Why Multi-Cloud Discount Optimization Is Harder Than Single-Cloud

One cloud is hard enough

Optimizing discount coverage on a single cloud provider is already a full-time job. AWS alone has Reserved Instances, Savings Plans (Compute and EC2), and Enterprise Discount Programs — each with different commitment lengths, payment structures, and coverage scopes. But at least the rules are consistent. You're working within one system, one billing model, one set of APIs.

Add a second cloud and the difficulty doesn't double — it compounds. Add a third and most organizations simply give up on coordinated optimization. They run three separate playbooks, managed by three separate teams, with no cross-cloud strategy. The result is predictable: one cloud gets optimized well, and the other two bleed money at on-demand rates.

Three discount models that don't translate

The fundamental problem is that AWS, Azure, and GCP each designed their discount instruments independently, with different flexibility tradeoffs:

AWS Reserved Instancesare tied to specific instance families, sizes, and regions. They offer the deepest discounts (up to 72% for 3-year all-upfront) but the least flexibility. AWS Savings Plans loosen the constraints — Compute Savings Plans apply across instance families, regions, and even Fargate/Lambda — but at a lower discount rate. The two instruments overlap, interact, and apply in a specific order that affects coverage calculations.

Azure Reservationsfollow a similar concept to AWS RIs but with different scoping rules. Azure lets you scope reservations to a single subscription, a resource group, or share them across the entire enrollment. Their exchange policy — which lets you swap one reservation for another — is more flexible than AWS, but the mechanics changed in 2024 and continue to evolve. Azure Savings Plans, introduced more recently, add another layer of coverage with their own priority rules.

GCP Committed Use Discounts come in two flavors: resource-based (committing to specific vCPU and memory quantities in a region) and spend-based (committing to a dollar amount per hour). Resource-based CUDs give deeper discounts but lock you into specific machine shapes. Spend-based CUDs offer flexibility similar to AWS Savings Plans. Neither maps cleanly onto the AWS or Azure model.

The coordination gap

Here's where multi-cloud optimization breaks down in practice. Each cloud's discount instruments require different data, different analysis, and different purchasing decisions. An AWS RI recommendation engine doesn't help you with GCP CUDs. Azure's reservation advisor doesn't account for workloads you could shift to AWS.

Consider a company spending $4M on AWS, $2.5M on Azure, and $1.5M on GCP. The AWS team achieves 75% coverage through a mature RI and Savings Plan strategy. The Azure team, smaller and stretched across other responsibilities, manages 35% coverage. GCP, treated as a secondary platform, sits at 15% coverage. The weighted average across $8M in total spend is about 52%.

At typical discount rates, the gap between 52% and 90% coverage represents roughly$1.3M in annual savingsthat nobody is capturing. Not because the discount instruments don't exist on Azure and GCP — but because the team with the expertise is focused on AWS.

Timing and lifecycle complexity

Each cloud's commitments expire on independent schedules. AWS RIs might expire quarterly. Azure reservations might be on annual cycles that don't align with the fiscal year. GCP CUDs have their own renewal windows. Without a unified calendar, expirations get missed, auto-renewals happen at stale terms, and gaps in coverage go unnoticed for weeks.

The pre-expiration window — the 60 to 90 days before a commitment ends — is when the real analysis needs to happen. Should you renew at the same terms? Convert to a more flexible instrument? Right-size the commitment based on current usage? Let it lapse because the workload is moving to another provider? These decisions require cross-cloud visibility that siloed teams don't have.

What coordinated optimization actually looks like

Solving this requires treating discount coverage as a single portfolio across all three clouds. That means one team (or one partner) with the tooling and expertise to analyze usage patterns, commitment utilization, and savings opportunities simultaneously across AWS, Azure, and GCP.

A managed discount service that covers all three providers can apply instruments with 30-day terms to capture savings that 1-year and 3-year commitments miss — workloads that are temporary, seasonal, or in flux between clouds. Coverage moves from the typical 40–55% range to 90–95%, because every dollar of spend gets evaluated regardless of which provider it sits on.

If you're running multiple clouds and your discount strategies are managed independently, you're already leaving significant savings on the table. The mechanics of each provider's discount model are genuinely different — and that's exactly why coordinating across them requires a deliberate, unified approach rather than three parallel efforts that never talk to each other.