Consolidate when one provider hosts the overwhelming majority of a workload's data and traffic, because depth of commitment and avoided egress dominate the math. Diversify when negotiation leverage, workload placement, or resilience is worth more than the duplicated effort. Most large estates land in between: a primary cloud for the bulk, with deliberate, bounded use of a second.
What does consolidation actually save?
Concentrating spend on one provider does three things to your bill. It pushes more usage under deeper commitment tiers, since AWS Savings Plans and Reserved Instances, Azure Reservations and the Azure Savings Plan, GCP Committed Use Discounts, and OCI Universal Credits all reward volume and term. It removes cross cloud egress, which is billed every time data moves between providers. And it strengthens an enterprise agreement: a larger single commitment unlocks better AWS EDP tiers or a more favourable Azure MACC, because the AWS EDP trades multi year spend for a discount tier and the MACC carries a shortfall clause where unspent commitment is still owed.
The risk is the mirror image. A deep single cloud commitment is the lock in. Your only real negotiation leverage with a provider is the credible option of placing workloads elsewhere, and consolidation quietly erodes that option just as your spend grows large enough to matter.
What does diversification actually cost?
Running two or more clouds preserves leverage and lets you place each workload where its economics are best, for example cheaper egress and flexible compute shapes on OCI, or a specific managed service on GCP. But it adds real cost: duplicated tooling and expertise, cross cloud egress and data gravity, thinner commitment coverage on each provider because your spend is split, and a harder allocation problem. Diversification pays when those costs are smaller than the leverage and placement value you gain. It loses when teams adopt a second cloud by accident rather than by decision.
How do you decide? A worked comparison
Consider an estate spending $1M a month, 85 percent on one provider and 15 scattered. The table below models the directional impact of consolidating the scattered 15 percent versus formalising a deliberate two cloud split. Figures are indicative, verified against billing data and anonymized.
| Lever | Consolidate the stray 15 percent | Deliberate two cloud split |
|---|---|---|
| Commitment depth | Deeper tier, plus 3 to 5 points | Coverage on both, neutral to slight loss |
| Cross cloud egress | Largely eliminated | Bounded but ongoing |
| Enterprise agreement | Stronger EDP or MACC position | Two smaller positions |
| Negotiation leverage | Weaker over time | Preserved |
| Operational overhead | Lower | Higher, needs governance |
The pattern most estates should follow is clear: consolidate the accidental sprawl, the 15 percent no one chose, to capture commitment depth and kill egress, while keeping a deliberate, governed second cloud only where it earns its place. That preserves leverage without paying for sprawl.
What governs a deliberate second cloud?
If you keep a second provider, it needs the same discipline as the first. Normalize billing data with the FinOps Foundation FOCUS specification so the two compare cleanly, set explicit placement rules by workload economics, and bound egress at the architecture level. A second cloud without governance is just more surface area for waste. A second cloud with governance is leverage you can take into every renewal.
The bottom line
Consolidate the spend that drifted, not the spend that is there on purpose. The question is not one cloud versus many in the abstract; it is which dollars belong where, judged on commitment depth, egress, leverage, and overhead. Decide it with your own numbers, not a vendor's slide.
Put a defensible number on your cloud spend.
No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.
The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.