A Google Cloud flexible CUD is a spend-based commitment: you agree to spend a minimum amount per hour on eligible compute for one or three years, and in return you take about a 28 percent discount on a one-year term or 46 percent on three years. One flexible commitment covers Compute Engine, GKE, and Cloud Run together, regardless of machine family or region. As of January 2026 the discount applies directly to eligible usage rather than as a separate credit. The rule is to commit only to your steady hourly floor, after rightsizing, so the discount never strands on waste.
Last updated: June 2026. Written by Fredrik Filipsson and reviewed by Morten Andersen, built on our See, Cut, Lock, Run method.
This article sits in our complete guide to Google Cloud cost optimization, the cluster pillar it links up to, and is part of the broader cloud cost optimization playbook. Buying commitments is the Lock step in our method, and it comes last on a clean baseline. Where the commitment lands depends on how you run compute, so pair this with our GKE Autopilot vs Standard cost comparison before you size a commitment.
Commit last, and commit to the floor. Rightsize and schedule compute first, then buy a flexible CUD only for the steady hourly spend that runs every hour of every day. Committing before rightsizing, or committing to peak rather than baseline, locks in waste for one to three years. The flexibility of the instrument is what lets you commit confidently to that floor without fearing a machine-family or region change will strand it.
What is a Google Cloud flexible CUD?
A Compute flexible committed use discount is a spend-based commitment to a minimum hourly dollar amount of eligible Google Cloud compute, in exchange for a discount on that spend. Unlike a resource-based CUD, which locks you to a specific machine family and region, a flexible CUD follows your spend: a single commitment applies across Compute Engine, GKE, and Cloud Run, across machine families and regions, per the Compute Engine committed use discounts documentation. You are not promising to keep a particular VM running; you are promising to keep spending a baseline dollar amount per hour, which is a far easier promise to keep as your estate evolves.
How much do flexible CUDs discount?
Flexible CUDs give roughly a 28 percent discount for a one-year commitment and about 46 percent for a three-year commitment on the committed spend. Those rates are lower than the deepest resource-based CUDs, which can reach higher discounts on a specific instance family, but the flexible model trades a little discount depth for the freedom to move spend without stranding the commitment. The table below summarizes the trade.
| Commitment type | Approx. discount | Flexibility | Best for |
|---|---|---|---|
| Flexible CUD, 1 year | About 28% | Any machine family, region, eligible service | Dynamic estates, steady baseline |
| Flexible CUD, 3 year | About 46% | Any machine family, region, eligible service | Very stable baseline you trust for 3 years |
| Resource-based CUD | Deeper on a fixed family | Locked to instance family and region | Predictable, unchanging workloads |
| Sustained use discount | Automatic, smaller | No commitment | Always-on Compute Engine, no planning |
The rates above reflect the Google Cloud model as we read it in June 2026. Confirm current discount percentages for your region in the linked documentation before committing, because they change.
What changed in the 2026 flexible CUD model?
As of January 2026, Google migrated spend-based CUDs to a multiprice, direct-discount model. Instead of appearing as a separate billing credit applied after the fact, the discount is now applied directly to the eligible usage on the invoice, per the Cloud Billing flexible commitment analysis documentation. The practical effect for a FinOps team is cleaner attribution: the discounted rate shows up against the project and service that generated the spend, so you can allocate the savings to the right team instead of reconciling a pooled credit. The economics of the commitment are unchanged; what improved is how readable the savings are.
How do you adopt flexible CUDs?
Adopt flexible CUDs after rightsizing, by committing only to your steady hourly floor. The sequence below is the Cut-then-Lock order from our method.
- Rightsize first. Rightsize and schedule compute before committing, so the baseline you cover is clean, steady spend rather than current waste.
- Find your steady hourly spend. Use Cloud Billing CUD analysis and recommendations to find the floor of eligible Compute Engine, GKE, and Cloud Run spend that runs every hour of every day.
- Commit to that floor. Buy a Compute flexible CUD for the steady hourly dollar amount on a one-year or three-year term, not for peak usage.
- Let the discount apply automatically. The flexible commitment applies as a direct discount across eligible compute, regardless of machine family, region, or service, with no manual attachment.
- Layer resource CUDs and Spot on top. Add resource-based CUDs on very stable instance families and Spot on fault-tolerant work to capture deeper savings above the flexible baseline.
Want your flexible CUD sized to the floor, not the peak?
Our Google Cloud cost audit rightsizes your compute first, then models the exact one-year and three-year flexible commitment that maximizes savings without stranding spend. On the performance model, you pay only from realized savings. No savings, no fee.
Book a Google Cloud cost audit →Flexible CUDs or resource-based CUDs?
Choose flexible CUDs for dynamic estates and resource-based CUDs only for a very stable core. Because a flexible commitment follows your spend across services and machine families, it carries far less risk of stranding when you change instance types, migrate regions, or shift workloads between Compute Engine, GKE, and Cloud Run. Resource-based CUDs discount more deeply but lock you to a specific family and region, so a single architecture change can leave a commitment paying for capacity you no longer use. The common 2026 posture is a flexible baseline covering the bulk of steady spend, with resource-based CUDs added only on the handful of workloads you are certain will not change for the full term. For the broader commitment strategy across clouds, see our Google Cloud guide.
The Google Cloud Cost Optimization Field Guide includes the flexible-versus-resource commitment model and the baseline-sizing worksheet referenced above. It is the downloadable companion to this explainer.
Frequently asked questions
What is a Google Cloud flexible CUD?
A Compute flexible committed use discount is a spend-based commitment where you agree to spend a minimum dollar amount per hour on eligible Google Cloud compute for a one-year or three-year term, in exchange for a discount on that spend. One commitment covers eligible Compute Engine, GKE, and Cloud Run usage regardless of machine type, region, or service.
How much do flexible CUDs save?
Compute flexible CUDs give roughly a 28 percent discount for a one-year term and about 46 percent for a three-year term on the committed spend. Resource-based CUDs on specific instance families can discount more deeply but are far less flexible. Confirm current rates in the Google Cloud documentation before committing.
What changed in the 2026 flexible CUD model?
As of January 2026, Google migrated spend-based CUDs to a multiprice, direct-discount model. The discount is applied directly to the eligible usage rather than as a separate billing credit, which makes the savings easier to read on the invoice and to attribute to the teams that generate the spend.
Flexible CUDs or resource-based CUDs, which is better?
Flexible CUDs are better for dynamic estates because one commitment follows your spend across services and machine families, reducing the risk of stranded commitment. Resource-based CUDs give a deeper discount but lock you to a specific instance family and region, so they suit only very stable, predictable workloads. Most estates buy a flexible baseline and add resource CUDs only on the most stable core.
The short version
A Google Cloud flexible CUD commits you to an hourly compute spend for about 28 percent off on one year or 46 percent on three, applied directly to eligible Compute Engine, GKE, and Cloud Run usage under the 2026 multiprice model. Commit last, to your rightsized floor, and the discount never strands. When you want the commitment sized and timed for you, that is what our Google Cloud cost optimization service delivers.
Cloud pricing and service behavior change frequently. Verify the specifics in this guide against the providers’ own current documentation and the FinOps Foundation: Google Cloud pricing ↗, Google Cloud documentation ↗ and FinOps Foundation Framework ↗. This article also reflects Cloud Cost Room’s hands-on, vendor-neutral engagement experience.