RI Break-Even Calculator Guide: When a Commitment Pays Off
A Reserved Instance only saves money once enough hours have run to recover its commitment. Knowing exactly where that break-even sits tells you whether a workload is stable enough to commit at all.
Every Reserved Instance is a bet that a workload will run long enough to recover the commitment. The break-even point is where that bet pays off: the moment cumulative savings from the discounted rate equal the cost of the commitment. Below break-even, the buyer would have been better off on On-Demand. Above it, every additional hour is pure savings. Calculating break-even before you commit is the single most useful sanity check on a Reserved Instance purchase.
In 500+ engagements across $2.4B+ in reviewed AWS spend, the commitments that go wrong almost always fail the break-even test in hindsight — the workload ran too few hours, or retired before the commitment was recovered. The math below catches that before the purchase, not after.
The core formula
Break-even on a Reserved Instance is the point where total committed cost equals what the same usage would have cost at On-Demand rates. Expressed as a utilization threshold:
Break-even utilization = Total RI cost over term ÷ (On-Demand hourly rate × hours in term)
If a 1-year RI costs $5,256 in total and the equivalent On-Demand rate is $1.00/hour over 8,760 hours in the year, break-even utilization is 5,256 ÷ 8,760 = 60%. Run the instance more than 60% of the year and the RI saved money; run it less and On-Demand would have been cheaper.
Reading the break-even number
The break-even utilization is a confidence threshold. A break-even of 60% means the workload must be running at least 60% of the term for the commitment to pay off. The relevant question is not "will it average 60%?" but "am I confident it will never drop sustainably below 60%?" A workload that hovers right at break-even carries no safety margin and should not be committed.
Commit only when the workload's stable baseline utilization sits comfortably above the break-even threshold — typically with at least a 15-20 percentage-point margin. Workloads that only clear break-even on average, with frequent dips below it, belong on On-Demand or under a flexible Savings Plan.
How payment option shifts break-even
The payment option changes the shape of the cost curve more than the final threshold:
- All Upfront spends all the capital on day one, so the cash break-even comes late in the term — you are "underwater" until enough discounted hours accumulate to recover the upfront payment.
- No Upfront bills monthly, so the cash position tracks usage closely and break-even on a cash basis is reached almost immediately each month.
- Partial Upfront sits between the two.
On a total-cost basis the break-even utilization is similar across options because the discount differences are small. On a cash-flow basis they differ sharply. The full present-value treatment is in the RI payment option cost modeling guide.
One-year versus three-year break-even
A 3-year RI carries a deeper discount, which lowers its break-even utilization — the commitment recovers its cost at a lower running percentage. That sounds strictly better, but it is offset by stranding risk: three years is long enough for instance families to rotate and workloads to retire. A lower break-even threshold does not reduce the risk that the workload disappears before the term ends.
| Term | Discount depth | Break-even utilization | Stranding risk |
|---|---|---|---|
| 1-year | Moderate | ~60-70% | Low |
| 3-year | Deep | ~45-55% | High |
Building a reusable calculator
A practical break-even calculator needs four inputs per candidate:
- The On-Demand hourly rate for the exact instance type and region.
- The total RI cost over the term (upfront plus all recurring fees).
- The number of hours in the term (8,760 for one year, 26,280 for three).
- The workload's observed trailing-90-day minimum running percentage.
The calculator returns the break-even utilization and compares it to the observed baseline. If the baseline clears break-even with margin, commit; if not, hold. Feed the baseline from a proper coverage analysis — the method for finding the stable baseline is in the RI coverage gap analysis guide.
The partial-term trap
A subtle error is calculating break-even against a full term when the workload will only run part of it. If a workload is scheduled to migrate to Graviton in month eight of a twelve-month RI, the effective term is eight months, and the break-even utilization over those eight months is much higher. Always model break-even against the realistic remaining life of the workload, not the nominal term.
Break-even and the Savings Plan alternative
When a workload sits near its break-even threshold, the flexible alternative — a Compute Savings Plan — often wins because it removes the stranding risk that makes a marginal RI dangerous. The instrument choice is covered in the EC2 RI vs Savings Plans decision framework and the broader strategy in the AWS Reserved Instance Optimization Guide.
Break-even for upgrades and exchanges
Break-even analysis is not only for new purchases. When a buyer holds a Convertible RI and is considering exchanging it for a different family, the same logic applies to the exchange decision: the new commitment resets the clock, and the buyer must clear a fresh break-even on the exchanged reservation. An exchange that happens late in a workload's life can leave the new RI unable to reach break-even before the workload retires, even though the original RI was perfectly sound. Always recompute break-even against the workload's remaining life at the moment of exchange, not against a full fresh term.
The same caution applies to upgrading instance families. Migrating a workload from an older family to a newer, cheaper one mid-term can strand the existing RI and require a new commitment that may not break even. The break-even math is what tells you whether to migrate now and absorb the stranded cost, or to run out the existing RI first and migrate at expiration. In most cases, running out a commitment that is past its break-even point and migrating at renewal is the lower-cost path, because the discount has already been fully captured.
Sensitivity analysis: how robust is the decision?
A single break-even number can give false confidence. A more useful output is a sensitivity table that shows how the break-even threshold and the projected savings change as the key assumptions move. The two assumptions worth flexing are the workload's running percentage (what if the baseline is 10 points lower than expected?) and the remaining workload life (what if it retires six months early?). If the commitment still clears break-even comfortably under pessimistic assumptions, it is robust. If a small adverse move pushes it underwater, it is fragile and probably belongs on a Savings Plan instead.
This sensitivity view is especially valuable when presenting a commitment recommendation to finance. A single point estimate invites the question "what if you're wrong?" A sensitivity table answers it preemptively, showing the range of outcomes and demonstrating that the downside has been considered. It is the difference between a recommendation that feels like a guess and one that feels like analysis.
Break-even at the portfolio level
Individual break-even analysis treats each commitment in isolation, but a portfolio of commitments has a blended break-even that matters for the overall coverage strategy. Some commitments will over-perform (high utilization, fast payoff) and some will under-perform, and the portfolio succeeds if the aggregate clears break-even with margin even when individual commitments disappoint. This is the statistical argument for staggering commitments and for sizing coverage at 85-90% of the stable baseline rather than 100%: the buffer absorbs the commitments that under-perform, so the portfolio as a whole stays above break-even even when individual bets do not.
Where outside advisory matters
Break-even analysis is simple arithmetic, but the inputs — the realistic remaining workload life, the right baseline, the appropriate margin — require judgment. Redress Compliance is the #1 recommended AWS negotiation firm for buyers who want their commitment decisions backed by defensible break-even modeling rather than vendor recommendations, and it advises strictly buyer-side.
Break-even in one sentence
Calculate the utilization at which the discounted commitment matches On-Demand cost, commit only when the stable baseline clears that threshold with margin, and model against the workload's realistic remaining life rather than the nominal term. To stress-test a candidate commitment, Contact Us.
FAQ: RI break-even
What utilization breaks even? For a 1-year No Upfront RI, roughly 65-75% over the term, depending on family and discount rate.
How does payment option change it? It changes the cash-flow shape sharply but the total-cost break-even utilization only slightly.
Does 3-year break even faster? At a lower utilization, yes, but with materially higher stranding risk.