EC2 Reserved Instance Types Explained

AWS EC2 Reserved Instances come in several forms, each with different discount profiles, flexibility characteristics, and use cases. Understanding the distinctions is essential before committing capital to any RI purchase.

Standard Reserved Instances

Standard RIs provide the highest discount on EC2 — up to 72% off on-demand for all-upfront, three-year terms — but commit the buyer to a specific instance type (family + size), operating system, tenancy, and region. The commitment is precise and inflexible: the RI applies its discount only when running usage exactly matches the reservation parameters. A Standard RI for an m5.large running Linux in us-east-1 does not apply to an m5.xlarge or a c5.large, even if those instances are running in the same account.

Standard RIs can be listed for sale on the AWS RI Marketplace if circumstances change — this provides an exit option that other commitment vehicles do not, though the Marketplace may not always have buyers for every RI type, and sale prices typically reflect a discount from remaining RI value. This salability is a meaningful risk-management feature for organisations with genuine uncertainty about multi-year architectural stability.

Convertible Reserved Instances

Convertible RIs offer lower discounts than Standard RIs — typically 5–10 percentage points lower at equivalent payment and term combinations — but can be exchanged for different instance families, operating systems, or tenancies as long as the exchange is of equal or greater value. Convertible RIs cannot be sold on the RI Marketplace.

The exchange mechanism is the key operational consideration: exchanges require that the new reservation's on-demand value is equal to or greater than the original. This means you can exchange a Convertible RI for a newer, more expensive instance family, but you cannot exchange down to a cheaper family and pocket the difference. For organisations planning Graviton migration or instance family transitions during a commitment window, Convertible RIs provide RI-level commitment coverage without requiring the specific instance type to remain fixed.

Zonal vs Regional Reserved Instances

Standard and Convertible RIs can be either Zonal or Regional in scope. Zonal RIs apply to a specific Availability Zone and provide a capacity reservation guarantee — the specified instance type is guaranteed to be available in that AZ when needed. Regional RIs apply to a full region, with the discount automatically shared across all AZs in that region for matching usage, but without capacity reservation. For most enterprise workloads, Regional RIs are preferable due to the broader coverage and automatic AZ optimisation. Zonal RIs are appropriate specifically when capacity reservation is a hard requirement for the workload.

EC2 RI Pricing: Payment Options and Discount Rates

The cost of an EC2 Reserved Instance depends on three variables: the instance type and specifications, the contract term (1 or 3 years), and the payment option. The interaction of these variables determines both the discount rate and the cash flow profile.

Payment Options Compared

AWS offers three payment structures for all RI types. The discount progression from no-upfront to all-upfront reflects the time value of money — AWS discounts its revenue in exchange for early receipt of committed capital:

  • All Upfront (AURI): Full payment at purchase. Highest discount rate — typically 35–40% off on-demand for 1-year terms, 60–72% for 3-year terms. Break-even against on-demand pricing occurs typically within 7–9 months on 1-year terms, 18–24 months on 3-year terms.
  • Partial Upfront (PURI): Approximately 50% paid upfront, remainder billed monthly. Discount is 1–3 percentage points lower than all-upfront at equivalent term. Break-even is slightly later but initial capital deployment is halved. Practical for organisations with capital constraints but multi-year stability confidence.
  • No Upfront (NURI): Monthly billing only, no upfront payment. Lowest discount — typically 30–35% off on-demand for 1-year terms, 45–50% for 3-year terms. Zero capital risk, but meaningfully lower savings versus AURI. Appropriate for organisations prioritising cash flow preservation over maximum discount.

Break-Even Analysis

Break-even is the point at which the cumulative cost of the RI commitment equals what you would have paid on-demand. For a 1-year AURI, the break-even calculation is straightforward: the upfront payment divided by the monthly on-demand cost. If the monthly on-demand cost is $200 and the 1-year AURI costs $1,200, break-even occurs at month 6. After that point, every month the instance runs represents pure savings.

The critical variable in break-even analysis is utilisation rate. A 1-year RI purchased for a workload that runs 8 hours per day rather than 24 hours has an effective utilisation rate of 33%. At that utilisation, the RI provides 33% of its potential discount while incurring 100% of its cost — the ROI may be negative compared to on-demand for sporadic workloads. RIs should only be purchased for workloads with utilisation rates at or near 100%.

"The most common RI mistake is purchasing based on average usage rather than baseline usage. A workload that averages 70% utilisation will have some periods at 20% utilisation — and an RI for that workload may underperform expectations."

ROI Framework: How to Model Your EC2 RI Return

A complete EC2 RI ROI model requires five inputs: current on-demand hourly cost for the target instance type, expected utilisation rate, RI payment option and term, cost of capital for upfront payment (if applicable), and architectural stability confidence for the commitment window.

Step 1: Identify Your Baseline Workloads

Use 60–90 days of CloudWatch and Cost Explorer data to identify EC2 instances that consistently run at near-100% utilisation. An instance that shows 95%+ hourly utilisation across the analysis window is a strong RI candidate. Instances with variable utilisation profiles — weekday spikes, batch processing, seasonal patterns — are not ideal RI candidates without careful analysis.

Step 2: Match Instance Types to Workloads

For each baseline workload, confirm the instance type expected to remain stable for the full commitment window. For 1-year terms, this requires 12-month architectural stability confidence. For 3-year terms, that extends to 36 months — a much higher bar. If you are planning an OS migration, Graviton adoption, or right-sizing exercise within the commitment window, a Standard RI for the current configuration may have reduced utilisation mid-term. Convertible RIs or Compute Savings Plans are more appropriate when any architectural change is planned. The broader context on when to use RIs versus Savings Plans is covered in our Reserved Instances vs Savings Plans 2026 guide.

Step 3: Calculate the ROI by Payment Option

For each candidate workload, model the economics across all three payment options. For a hypothetical m5.xlarge in us-east-1 with an on-demand cost of $0.192/hour ($1,683/year at 100% utilisation), the approximate 1-year RI costs are as follows: AURI at 38% discount equals approximately $1,044 ($639 savings per year); PURI at 36% equals approximately $1,078 ($605 savings); NURI at 33% equals approximately $1,127 ($556 savings). The AURI saves an additional $83 per year compared to NURI — at scale, this matters, but the capital deployment ($1,044 upfront versus $0) must be weighed against your cost of capital.

Step 4: Scale to Portfolio Level

Individual RI ROI calculations are useful as checks, but enterprise decision-making operates at portfolio level. Aggregate your RI candidates by instance family and region to identify natural groupings. At portfolio level, Convertible RIs for an entire instance family may offer better risk-adjusted returns than Standard RIs for individual instance types, because the exchange flexibility protects against selective underutilisation when some instances change while others remain stable.

Want an independent EC2 RI portfolio assessment?

We identify RI opportunities, model break-even, and align RI strategy with your EDP commitment.
Speak to AWS Cost Specialists →

EC2 RI and EDP Interaction

EC2 Reserved Instances count toward your AWS EDP commitment, making RI purchases a dual-function instrument: they reduce compute costs while simultaneously accelerating EDP draw-down. This interaction has both strategic and tactical dimensions.

Strategically, building a robust RI portfolio is one of the most reliable ways to ensure consistent EDP commit draw-down on stable compute workloads. RI purchases are typically booked as costs at purchase time or spread as amortised monthly charges, both of which contribute to eligible EDP spend. Including projected RI portfolio purchases in your EDP commit modelling ensures your annual commitment is credible and appropriately sized.

Tactically, RI purchases are a rapid shortfall remediation mechanism. A large Q4 RI purchase can close an EDP shortfall gap quickly and convert potential shortfall charges into committed compute discounts that deliver genuine forward-period value. The mechanics of this tactic are covered in our AWS RI and Savings Plan optimisation guide and the EDP shortfall risk management guide.

AWS Cost Intelligence — Monthly

EC2 pricing updates, RI optimisation tactics, and EDP alignment strategies for enterprise buyers.

Managing Your RI Portfolio Operationally

EC2 RI portfolios require active management to maintain high utilisation rates and capture maximum ROI. Several operational practices are essential.

Utilisation Monitoring

AWS Cost Explorer provides RI utilisation reporting that shows the percentage of purchased RI capacity that was actually applied to running instances. Any utilisation consistently below 85% warrants investigation: either the workload has changed, the instance was right-sized, or the RI was purchased speculatively without sufficient analysis. Low utilisation is the primary value destroyer in RI portfolios — it should be reviewed monthly, not annually.

Coverage Reporting

RI coverage reports show what percentage of your total on-demand usage is covered by RIs (or Savings Plans). Low coverage means you have eligible workloads that are not benefiting from commitment discounts — this is an undercommitment problem rather than a waste problem. Coverage analysis identifies RI expansion opportunities, while utilisation analysis identifies waste. Both metrics should be part of your monthly FinOps review cycle.

Expiration Management

Expired RIs revert all usage to on-demand pricing without warning. AWS does not proactively alert you to expiring RIs in a way that ensures timely renewal — expiration management must be built into your procurement calendar. Set alerts at 90 and 30 days before expiration for all significant RI commitments. The renewal decision should align with any EDP renewal activity and be modelled alongside Savings Plan alternatives, as RI renewal is an appropriate moment to re-evaluate whether RIs or Savings Plans are the better vehicle for each workload.

Convertible RI Exchange Execution

For Convertible RIs, exchanges must be executed through the AWS console before the original RI expires. The exchange process creates a new RI with a new term — typically extending the commitment. Modelling exchange timing is important: an exchange executed 6 months into a 1-year Convertible RI resets the commitment term to the new RI's full duration. Organisations running active migration programmes should map their Convertible RI exchange schedule to their infrastructure migration timeline.

For organisations with complex RI portfolios, or those approaching an EDP renewal where RI strategy needs alignment with the new commitment level, our AWS cost optimisation advisory team provides independent portfolio reviews. For questions about data egress costs which affect the total AWS cost picture alongside EC2, see our AWS data transfer and egress negotiation guide. To discuss your EC2 RI portfolio and broader AWS cost strategy, contact us here.