Egress rarely exceeded 3 to 5 percent of a year's commitment, and the exit plan earned 8 to 15

Lock in is real, measurable, and mostly reversible. The cost of leaving is consistently smaller than teams fear, and the cost of staying without a credible alternative is consistently larger, which makes the exit plan a negotiating instrument rather than a migration project.

Prepared by Redress Compliance · August 15, 2026 · AWS advisory. Based on 35 to 45 AWS commitment and exit reviews, 2024 to 2025.

• One portable workload, chosen for low data gravity and kept genuinely movable, which buys most of the leverage of full multi cloud at a fraction of the cost.
• A costed migration path, with egress, rebuild effort and a timeline documented for that workload rather than described in principle.
• Surgical dual sourcing, used only where it lowers a specific risk, since multi cloud adds tooling and skills cost wherever it is applied.
• Egress waivers and portability terms requested explicitly and captured in writing in the agreement, per the egress brief.
• A right sized commitment, because a commitment set to optimistic growth converts a slowdown into a shortfall you pay in cash.
• Clean usage data, which lets you negotiate from facts rather than from the vendor's account of your consumption.

The standard consultancy answer to lock in is broad multi cloud. In roughly 25 of the 40 AWS estates we reviewed, that added more operational cost than the lock in it was meant to cure, because teams paid twice for skills, tooling, and undifferentiated work. The instinct is right that leverage requires an alternative. The execution is wrong about how much alternative is needed.

What the engagement file shows is a consistent gap between the feared cost of leaving and the measured one. Egress is the line everyone points to, and it is genuinely the largest exit cost, yet it rarely exceeded 3 to 5 percent of a single year of committed spend. That is a real number and not a trivial one, but it is nothing like the barrier it is treated as in renewal conversations. The asymmetry matters because the alternative cost, staying without leverage, showed up every year rather than once.

This is why the plan outperforms the move. Customers with a written, costed exit plan secured 8 to 15 percent better renewal terms than those without one, and most of them never migrated anything. A renewal conversation changes character when the buyer can put a costed, time bound path for a named workload on the table, because the discussion stops being about whether you could leave and becomes about what it would take. Nothing produces that shift except specificity: a general willingness to consider alternatives is not leverage, and both sides know it.

The commitment is the part buyers control most directly and manage worst. Sizing to growth rather than to floor demand added 10 to 20 percent of phantom cost, and it does something subtler as well: it converts the exit option into an expensive one, because leaving now means paying a shortfall on top of the egress. A commitment sized to demand you are certain of keeps the option affordable, which keeps the leverage real. Pick one low gravity workload, cost its move honestly, right size the commitment, and take the documented plan into the renewal. The commitment sizing sits in the renewal strategy and the year round controls in the vendor management playbook.

Across roughly 35 to 45 AWS commitment and exit reviews in 2024 and 2025, the cost of leaving was almost always lower than customers feared:

The patterns: exit costs estimated rather than calculated, multi cloud bought broadly to cure a narrow exposure, and commitments sized to a growth curve that made leaving expensive.

The buyer side move is to make one alternative real and documented. The wider library sits in the AWS practice.

What actually creates AWS lock in?

Three forces with different cures. Data gravity, where large datasets are slow and costly to move. Proprietary services with no clean equivalent elsewhere, which raise re engineering cost. And committed spend, where the discount always carries a matching obligation that penalizes leaving early.

How expensive is an AWS exit really?

Lower than most teams assume. Egress was the single largest exit line and rarely exceeded 3 to 5 percent of one year of committed spend in the estates we reviewed. The number that matters is your own, calculated from current published rates against your largest datasets, not an industry anecdote.

Is the exit plan worth building if you intend to stay?

Yes, and that is the main reason to build one. Customers with a written, costed exit plan secured 8 to 15 percent better renewal terms than those without one. The plan changes the renewal conversation whether or not you ever act on it, which makes it the cheapest leverage available.

Should you go multi cloud to escape lock in?

Rarely as a blanket strategy. In roughly 25 of the 40 estates we reviewed, broad multi cloud added more operational cost than the lock in it was meant to cure, because teams paid twice for skills and tooling. Keeping one workload genuinely portable delivers most of the leverage at a fraction of the cost.

Can egress waivers be negotiated?

Sometimes, and it is always worth asking explicitly. Waivers have been granted under specific conditions and regulatory pressure has widened the options. Whatever is agreed has to appear in the agreement in writing, because a waiver described in a conversation is not a waiver.

How does the commitment affect exit exposure?

It deepens it directly. Commitments sized to growth rather than to floor demand added 10 to 20 percent of phantom cost in the estates we reviewed, because a slowdown triggers a shortfall you pay in cash. Sizing to demand you are certain of keeps the exit option affordable.

Which workloads move most easily?

Stateless compute and containerized workloads, because they carry little data gravity. Databases and analytics platforms with large stored datasets are the hardest and most expensive to relocate, which is exactly why the portable workload you keep ready should be chosen for low gravity rather than for size.

What makes an exit plan credible rather than theoretical?

Costed figures and a timeline for a specific workload. Egress, rebuild effort and dates for one named workload with low data gravity, documented and shared internally. A general intention to consider alternatives changes nothing. A costed, time bound path for one real workload changes the conversation.