Microsoft added the Agent Pre-Purchase Plan in February 2026 using Agent Commit Units (ACUs). ACUs work like CCCUs (5–20% volume discount on annual commit) but draw down against eligible consumption across Copilot Studio, Microsoft Foundry, and eventually Fabric and GitHub agentic services. ACUs are the right answer for organisations whose 2026–2028 roadmap spans multiple Microsoft agent platforms. For Copilot-Studio-only organisations, CCCUs provide the same discount with simpler commercial structure. Default to CCCU unless the cross-platform need is real and immediate.
What ACU is
ACU stands for Agent Commit Unit, introduced in February 2026 as part of Microsoft’s unified agent commercial framework. ACUs are pre-purchased in annual pools at discounted pricing, consumed against eligible agent operations across the supported platforms, and fall back to pay-as-you-go pricing if the pool exhausts.
The structural difference from CCCU is the scope of eligible consumption. CCCUs cover Copilot Studio specifically. ACUs cover Copilot Studio AND Microsoft Foundry (Microsoft’s agent development platform), with planned expansion to Fabric agent services and GitHub agentic services. The ACU is positioned as the future unified commit unit for the Microsoft agent ecosystem.
ACU vs CCCU comparison
| Aspect | CCCU | ACU |
|---|---|---|
| Coverage | Copilot Studio only | Copilot Studio + Foundry (Fabric, GitHub planned) |
| Discount tiers | 5–20% | 5–20% (parallel structure) |
| Annual pool | Yes | Yes |
| Spillover to pay-as-you-go | Yes | Yes (across covered platforms) |
| Cross-platform flexibility | No | Yes |
| Commercial complexity | Lower | Higher (cross-platform allocation) |
When ACU is the right answer
Three customer profiles consistently benefit from ACU over CCCU.
Profile 1: Multi-platform agent deployment in progress. Organisations actively building agents in both Copilot Studio and Microsoft Foundry simultaneously benefit from a unified commit pool that covers both platforms. The flexibility to shift consumption between platforms as deployment evolves is meaningful.
Profile 2: Strategic Microsoft customers with broad agent roadmap. Large enterprise customers whose 2026–2028 strategy explicitly spans the Microsoft agent stack — Copilot Studio for business agents, Foundry for developer agents, Fabric for analytical agents, GitHub for code agents — need a unified commercial framework. ACU is that framework.
Profile 3: Organisations consolidating from multiple AI platforms. Organisations migrating from non-Microsoft AI platforms to Microsoft’s agent ecosystem benefit from the optionality ACU provides during migration. Consumption can shift between Microsoft platforms based on where workloads land.
When CCCU is the right answer (despite ACU existing)
Most enterprises in 2026 are still Copilot-Studio-only. For these organisations, CCCU provides the same discount with simpler commercial structure and no need to allocate consumption across platforms. The cross-platform flexibility ACU provides is not actionable.
Three signals indicate that CCCU is the right answer despite ACU’s existence: no current Foundry deployment, no plans for Foundry in the next 12 months, and no roadmap consolidation across multiple Microsoft agent platforms. Organisations matching all three should default to CCCU.
Negotiation considerations specific to ACU
Two ACU-specific negotiation considerations.
Platform allocation flexibility. ACU commits should explicitly preserve the customer’s right to allocate consumption across covered platforms without prior Microsoft approval. Default ACU language sometimes includes allocation reporting requirements that limit flexibility in practice. The protective clause: “Customer may allocate consumed ACUs across covered platforms at its discretion, without prior approval, subject to standard usage reporting.”
Future platform inclusion. Microsoft has stated that Fabric and GitHub agent services will be added to ACU coverage. The timing of additions is not contractually guaranteed unless specified. For organisations with explicit Fabric or GitHub agent roadmaps, the contract should specify expected inclusion timing and provide adjustment mechanisms if Microsoft’s timeline slips.
Action plan for ACU evaluation
- Map your current and planned agent platforms. Copilot Studio only? Add Foundry? Plus Fabric? Plus GitHub? The platform scope determines the right commit unit.
- Default to CCCU unless multi-platform. Simpler commercial structure for the same discount when only Copilot Studio is in scope.
- If considering ACU, negotiate allocation flexibility. The platform-allocation rights are negotiable and important.
- Document the platform-expansion roadmap. If signing ACU with future Fabric/GitHub expectations, ensure the contract reflects timing.
- Engage independent advisory. ACU vs CCCU is a real decision and benefits from external perspective. Book a scoping call.