A missed cloud spend commitment is billed at 100 percent of the unconsumed balance in most AWS, Microsoft, and Google contracts, which turns a 20 percent shortfall on a $5M commitment into a $1M true-up invoice with nothing delivered in return. The penalty is not a fee on top of usage; it is the gap between what you committed and what you spent, charged in full. Because commitments are almost always sized to an optimistic forecast, shortfall is the most common and most avoidable cloud overspend. This guide explains how the penalty works at each hyperscaler and the contract terms that cap it.
Cloud spend commitments trade a discount for a promise. You agree to spend a fixed dollar amount over one to five years, and in return you receive a discount, credits, or preferred pricing. The vehicles are the AWS Enterprise Discount Program and Private Pricing Agreements, the Microsoft Azure Consumption Commitment, and Google Cloud commitment-based discounts. The discount is real. The risk is that demand falls short of the forecast that justified the commitment. The wider negotiation context sits in the software contract negotiation guide.
Why this is the most avoidable cloud cost
Unlike compute waste, which requires continuous engineering effort to control, shortfall is avoided once, at the negotiating table, by sizing the commitment correctly and securing carryforward or true-forward terms. A buyer who gets the structure right at signing rarely pays a penalty regardless of how demand moves, which makes the few hours of modeling before signature among the highest-return work in cloud cost management.
How shortfall penalties are calculated
Across all three providers the core formula is the same: penalty equals committed amount minus eligible consumption, with the difference invoiced at term-end or trued up at each anniversary. The differences are in what counts as eligible consumption and whether any shortfall can carry forward.
| Provider | Commitment vehicle | Shortfall treatment |
|---|---|---|
| AWS | EDP / Private Pricing Agreement | Unmet commitment billed at term-end; limited carryforward if negotiated |
| Microsoft Azure | MACC | Drawdown tracked over term; unconsumed balance owed; broad eligible-service list |
| Google Cloud | Spend-based / committed use | Monthly or term shortfall billed; some SKUs excluded from drawdown |
The detail that costs buyers money is the eligible-service definition. Spend on marketplace purchases, certain support tiers, or third-party SKUs may not count toward drawdown even though it is real cloud spend, which means a buyer can spend the full committed amount and still register a shortfall. The Azure rules are unpacked in Azure MACC explained.
A worked example
Consider a three-year, $15M AWS commitment, $5M a year. The buyer forecasts steady growth but a business unit divests in year two, cutting demand 20 percent. Actual spend lands at $4M in year two. The $1M gap is billed in full at the anniversary. Over a three-year deal where two of the three years run 15 to 20 percent short, the cumulative shortfall true-up can exceed the entire discount the commitment was meant to deliver, leaving the buyer worse off than pay-as-you-go.
The asymmetry to watch: Overspending past a commitment usually earns only the same discount, while underspending is penalized at 100 percent of the gap. The risk is entirely one-directional. Size the commitment to your conservative demand case, not your forecast, and push the upside spend into a separate uncommitted tranche that still earns the discount.
The protections to negotiate
Four contract terms turn an unforgiving commitment into a survivable one. A carryforward right lets an underspend in one period offset a future period instead of being billed. A true-forward structure bills only growth above the commitment and never penalizes a shortfall. A ramped commitment sets lower targets in early years when demand is uncertain. An off-ramp or reduction right allows the commitment to be reduced on a defined trigger such as divestiture or a material business change. Buyers who secure two or more of these cut shortfall exposure by 50 to 70 percent. The mechanics of building these into a deal are in cloud FinOps negotiation and Azure committed use.
Sizing the commitment correctly
The discipline that prevents shortfall is sizing to demand you are confident in, then laddering additional commitments as usage proves out rather than committing the full forecast on day one. A laddered approach signs a smaller base commitment with the right to add tranches at the same discount, capturing the price benefit without betting the full forecast. This pairs with ongoing rightsizing so the committed baseline reflects optimized, not wasteful, consumption. See cloud cost optimization for the rightsizing methods and contract red flags for the language to strike.
The shortfall penalty also interacts with the discount in a way buyers rarely model: the deeper the discount, the larger the commitment that earned it, and the larger the commitment, the bigger the absolute dollar shortfall a given percentage miss produces. A 30 percent discount that required a $20M commitment turns a 15 percent demand miss into a $3M penalty, while a smaller commitment at a 22 percent discount would have produced a far smaller absolute exposure for the same percentage miss. The discount and the risk scale together, which is why the deepest discount is not automatically the best deal once the shortfall exposure is priced in.
Why commitments get over-sized
Commitments are over-sized because the discount scales with the commitment, which gives both the vendor's salesperson and the buyer's own champion an incentive to commit more. The salesperson is compensated on committed value, so a larger number is a better deal for them. The buyer's cloud team wants the deepest discount, which means the largest commitment, and the forecast that justifies it is built in optimistic conditions before any business disruption is priced in. The result is a commitment anchored to a growth case that assumes nothing goes wrong, signed by people who benefit from a bigger number, against demand that is inherently uncertain. The corrective is to size against a conservative demand case and treat the optimistic forecast as upside that earns the same discount through laddered tranches rather than a single large commitment.
A second driver is multi-year term pressure. Vendors offer the best rate on the longest term, and a three or five year commitment locks in a discount but also locks in a forecast across years where the business is least predictable. The further out the commitment runs, the wider the band of plausible demand, and the higher the shortfall risk. A shorter base commitment with renewal options often costs only a few points more in discount while removing most of the long-horizon shortfall exposure.
Reading the eligible-spend definition
The eligible-spend definition decides what counts toward drawdown, and it is where buyers register a shortfall despite spending the committed amount. Marketplace purchases, certain premium support tiers, third-party SKUs sold through the provider, and some professional services may be explicitly excluded from drawdown even though they are billed by the same provider. A buyer who assumes all provider spend counts can spend the full committed dollar amount across a mix of services and still owe a shortfall true-up because a portion of that spend was ineligible. The defense is to read the eligible-service list before signing, model your planned spend mix against it, and negotiate the broadest possible definition, ideally one that counts marketplace and support spend toward the commitment.
| Spend category | Usually eligible | Often excluded |
|---|---|---|
| First-party compute and storage | Yes | Rarely |
| Premium support tiers | Sometimes | Often |
| Marketplace third-party software | Sometimes | Often |
| Professional services | Rarely | Usually |
| Reserved capacity prepayments | Varies | Varies |
Tracking drawdown in-year
A shortfall is far cheaper to prevent than to remediate, which makes in-year drawdown tracking the single most valuable operational control on a commitment. The pattern that gets buyers in trouble is discovering the shortfall at the anniversary, when there is no time left to consume the gap. A monthly drawdown report that compares consumed eligible spend against the linear path to the commitment gives the cloud team months of warning, enough time to accelerate a planned migration, reallocate workloads into the committed account, or convert eligible purchases to close the gap. Organizations that track drawdown monthly almost never pay a shortfall penalty; those that check at renewal frequently do. The reporting discipline is part of the wider cloud FinOps approach.
Modeling the break-even discount
Every commitment has a break-even point: the consumption level below which the shortfall penalty erases the discount and the buyer would have been better off on pay-as-you-go. Modeling that point before signing tells you how much demand cushion the commitment carries. If a 25 percent discount requires 90 percent consumption to beat pay-as-you-go after penalties, the commitment only pays off if demand lands within 10 percent of forecast, which is a thin margin for a multi-year bet. If the same discount breaks even at 70 percent consumption, the commitment is far safer. Buyers should require this break-even analysis as a condition of signing, because it converts the commitment from a leap of faith into a quantified risk. The total-cost framing belongs in a software TCO model alongside the contract terms that matter.
Why multi-year terms compound the risk
A longer commitment term widens the band of plausible demand, and the wider that band, the higher the shortfall risk, which is the trade hidden inside the deeper discount a multi-year term earns. Forecasting cloud demand one year out is hard; forecasting it three or five years out, across acquisitions, divestitures, re-architectures, and business cycles, is closer to a guess. A five-year commitment locks the buyer to a number set when the future was least knowable, and the discount that justified it can be erased by a single year of unexpected contraction billed at full shortfall. The deeper discount on a longer term is real, but it is compensation for risk the buyer is taking on, not a free benefit.
The structural answer is to prefer a shorter base term with renewal options, or a laddered set of shorter commitments, accepting a slightly smaller discount in exchange for far less exposure. A buyer who can demonstrate steady consumption after two years is in a stronger position to commit further than one who guessed at five years on day one. Model the discount difference against the shortfall risk explicitly, using the break-even analysis, before choosing the longer term. The renewal framing is in cloud renewal strategy.
If you are already in a shortfall
A buyer heading for a shortfall has more options before term-end than after. Reallocating eligible workloads into the committed account, accelerating planned migrations, and converting marketplace purchases into drawdown-eligible spend can close a gap. Where the shortfall is structural, the renewal is the moment to renegotiate the next term down rather than letting the provider roll the same number forward. Providers prefer a renewed customer at a lower commitment to a penalized one who leaves. Our cloud contract negotiation and software licensing advisory teams run this remediation. For estates spanning AWS and Microsoft, the commitments should be modeled together so a shortfall on one is not hidden by overspend on the other.