Committed use discounts (CUDs) trade a one or three year commitment for a lower rate on Google Cloud, and they are where the largest savings on a steady workload live. The catch is that Google offers two structures that behave differently: resource-based CUDs tie you to specific machine resources in a region, and spend-based CUDs tie you to a dollar amount per hour. Choosing the wrong one, or buying before the footprint is clean, locks in cost you should have cut first.
This article links up to our complete guide to Google Cloud cost optimization, the pillar for this cluster. Commitments sit in the Cut step of our See, Cut, Lock, Run method, and the rule is the same one we apply everywhere: a commitment comes last, on a baseline you have already rightsized. The same coverage discipline shows up in our cross-cloud comparison of Reserved Instances, Savings Plans and CUDs.
Rightsize and clear idle workloads first, then commit on the smaller, stable baseline. A CUD bought on an oversized fleet locks the waste in for the full term.
What is a committed use discount?
A committed use discount is an agreement to use a minimum amount of Google Cloud over a one or three year term in exchange for a lower price. Unlike AWS Reserved Instances, GCP CUDs do not require you to reserve a specific instance up front in the old sense; they apply automatically to matching usage. The discount is substantial: resource-based commitments can reach up to 55 percent off on-demand for most machine types, and up to 70 percent for some, while the newer flexible spend-based commitments offer a flat 28 percent for one year and 46 percent for three years across many machine families and regions. These figures reflect Google Cloud pricing as of May 2026; verify current rates in the billing console before committing.
Resource-based CUDs: commit to specific resources
A resource-based CUD commits you to a minimum quantity of a specific resource, such as vCPUs and memory of a given machine family, in a particular region. In return you get the deepest discount, up to 55 to 70 percent. The trade-off is rigidity: the commitment is tied to that family and region, so if your workload migrates to a different machine type or region, the commitment can strand. Resource-based CUDs suit predictable, steady-state workloads that you are confident will keep running on the same shape of hardware for the full term, such as a stable database tier or a long-lived application fleet.
Spend-based CUDs: commit to an hourly dollar amount
A spend-based CUD commits you to a minimum amount of spend per hour on a product rather than to specific hardware. You keep receiving the discounted rate until your hourly spend on eligible resources reaches your committed amount, and anything above that bills at on-demand. Spend-based CUDs are more flexible because the discount follows your spend across eligible resources rather than a single machine type. The flexible variant, Compute Flexible CUDs, extends this further: a flat 28 percent for one year and 46 percent for three years that applies across multiple VM families and regions, so the commitment does not strand when workloads move. As of January 2026 Google migrated spend-based CUDs to a direct discount model rather than the older credit-based approach.
Not sure which commitment to buy?
Our Google Cloud cost audit reads your usage, rightsizes the fleet first, then models the resource-based and spend-based mix that maximizes coverage without stranding. On the performance model, you pay only from realized savings. No savings, no fee.
Book a GCP cost audit →Resource-based vs spend-based: how to choose
The decision comes down to predictability versus flexibility. If a workload is stable in shape and location, the deeper resource-based discount wins on raw savings. If your usage is steady in dollars but moves across machine families or regions, a flexible spend-based CUD captures most of the saving with far less stranding risk. Many estates use both: resource-based commitments for the predictable core, and a flexible spend-based layer on top for the part of the bill that drifts.
| Resource-based CUD | Spend-based / Flexible CUD | |
|---|---|---|
| You commit to | Resources (vCPU, memory) in a region | Hourly spend on a product |
| Typical discount | Up to 55%, up to 70% on some types | 28% (1yr) / 46% (3yr) flexible |
| Flexibility | Tied to family and region | Spans families and regions |
| Best for | Predictable, fixed-shape workloads | Steady spend that moves around |
| Main risk | Stranding if workload changes | Lower discount on the same spend |
Note that Google also offers sustained use discounts, which apply automatically without any commitment, and which we cover in sustained use discounts on Google Cloud. CUDs stack on top of the analysis of your actual run rate, so always model both before buying.
The Google Cloud Cost Optimization Field Guide includes the CUD coverage model we use to size resource-based and spend-based commitments side by side. It is the downloadable companion to this guide.
Common questions about committed use discounts
Do committed use discounts require an upfront payment?
No. Google Cloud CUDs are billed monthly across the term rather than paid all at once, so they do not tie up cash the way a fully prepaid reservation can. You commit to the usage or spend; the charge spreads over the one or three years.
Can I cancel a CUD if my workload changes?
Generally no. Commitments run the full term you signed up for, which is exactly why you rightsize and prove the baseline before committing. Flexible spend-based CUDs reduce the risk because the discount follows your spend across families and regions rather than stranding on one machine type.
Do CUDs stack with sustained use discounts?
Resource-based CUDs and sustained use discounts apply to different parts of your usage and are not simply added together; the platform applies the discounting that yields the lower price. Model your actual run rate with both in view before you buy, rather than assuming they compound.
The short version
Committed use discounts cut a Google Cloud bill the most, but only after rightsizing. Use resource-based CUDs for the predictable core where the deeper discount pays off, and flexible spend-based CUDs for the spend that moves across families and regions. Model coverage so you commit the steady baseline, not the peak. When you want the whole commitment mix sized and managed without guesswork, that is what our Google Cloud cost optimization service delivers.