AWS Budgets vs Cost Explorer (2026): What Each Actually Does
AWS Budgets and Cost Explorer get compared as if you have to pick one. You don’t. They do different jobs: Budgets watches the future and alerts you before spend crosses a line, Cost Explorer explains the past so you can see where the money went. Most accounts need both, and both are effectively free at normal usage.
The real question is not which one to use. It is what each one cannot do, because the expensive surprises live in that gap. This guide covers the practical differences, the exact pricing, a setup that takes 20 minutes, and what neither tool will ever tell you.
Quick Answer: Which One Do You Need?
| You want to… | Use |
|---|---|
| Get alerted before monthly spend exceeds a limit | AWS Budgets |
| Understand why last month’s bill was higher | Cost Explorer |
| Automatically stop resources when spend crosses a line | AWS Budgets (budget actions) |
| Break down spend by service, account, tag, or region | Cost Explorer |
| Track Savings Plans or RI utilization against a target | AWS Budgets (utilization budgets) |
| Forecast next month’s bill | Either (both forecast) |
| Find the specific idle resource wasting money | Neither (more on this below) |
What AWS Budgets Actually Does
AWS Budgets is a guardrail. You define a limit, and AWS tells you when actual or forecasted spend approaches it. The forecasted part matters: a budget can fire mid-month when you are merely on track to overshoot, not after the damage is done.
You can budget four things:
- Cost budgets: the common case. “Alert me at 80% of $5,000/month.”
- Usage budgets: limits on usage quantities, like EC2 running hours.
- RI and Savings Plans utilization budgets: alert when your commitment utilization drops below a threshold, which means you are paying for discount capacity you are not using.
- RI and Savings Plans coverage budgets: alert when too much of your usage runs on-demand instead of under a commitment.
Alerts go to email or SNS, which means they can reach Slack, PagerDuty, or anything else with an endpoint.
Budget actions are the part most teams skip: a budget can do more than notify. When a threshold is crossed, it can attach a restrictive IAM policy or SCP, or stop EC2 and RDS instances directly. For dev and staging accounts, “stop the instances at 120% of budget” is a real safety net.
Pricing, verified June 2026: monitoring-only budgets are free. Your first two action-enabled budgets are free each month; each additional action-enabled budget costs $0.10 per day. Scheduled budget reports cost $0.01 per delivered report.
What Cost Explorer Actually Does
Cost Explorer is the autopsy tool. It holds 13 months of cost and usage history, lets you group and filter by service, linked account, region, instance type, or tag, and forecasts up to 12 months forward. When the bill jumps, this is where you find out which service, which account, and which day.
It also produces recommendations: rightsizing suggestions for underutilized EC2 instances and Savings Plans or RI purchase recommendations based on your actual usage. They are worth reviewing, but they are suggestions in a console page, not alerts. Nobody pushes them to you.
Pricing, verified June 2026: the console is free. The API costs $0.01 per request against your primary billing view, which matters if you build anything that polls it. Hourly granularity is a paid add-on at $0.00000033 per usage record per day, which works out to about $0.01 per 1,000 usage records monthly.
If you want the bill itself decoded line by line, we wrote a separate walkthrough: how to read your AWS bill.
The Differences That Matter
| Dimension | AWS Budgets | Cost Explorer |
|---|---|---|
| Direction | Forward: thresholds on future spend | Backward: analysis of past spend |
| Trigger model | Pushes alerts to you | Waits for you to look |
| Automation | Can act (IAM, SCP, stop EC2/RDS) | Read-only |
| Granularity | Budget per scope you define | Slice by service, account, tag, region, day |
| Forecasting | Alerts on forecasted overage | Visual forecast up to 12 months |
| History | Current and prior periods per budget | 13 months |
| Typical cost | $0 for monitoring | $0 in the console |
The one-line version: Budgets is an alarm, Cost Explorer is a map. An alarm with no map tells you something is wrong but not what. A map with no alarm means you find problems only when you happen to look.
The 20-Minute Setup That Covers Most Teams
- One cost budget per account at your expected monthly spend, alerting at 80% actual and 100% forecasted. Two thresholds, one budget, free.
- A stop-instances action budget on dev and staging accounts. Non-production environments are where forgotten experiments run up bills. Your first two action-enabled budgets are free, so this costs nothing.
- A utilization budget if you own Savings Plans or RIs. Below 90% utilization, you are paying for discounts you are not collecting.
- A monthly 15-minute Cost Explorer review. Group by service, compare to last month, click into whatever moved. That is the whole ritual.
If you also want anomaly alerts between budget thresholds, AWS Cost Anomaly Detection is free and watches for unusual spend patterns. It detects spend changes, though, not structural waste, and that distinction is the next section.
What Neither Tool Does
Here is the gap. Budgets tells you THAT spending crossed a line. Cost Explorer tells you WHERE the money went. Neither tells you WHICH resource is wasted or what to run to fix it.
A concrete example from a real scan: an EBS volume sat unattached for 1,790 days, burning money the entire time as part of $284/month of waste found in a single region. No budget ever fired, because the spend was steady and predictable. Cost Explorer showed an “EC2-Other” line item every single month, and it looked normal because it was always there. Steady waste is invisible to both tools by design: one watches for change, the other reports totals.
The same applies to idle RDS instances, gp2 volumes that should be gp3, NAT Gateway traffic that could route through free VPC endpoints, and unattached Elastic IPs. None of these are anomalies. They are structure, and they recur in almost every account we scan. For a full worked example, see the anatomy of one organization leaking $104,724 a year with neither tool firing once.
You can find all of them by hand: we keep a list of 13 CLI commands that find this waste for free. Or you can have it done for you: CostPatrol runs 123 rules against your account through a read-only role and returns each finding with the dollar amount and the exact fix command. Rules like EBS-O003 (unattached volumes with their accumulated cost), RDS-O001 (databases with zero connections for 14+ days), and NAT-O001 (traffic routable through free VPC endpoints) cover exactly the territory Budgets and Cost Explorer leave open. Free for accounts under $5K/month.
Summary
| Factor | AWS Budgets | Cost Explorer |
|---|---|---|
| Job | Guardrail for future spend | Explanation of past spend |
| Best feature | Forecast-based alerts and actions | 13 months of sliceable history |
| Blind spot | Steady waste under the threshold | Anything you don’t look for |
| Cost | Free for monitoring | Free in the console |
| Our recommendation | Set up in every account, today | Review monthly, 15 minutes |
Use both. They are free, they take 20 minutes to set up together, and they cover spend changes and spend visibility. Then close the third gap, structural waste, with either the manual CLI route or an automated scanner, because that is the category where the quiet money disappears.