The Cloud Waste No One Is Looking At

How one command finds the money hiding in your AWS bill

Cloud waste is the electricity bill no one reads.

The total gets attention, but the individual line items often do not. An EC2 instance was sized for a launch that ended months ago. A Fargate service still has capacity for a traffic peak that never returned. A database is technically running but barely queried. Snapshots and object versions accumulate because nobody owns the cleanup decision.

None of these choices looks unreasonable when it is made. Together, they can become a quiet, recurring cost.

Industry surveys consistently estimate that 27–32% of cloud spend is wasted. That does not mean every company can remove a third of its bill tomorrow. It means many teams could benefit from a systematic way to find resources worth reviewing.

Waste hides behind ordinary line items

AWS gives teams extraordinary flexibility. It also distributes spending across thousands of resources, usage types, and decisions. Billing dashboards can show where money went without always explaining which specific resource no longer delivers enough value.

The difficult questions live below the account total:

• Which compute instances are idle?
• Which Fargate services are consistently over-provisioned?
• Which Lambda functions are no longer invoked?
• Which databases are quiet enough to review?
• Which S3 buckets lack lifecycle policies?
• Which snapshots and machine images have outlived their purpose?
• Which workloads might be candidates for more efficient architectures?

Answering those questions manually takes time, service knowledge, and context. That is why the review is so easy to postpone.

A useful second opinion in one command

GreenOps Scan is a free, open-source CLI that inspects AWS resource metadata and usage signals for potential cost and carbon opportunities.

Run:

npx greenops-scan

Choose an AWS profile and region, then let the scanner check services including EC2, S3, ECS, EKS, Lambda, RDS, ElastiCache, CloudFront, and EBS snapshots.

The result is a terminal summary plus JSON and a branded PDF report. Each finding identifies the affected resource, explains the issue, estimates the potential cost and carbon impact, and provides remediation guidance for review.

The estimates are a starting point, not a promise. Workload requirements, contractual pricing, architecture, and business context determine whether a finding can become a realized saving.

What one real scan surfaced

On one real AWS account, GreenOps surfaced:

• 130 issues across nine services
• approximately $2,209 per month in estimated savings opportunities
• approximately $26,510 per year in estimated savings opportunities
• approximately 1,210 kg CO2 per year in modeled carbon-reduction opportunities

The largest groups included over-provisioned ECS Fargate services, S3 buckets without effective lifecycle policies, idle or older RDS resources, and Lambda functions with low usage or efficiency opportunities.

Those figures describe one account at one point in time. They do not guarantee that another account will produce the same result, or that every surfaced estimate should be acted on. The value of the scan is that it turns an abstract concern into a concrete review list.

Even when only a small number of findings survive technical and business review, they might be enough to justify the few minutes required to run the command.

Safe by design: read-only and local

The first question for any infrastructure tool should be: what can it access, and where does the data go?

GreenOps Scan is designed for a low-risk first look:

• it uses read-only AWS APIs
• it runs locally on your machine
• your AWS credentials never leave your machine
• it does not install an agent in the account
• it cannot modify, stop, resize, or delete resources
• the report stays local unless you choose to share it

The scanner reports; it does not remediate automatically. Every recommendation should be reviewed by someone who understands the workload, its owners, and its recovery requirements before a change is made.

That separation matters. A team can begin with a dev or sandbox account, inspect exactly what the scanner finds, and decide whether the evidence is useful without granting write access or onboarding a SaaS platform.

Cost efficiency and carbon efficiency are connected

Unused and over-provisioned infrastructure is not only a financial inefficiency. Compute, storage, and data movement require hardware and energy even when they deliver little useful work.

GreenOps models carbon opportunities using principles from the Green Software Foundation methodology. The estimates help translate technical findings into a second dimension: the potential emissions associated with avoidable resource use.

These figures are modeled estimates, not audited emissions measurements. They do not make an organization CSRD compliant and should not be presented as guaranteed carbon reductions.

They could still be useful. CSRD and other sustainability programs increase the need for engineering, finance, and sustainability teams to understand where digital infrastructure consumes resources and how improvement opportunities are identified. A documented, repeatable scan can contribute evidence to that internal conversation.

From a one-time scan to continuous monitoring

A one-time scan answers an immediate question: what might be worth reviewing today?

Cloud environments do not stand still. Teams deploy new services, traffic changes, temporary resources become permanent, and yesterday's reasonable allocation can become tomorrow's waste. The same account can look different a month later.

That is where GreenOps Cloud fits. It extends the free CLI approach with continuous monitoring, scan history, shared visibility, and recurring detection for teams that need more than a point-in-time report.

Start with the lowest-friction step:

npx greenops-scan

Review the findings with the people who own the workloads. If the scan surfaces useful opportunities and the team wants ongoing visibility, explore GreenOps Cloud for continuous monitoring.

You might find a meaningful saving. You might find only a few cleanup tasks. Either way, you will have replaced a vague suspicion with evidence.