Skip to content
- Are EC2 Savings Plans Bad? The Hidden Costs AWS Won’t Tell You About
Are EC2 Savings Plans Bad? The Hidden Costs AWS Won’t Tell You About
AWS is constantly evolving its pricing models, but that doesn’t mean every new option is an upgrade. When Savings Plans launched in 2019, they promised to simplify long-term discounting with broader flexibility than Reserved Instances. Between Compute Savings Plans and EC2 Instance Savings Plans, the idea was to cover more of your usage with less effort.
But the reality is that Savings Plans alone leave critical gaps—no way to adjust commitments when your usage shifts, higher risk of overcommitment with dollar-based planning, and no option to offload unused commitments.
For the vast majority of organizations, we argue that RIs — or a combination of SP and RI — are the better choice. In this post, we’ll break down why relying only on Savings Plans is almost always a mistake and how to build a better commitment strategy.
Amazon EC2 pricing
Firstly, let’s quickly review the basics. Amazon has 4 main pricing models:
On-Demand Instances | Spot Instances | Savings Plans (SPs) | Reserved Instances (RIs) |
AWS defaults to On-Demand pricing, where compute resources are billed per second or hour with no upfront commitment. While this provides flexibility, it also comes at a premium. | Spare AWS capacity offered at deep discounts (<90%). A cost effective approach to interruptable workloads. AWS can reclaim Spot with a 3 minute warning. Availability forecasting and automation are key. | Provide a flexible pricing model offering lower prices than On-Demand pricing in exchange for a specific usage commitment (measured in $/hour) for a one- or three-year period with savings up to 72%. | Provide the highest savings (up to 75%) for predictable workloads by committing to specific instance types and regions, in contrast to Savings Plans, which offer more flexibility for slightly lower discounts |
Each model comes with various tradeoffs. Leaving aside the Spot option for now, we’ll discuss how Savings Plans compare to Reserved Instances and why your team might want to avoid relying on them — particularly if you’re spending at least 100k annually on AWS.
AWS EC2 Savings Plans vs Reserved Instances
When choosing between commitments, consider the following 8 factors:
Flexibility
Compute Savings Plans (SPs) were designed to maximize coverage in dynamic environments. They apply to EC2, Lambda, and Fargate across all instance types, regions, and operating systems without requiring any manual adjustments. This makes them appealing for organizations with constantly evolving infrastructure.
But that flexibility comes at a price. SPs are immutable—once you lock in a dollar-per-hour commitment, you can’t adjust it if usage drops or shifts. As ProsperOps warns, this rigidity has led some customers to negative Effective Savings Rates when overcommitted.
Reserved Instances (RIs) seem more restrictive, especially Standard RIs which require accurate forecasting. However, Convertible RIs (CRIs) shift the balance. With CRIs, you can adapt your commitments to new instance families, sizes, regions, and operating systems as your workloads change—all while retaining your original expiration date.
Bottom line: In the short term SPs are more flexible. In the long term, CRIs can provide a huge amount of flexibility for teams managing longer-term usage patterns.
Ease of Use
Savings Plans are often described as “set it and forget it.” AWS automatically applies SP discounts to eligible usage without requiring you to model specific instances or families. Duckbill notes this simplicity is one of their biggest advantages: no manual exchanges, no resale, no tracking which workload is covered.
However, this simplicity can be deceptive. Because SP commitments are made in post-discount dollars, miscalculations are common and it’s easy to accidentally overcommit — which can be worse than just paying the On-Demand rate in the first place.
Without automation, RIs take more effort upfront. Standard RIs demand precise selections of instance type, region, and term. CRIs reduce operational burden but still require some management (or automation tools) to ensure optimal alignment with actual usage. But because you can “grow” and “squish” them, they are more forgiving of changes in usage patterns.
Bottom Line: SPs can be simple for teams with a small amount of AWS spend, but they aren’t as simple as they seem at first glance.
Discounts
At first glance, Reserved Instances might seem like a worse deal—you trade flexibility for only slightly better discounts. Standard RIs offer up to 75% off On-Demand pricing, while EC2 Instance SPs max out at 72%, and Compute SPs slightly lower at 66%.
But the discount percentages can be misleading. SPs offer nearly equivalent savings without the complexity of manual exchanges or resales. However, CRIs often match Standard RI discounts over time by avoiding underutilization through mid-term exchanges.
Bottom Line: SPs deliver broad discounts. CRIs, when managed properly, can achieve the same or better discounts by keeping commitments aligned with evolving usage.
Predictability
For steady-state workloads, RIs—particularly Standard RIs—offer the highest possible discounts and predictable monthly costs. This predictability is valuable for teams needing stable budgets over a 1- or 3-year term.
Savings Plans provide similar predictability if dollar-based forecasting is accurate. But, SPs don’t cover every compute element. They apply automatically in “deepest discount first” order, which can leave some workloads uncovered—often leading to uncomfortable chargeback conversations between teams.
Bottom line: Both SPs and RIs can deliver predictable costs, but only CRIs let you adjust commitments mid-term when forecasts miss the mark.
Exchangeability
Once purchased, Savings Plans are fixed. If your usage shifts significantly—like migrating workloads to another region, changing instance families, or scaling down—you’re stuck paying for unused commitments. There’s no equivalent to the AWS Reserved Instance Marketplace for SPs, meaning any overcommitment is pure sunk cost.
Reserved Instances provide mitigation options. Convertible RIs (CRIs) allow you to exchange mid-term for new instance families, sizes, regions, or operating systems without resetting your expiration date. Standard RIs don’t have this flexibility, but they do offer an escape hatch: the AWS Reserved Instance Marketplace.
This marketplace lets you list unused Standard RIs for sale to other AWS customers, allowing you to recover some of your sunk costs. There are caveats—it only works for RIs purchased at full price (not discounted EDP or PPA agreements), you can’t sell RIs within the first 30 days or the final month of their term, and market liquidity varies by region and instance type. But for commonly used configurations, it’s a practical safety net.
Bottom Line: SPs are one-way doors, RIs come with escape hatches.
Coverage
Savings Plans cover EC2, Lambda, and Fargate. For compute-heavy environments, this may be enough. But for organizations leveraging other AWS services like RDS, ElastiCache, Redshift, DynamoDB, and OpenSearch, SPs leave significant gaps.
Reserved Instances extend coverage to all these services. This broader applicability is critical for mixed-service architectures where compute isn’t the only driver of costs.
Bottom Line: SPs are compute-focused. RIs bring organization-wide savings across multiple AWS services.
The Verdict
Savings Plans may seem like the safer, easier option, but they’re also the riskiest when your usage shifts. With no way to modify or resell commitments, overcommitting can lead to wasted spend that’s impossible to recover. Reserved Instances—especially Convertible RIs—offer crucial safety nets. For most organizations, Savings Plans are absolutely not enough.
Automatically Manage SPs and CRIs
Automating the management and exchange of your AWS commitments can significantly enhance their value. AWS Marketplace offers tools like nOps that can automate tracking and exchanges, ensuring businesses are always optimizing costs.
Here are the benefits of letting nOps manage your CRIs:
- More Cost Savings: Intelligently analyzes usage patterns and continually selects optimal exchanges based on workload changes
- No Overhead: integrating with AWS APIs to handle the full commitment management process, nOps frees your team to focus on building and innovating
- 100% Visibility and Utilization Guarantee: See your actual savings at any time. nOps backs you with a 100% utilization guarantee* eliminating your risk.
With this closed‑loop system, customers typically see the following 90-day results:
| Metric | Before nOps | After nOps |
| Average Coverage Accuracy | ±18 % | ±1 % |
| Commitment Utilization | 78 % | 96-98 % |
| Effective Discount vs On‑Demand | 34 % | 52 % |
nOps was recently ranked #1 with five stars in G2’s cloud cost management category, and we optimize $2 billion in cloud spend for our customers.
Join our customers using nOps to maximize your commitment savings and leverage automation with complete confidence by booking a demo with one of our AWS experts!
*Subject to terms and conditions
