renewal?
Most enterprises don't discover the commercial impact of SAP's July 2025 restructure until their account team presents a renewal quote. Our SAP contract negotiation team provides pre-renewal analysis of how the new AI and data licensing framework affects your specific position and what commercial protections to demand.
Get Pre-Renewal AnalysisSAP Datasphere Licensing Changes
SAP Datasphere (formerly SAP Data Warehouse Cloud) underwent a significant licensing restructure as part of the July 2025 changes. SAP moved Datasphere towards a multi-dimensional pricing model that combines three metrics: user-based access, data volume capacity, and compute hours for AI-driven data workflows.
Before July 2025, most Datasphere customers were on capacity-unit pricing from the Data Warehouse Cloud era — a relatively straightforward metric based on storage and compute capacity. After July 2025, the picture became more complex:
- User licences — Business User, Modeller, and Data Integration licences at different price points, replacing the previous unified capacity unit model for user access
- Storage and compute capacity — SAP HANA Cloud capacity (measured in compute and storage blocks) underpins Datasphere and is priced separately from user licences
- AI Data Processing Units — a new metric introduced in July 2025 for AI-powered data transformation, semantic enrichment, and ML model training workflows within Datasphere
The third metric — AI Data Processing Units — is the one that caught most enterprises by surprise. Any Datasphere workflow that uses SAP's embedded AI capabilities (intelligent data discovery, automated data quality, ML-based anomaly detection in data pipelines) now consumes AI Data Processing Units in addition to standard compute capacity. Organisations that built sophisticated data pipelines in Data Warehouse Cloud and migrated to Datasphere found that their workflows were consuming a new metric that their existing contracts didn't cover.
Transition exposure for former Data Warehouse Cloud customers: If your organisation migrated from SAP Data Warehouse Cloud to Datasphere before July 2025, your migration contract may have been structured under the old capacity unit framework. SAP's transition offer for these customers typically involves a credit conversion at rates that favour SAP's revenue recognition. Before accepting any Datasphere contract revision, get an independent assessment of whether the credit conversion rate is commercially fair. Our SAP licence optimisation team has reviewed multiple Datasphere transition proposals and consistently found conversion rates 20–40% below what independent analysis supports.
BTP AI Foundation and Service Changes
SAP's AI Foundation service — the BTP layer that provides enterprise AI capabilities including document processing, entity extraction, machine translation, and the AI API framework — was restructured in July 2025 to align with the AI Units model.
Before July 2025, individual AI Foundation services were priced as BTP service credits at rates published in SAP's BTP Service Catalogue. Enterprise architects could calculate exactly how many BTP credits each AI-powered process step would consume and model full workflow costs with precision.
After July 2025, AI Foundation services consume AI Units through a conversion model where BTP service credits for AI workloads are converted to AI Units at rates that SAP controls and can revise. This creates two layers of indirection between what you buy (AI Units) and what you get (business outcomes from AI Foundation services) — with SAP controlling the conversion rate at both layers.
The specific AI Foundation services affected by the July 2025 restructure include:
- Document Information Extraction (INVOICE, PAYMENT ADVICE, PURCHASE ORDER processing)
- SAP AI Core (model training, inference endpoints for custom models)
- Business Entity Recognition and data extraction services
- Machine Translation and language services
- Anomaly Detection and predictive analytics services
Organisations with significant document processing volumes — particularly those using Document Information Extraction for accounts payable automation — need to re-model their consumption under the new AI Units framework. The rates that applied under the BTP service credit model do not translate directly to AI Units, and in most cases the effective cost per document processed increased between 10–25% in the transition.
Impact on RISE and GROW Entitlements
SAP's July 2025 restructure materially changed the AI entitlements bundled with RISE with SAP and GROW with SAP subscriptions — in ways that disadvantaged customers who did not proactively renegotiate their subscriptions at the time of the change.
RISE with SAP subscriptions signed before July 2025 typically included access to Joule as an embedded feature of S/4HANA Cloud, with no separate AI Unit consumption charge for standard copilot interactions. After July 2025, SAP's position on RISE renewals changed: AI Unit consumption was introduced as a variable cost layer on top of the subscription fee, with an initial "included allocation" designed to cover light copilot usage but insufficient to support any meaningful agentic deployment.
The commercial impact at RISE renewal is significant: organisations that signed multi-year RISE agreements before July 2025 may find that their renewal quote includes a mandatory AI Unit add-on that was not in the original deal structure. SAP's account teams will present this as customers "upgrading" to the new platform capabilities — which technically they are. What they will not present is the fact that the AI Unit add-on cost was implicitly included in the older RISE pricing and is now being separated out as an incremental revenue line.
Our RISE with SAP advisory team negotiates AI Unit inclusions as a component of every RISE renewal, challenging SAP's framing that AI Units are a new cost that wasn't in the original contract value.
Transition Risks for Existing Customers
Three categories of existing customers face the highest transition risk from the July 2025 changes:
1. BTP-Heavy AI Users
Organisations that built significant AI automation on BTP using AI Foundation services under the pre-July 2025 service credit model may find their existing processes have become materially more expensive under the AI Units framework. If your organisation has an annual BTP AI Foundation spend that translated to significant document processing or inference volume, a cost audit of your post-July 2025 actual spend vs. pre-July 2025 projected spend should be a priority action.
2. Datasphere / Data Warehouse Cloud Migrators
As described above, organisations that migrated from Data Warehouse Cloud to Datasphere under transition offers need to verify that their current contract covers AI Data Processing Units. Uncontrolled AI data processing in Datasphere pipelines can generate significant unbudgeted consumption.
3. SuccessFactors and Joule Early Adopters
Organisations that adopted Joule early within SuccessFactors — particularly those using Joule for HR process automation — may have done so under promotional or pilot terms that pre-date the AI Units model. When those arrangements roll into standard commercial terms, the consumption charges will apply retroactively to the activation patterns already established in production. There is no grandfather provision in SAP's standard terms for pre-commercial Joule usage.
Renewal Implications and Action Plan
For enterprises with SAP renewals in 2026 or 2027, the July 2025 changes create specific commercial risks that can be mitigated through proactive preparation.
Twelve months before renewal: Commission an independent AI and data consumption audit. Map every AI feature currently active in your environment against the July 2025 commercial framework. Quantify your actual AI Unit consumption (from BTP Cockpit and SAP for Me) and compare against your current entitlements. Identify any consumption occurring outside your contracted allocation.
Six months before renewal: Engage SAP in a pre-renewal commercial discussion before SAP's account team presents a formal renewal quote. Entering the renewal conversation with your own consumption analysis gives you the ability to challenge SAP's framing of what you owe. Without your own data, you are negotiating blind against SAP's telemetry.
At renewal: Negotiate AI Unit allocations, rate locks, and consumption caps as core elements of the renewal deal — not as add-ons negotiated after the core terms are agreed. Our complete framework for negotiating SAP AI contracts provides the specific commercial language you need. The renewal moment is when you have maximum leverage — SAP needs the renewal revenue and will accept better terms at close than they will six months later when you are mid-contract with no negotiating pressure.
Key Takeaways
- SAP's July 2025 restructure introduced AI Units as the universal AI pricing currency, replacing fragmented BTP service credit and feature-flag models — with commercial terms that benefit SAP's revenue flexibility
- Business AI Base and Premium tiers replaced previous Joule entitlement structures; Premium is required for agentic capabilities but rates per operation remain opaque
- SAP Datasphere introduced AI Data Processing Units — a new metric that catches organisations with data pipeline AI automation off-guard
- BTP AI Foundation services moved to AI Units; effective cost per document or inference increased 10–25% for most organisations in the transition
- RISE and GROW renewals post-July 2025 include AI Unit add-ons that SAP presents as new capability but were implicitly included in original pricing
- Organisations that signed RISE or BTP agreements before July 2025 should audit their current consumption against their contracted entitlements before the next renewal
- Pre-renewal consumption modelling is the most effective way to protect your commercial position in an AI-heavy SAP environment
Independent SAP Licensing Advisory
Audit defence, contract negotiation, licence optimisation — buyer-side only, zero SAP affiliation.
Explore All Services → Case StudiesReal Results for Enterprise Buyers
See how we've helped enterprises reduce SAP spend by 30–60% and win audit disputes.
Read Case Studies →