Your FinOps roadmap is a modernization roadmap

 |  Milen Kohli

At June 2026’s AWS Summit Toronto, two conversations were happening in parallel.

In the breakout rooms, AWS was talking about modernization: moving away from outdated database versions, adopting newer instance generations, migrating x86 workloads to Graviton, introducing serverless architectures and upgrading Kubernetes clusters before Extended Support fees appeared.

A few metres away, we were showing customers Stable recommendations based on their own AWS environments. The lists looked remarkably similar:

  • Upgrade an Amazon RDS database that has reached Extended Support
  • Migrate an x86 workload to Arm
  • Move Amazon EBS volumes from gp2 to gp3
  • Tier old Amazon S3 data
  • Upgrade an Amazon EKS cluster before additional fees apply

AWS called it modernization. Others initially called it cloud cost savings. The customer was looking at the same work either way. That overlap reveals something important about cloud cost optimization: much of what companies describe as FinOps is really modernization translated into financial terms.

Your AWS bill measures how far the platform has evolved

An AWS environment can run perfectly well even as the services and architecture behind it become dated. The applications still work, the databases remain stable and there is rarely an immediate reason to interrupt product development for an infrastructure upgrade. Meanwhile, AWS continues to release newer instance generations, managed services and pricing models that can make the existing environment more expensive by comparison.

AWS keeps introducing newer and often more efficient ways to run the same workloads, whether through updated instance families, managed services, serverless options or lower-cost storage tiers. At the same time, older database and Kubernetes versions eventually fall out of standard support and become more expensive to keep. 

The architecture may be no worse than it was when the team built it, but its position relative to the AWS platform keeps changing. Over time, the cost difference between the architecture you have and the architecture you could have becomes visible. 

However, a code review will not flag an older instance family when the application still performs correctly. Monitoring will not identify a database version as a financial problem while the database remains healthy. Architecture reviews may catch it, but those reviews are periodic and often postponed.

The AWS bill keeps reflecting the gap every month. That is one reason FinOps has become such a useful operating function. It turns the financial effect of an aging architecture into a visible list of engineering decisions.

FinOps Lite and the work underneath it

Most cloud cost programs begin with two familiar categories.

The first is commercial optimization: Savings Plans, Reserved Instances, enterprise agreements and other discounts tied to predictable usage. These can create substantial savings without requiring architectural changes.

The second is cloud hygiene: deleting unattached Amazon EBS volumes, removing idle load balancers, shutting down unused instances, releasing abandoned Elastic IP addresses and cleaning up resources that no longer serve a purpose.

Both categories matter. Both should be part of a well-managed AWS environment.

Nevertheless, once commitments and unused resources have been addressed, the remaining opportunities tend to require engineering judgment.

Should a workload change to Graviton? Should a database use provisioned capacity or a serverless model? Could Redis be replaced with Valkey? Should infrequently accessed data shift to another Amazon S3 storage class? Is the team better off upgrading a database now or continuing to pay Extended Support fees?

These decisions cannot be made from a billing table alone. They require an understanding of the application, its traffic patterns, operational requirements, migration risk and expected lifespan.

This is where FinOps becomes modernization. Commercial discounts lower the price of the infrastructure a company already runs. Modernization changes the infrastructure itself.

Extended Support makes modernization debt explicit

Extended Support fees are one of the clearest examples.

AWS charges additional fees when customers continue running certain Amazon RDS, Amazon Aurora and Amazon EKS versions after standard support has ended. The resource may continue operating normally, but maintaining an outdated version now carries a visible premium.

In a FinOps dashboard, that premium may appear as an annualized savings opportunity. Operationally, it is closer to a receipt for deferred modernization.

The customer is paying more because an upgrade did not happen before the support deadline. Removing the charge requires engineering work: assessing compatibility, testing the new version, planning the migration and managing the change safely. In other words, the recommendation is financial. But the solution is technical.

Extended Support is simply the most direct version of a broader pattern. AWS has always encouraged customers to head toward newer services and architectures through performance, pricing and support policies. Older technologies gradually become less competitive, less efficient or more expensive to maintain.

Extended Support turns that gradual pressure into a line item. Calling it a savings opportunity is technically accurate, but incomplete. The company is not finding a clever discount. It is addressing modernization debt after that debt has begun appearing on the invoice.

The role FinOps actually plays

FinOps gives organizations a practical way to connect engineering decisions with financial outcomes. That connection is often what allows modernization work to compete for time and budget.

An engineering team may struggle to prioritize a migration because the existing environment still works. The business case becomes clearer when the same migration is connected to a recurring annual cost, an upcoming support fee or a measurable improvement in unit economics.

The financial lens helps answer questions that architectural arguments alone may not.

How much is the current design costing the company? What is the expected return from changing it? Which migration should happen first? How long will it take for the savings to recover the implementation cost? Is the engineering effort worthwhile for a workload that may be retired next year?

FinOps gives finance and engineering a way to evaluate the same problem from both sides. The cost is visible, the technical work is clear and leadership has enough context to decide whether the effort belongs on the roadmap.

A FinOps report should be a modernization roadmap

A useful FinOps report should do more than produce a long list of possible savings.

A useful FinOps report does more than group recommendations by savings. It helps teams separate quick wins from larger technical projects and weigh each opportunity against the effort, risk and expected payback.

A recommendation to migrate an instance to Graviton is not useful simply because Graviton may be less expensive. The team needs to know whether the workload is compatible, how much it currently costs, what the expected savings are, what must be tested and whether the migration deserves priority over other opportunities.

The same applies to database modernization, storage tiering, serverless adoption and Extended Support remediation.

Moving before the bill forces the decision

For many companies, FinOps becomes urgent only when the AWS bill forces the issue. A budget is missed, an unexpected charge appears or leadership starts asking questions. The modernization work then has to happen under pressure instead of as part of a planned roadmap.

A stronger approach is to treat cost efficiency as a continuous architectural practice.

That means identifying aging instance families before they become a material disadvantage. Planning database upgrades before Extended Support begins. Reviewing whether workloads are still using the right compute, storage and database models. Evaluating new AWS services based on their operational and financial fit rather than adopting them simply because they are new.

Not every recommendation is worth pursuing. Migration work takes time, carries risk and has to compete with other priorities. The better approach is to make those trade-offs before higher AWS costs force the issue.

When modernization happens on the company’s schedule, the financial savings still appear. They become evidence that the architecture is becoming more efficient rather than the emergency justification for finally updating it.

More than just easy cost savings

Savings Plans, Reserved Instances and unused-resource cleanup can reduce waste quickly. They do not answer the larger question of whether the environment is still built on the right AWS services, versions and operating models.

That is where the deeper FinOps work begins. The recommendations may appear on a cost report, but acting on them often means upgrading, migrating or redesigning part of the environment.

Used well, FinOps does not simply show where money is being spent. It helps companies decide where the architecture needs to move next.