The median enterprise enters a cloud commitment with a forecast 18–34% above what they actually consume in year one. True-down rights are the safety net — and the mechanic exists at all three hyperscalers in some form. The customer has to raise it, has to raise it before signing, and has to model it alongside the commitment sizing.
A cloud commitment — AWS EDP, Azure MACC, GCP CUD — is a take-or-pay contract for forward consumption. The structural risk is identical across all three: the commitment is sized at the high end of the customer's forecast, signed for one to five years, and very difficult to true down in the default contract. The forecast miss is not hypothetical — across our engagement portfolio, the median enterprise enters the commitment with a forecast 18–34% above what they actually consume in year one. The customers who do not negotiate true-down rights are signing for a forecast miss that is almost statistically guaranteed.
The conversation is almost never offered by the hyperscaler. Account teams are compensated on commitment value and have no incentive to introduce flex mechanics that reduce the contract dollar amount. The customer therefore has to raise it, and has to raise it before signing — true-down rights are essentially impossible to negotiate mid-term. The mechanic exists at all three hyperscalers in some form. AWS EDP true-down is rare but precedented at large commit tiers. Azure MACC has a soft-flex re-base mechanism. GCP has the strongest published true-down language. None of them appear in the default contract template.
AWS EDP true-down concessions are granted at very large commit tiers, typically $25M+ annual commitment value. The mechanic is usually a defined re-base window mid-term — a single opportunity, often at the 24-month mark on a 3-year deal, to renegotiate the remaining commitment based on actual consumption pattern. The discount tier is preserved on the re-based commitment. The mechanic is not free — AWS will typically extract a longer renewal commitment, a deeper Marketplace contribution requirement, or a public reference commitment in exchange. The economics still favour the customer in any scenario where the original commitment was sized higher than actual consumption. Securing that re-base window is exactly the kind of structural concession a buyer-side AWS EDP negotiation is built to extract before signature, not after.
True-down rights are almost impossible to negotiate mid-term. They must be raised before signing — and almost no customer does.
Microsoft Azure Consumption Commitment (MACC) has a softer flex mechanism than AWS EDP. The standard concession at large commit tiers is an annual re-base review where the customer can adjust the year-two and year-three commitments based on actual year-one consumption. The mechanic is not a true-down to actual — there is usually a floor at 70–80% of the original commitment — but it provides meaningful relief for the customer whose forecast misses by 20%+. The concession is granted at commit tiers from roughly $10M annually, lower than the AWS threshold.
The negotiation pattern for MACC re-base is to position it as a precondition for any multi-year commitment, not a separate ask. Microsoft account teams have authority to grant the re-base provision at deal level without escalation in many cases, and the inclusion in the original contract has minimal financial impact on the deal economics. The mechanism is meaningful, common, and missed by the majority of MACC customers because they do not ask.
Google Cloud has the cleanest published true-down language of the three hyperscalers. CUDs at the spend-based level have explicit flex mechanisms in the standard contract. The practical execution is the weakest of the three — Google account teams have less commercial authority than AWS or Azure equivalents, and the actual deployment of the flex mechanism often requires deal escalation. The contract is in your favour. The operationalisation is not. The mitigation is to negotiate the operational process for invoking the flex in the original deal, not just the underlying right.
True-down patterns across AWS, Azure and GCP, commitment sizing model, re-base mechanics and the negotiation timeline.
True-down rights are the safety net for the inevitable forecast miss. The primary defence is sizing the commitment correctly to begin with. The recipe is bottom-up build of stable workload commitment from actual run-rate, plus explicit ramp curves for new workloads with conservative assumptions, plus explicit decay curves for sunset workloads, minus FinOps-driven savings expected over the term, plus an explicit 8–15% buffer. The commitment number that emerges is typically 18–28% lower than the number the account team will propose, and it is almost always closer to actual consumption.
The customers who size correctly and also negotiate true-down rights occupy a fundamentally different commercial position than the majority. They optimise for the most likely outcome, hedge for the downside, and walk into the renewal with a credible track record of forecast accuracy. The customers who over-commit and have no flex mechanism arrive at renewal with shortfall liability and zero negotiating leverage. The structural asymmetry of cloud commitments compounds every renewal cycle for the customer who does not actively manage it.
Almost never mid-term. The flex provisions must be negotiated into the original contract. Customers who do not raise true-down rights before signing have effectively no recourse for forecast miss until renewal.
Rare but precedented at commit tiers above $25M annually. Typically a single re-base window at month 24 of a 3-year commit, often in exchange for renewal commitment or Marketplace contribution requirements.
Yes, in the form of an annual re-base mechanism granted at commit tiers from roughly $10M. Floor is typically 70–80% of original. Microsoft account teams can grant the provision at deal level without escalation.
On paper, yes — GCP has the cleanest published flex language. In practice, the operational execution is the weakest of the three hyperscalers. Negotiate the operational process for invoking the flex, not just the right.
8–15% between the central forecast and the commitment level. Large enough that normal forecasting error does not produce shortfall, small enough that the commitment is not financing the hyperscaler's revenue plan.
We have advised on cloud contracts from $2M to $120M annually across AWS, Azure and GCP. The leverage is in the structure, not the rate card.
Weekly compliance intelligence for IT leaders.