Look, we’ve all been there. You open the AWS Cost Explorer, hoping for a steady, predictable line, but instead, you see a vertical spike that looks like a mistake. Your first instinct is to blame a sudden surge in traffic, but usually, the truth is much more boring – and much more expensive.
After digging through hundreds of environments, we’ve realized that most wasted spend isn’t about high-level architecture. It’s about housekeeping. It’s the silent leaks that build up over time.
Here are the 5 black holes actually draining your budget while you’re busy building features.
1. The EBS Graveyard
When you terminate an instance, AWS asks if you want to delete the root volume. But what about the other four volumes attached to it? Usually, they just sit there in “available” state, doing absolutely nothing but racking up charges every hour. It’s essentially a graveyard of dead data that no one is using.
- The fix: Stop playing detective. Filter for “Unattached” volumes in your console and kill them. Better yet, automate the cleanup so you don’t have to go on a “search and destroy” mission every month.
2. The Lazy Tax (Legacy Families)
Are you still running M4 or T2 instances in 2026? If you are, you’re paying a premium for old hardware. AWS wants you to move to Graviton (m7g) for a reason – it’s cheaper for them to run and faster for you. Staying on old families is basically paying a “lazy tax” for not updating your Launch Templates.
- The fix: Modernize. Switching to Graviton isn’t just a performance boost; it’s a direct 20-40% cut to your compute bill.
3. Snapshot Hoarding
There is a fine line between Data Retention and Hoarding. We’ve seen companies keeping daily snapshots from their MVP days four years ago. You’re paying S3 prices for data that is literally useless for any recovery scenario.
- The fix: Set a hard policy. If the business doesn’t need it for compliance, delete it. Use Lifecycle Manager (DLM) to automate the “trash” so your storage costs don’t snowball into a mountain of waste.
4. Idle Load Balancers (The Silent Base Fee)
An Application Load Balancer costs money just for existing. That’s fine for production, but what about those ten ALBs created for a POC that ended last quarter? They don’t care if they aren’t routing traffic – they’ll keep billing you as long as they are on.
- The fix: If the Request Count is zero, the ALB shouldn’t be there. Go to Trusted Advisor, find the idlers, and shut them down before the next billing cycle.
5. The AWS Toll Road (Data Transfer)
This is the one that blindsides everyone. You see a massive charge for “Data Transfer Out” and have no idea where it’s coming from. It’s usually inefficient cross-AZ traffic or moving logs where they don’t need to go. It’s like a toll road with no signage – you only realize how much you’ve spent when the bill hits your desk.
- The fix: Keep it in the family. Use VPC Endpoints to avoid NAT Gateway charges and keep as much data as possible within the same region or AZ to avoid unnecessary tolls.
FinOps isn’t a one-time project, it’s a hygiene habit.
If you haven’t looked at these five black holes in the last 30 days, you’re almost certainly overpaying for resources you don’t even use.
Is your environment optimized, or just running? Our team can spot your blind spots and help you fix them. Reach out for a FinOps review.
