How do sustained-use discounts work?

GCE automatically applies sustained-use discounts (SUDs) when instances run more than 25% of the month. The discount increases incrementally: the first 25% is full price, the next 25% is 20% off, then 40% off, then 60% off. A full-month instance receives approximately 30% average discount. SUDs apply to N1, N2, and N2D but not E2 or T2D.

What is the difference between preemptible and spot VMs?

Spot VMs are the successor to preemptible VMs with the same 60-91% discount but without the mandatory 24-hour maximum runtime. Both can be terminated with 30 seconds notice when GCE needs the capacity. Spot VMs also support custom pricing caps. Use either for fault-tolerant batch, CI/CD, and distributed computing workloads.

How are custom machine types priced?

Custom machine types are priced per vCPU and per GB of memory at slightly higher rates than equivalent predefined types. However, they can be more cost-effective when predefined types would waste resources. Extended memory (beyond 8 GB/vCPU) carries a premium surcharge.

GCP GCE Cost Estimator

ComputeGCP

Estimate Compute Engine costs with sustained-use and committed-use discounts.

Last verified: May 2026

GCE Instance Configuration

Machine Type

Number of Instances

Hours Per Month

730 hours = average month. Max 744 hours (31 days).

Operating System

Sustained Use Discount(up to 30% off for N1/N2/C2 families)

Commitment Type

Preemptible / Spot VM(~60-80% off, may be terminated)

Raw Output

Output will appear here...

See It in Action

Your team is comparing GCP costs against your existing AWS deployment of 100 m5.xlarge instances ($14,400/month on-demand). The estimator shows the GCP equivalent (n2-standard-4) at $108/month list × 100 = $10,800, dropping to $7,560 with SUDs (full-month operation). With 1-year CUDs, that drops further to $5,940/month. Multi-year savings projection: $100K+/year vs your current AWS spend. The CFO greenlights a migration POC.

What This Tool Does

The GCP GCE Cost Estimator calculates monthly Compute Engine costs with sustained-use and committed-use discounts factored in. It covers predefined and custom machine types, boot and additional persistent disks, network egress, and preemptible/spot instance pricing. The tool shows the effective hourly and monthly rates after applicable discounts.

Technical Details

The estimator computes monthly Compute Engine cost as: machine type rate × hours × (1 - SUD discount for eligible families) for on-demand, or × (1 - CUD discount) for committed-use, plus boot disk and additional disk costs at per-GB rates. For Spot pricing, it applies the current spot discount (typically 60-80%) to the equivalent on-demand rate.

Common Use Cases

1Estimating monthly Compute Engine costs for a fleet of VMs with specific machine types and disk configurations.
2Comparing on-demand vs. spot vs. committed-use pricing for a production workload.
3Calculating the savings from committed-use discounts on a projected VM fleet over 1 or 3 years.
4Modeling the cost of custom machine types vs. predefined types for workloads with unusual CPU/memory ratios.

Common Questions

How do sustained-use discounts work?: GCE automatically applies sustained-use discounts (SUDs) when instances run more than 25% of the month. The discount increases incrementally: the first 25% is full price, the next 25% is 20% off, then 40% off, then 60% off. A full-month instance receives approximately 30% average discount. SUDs apply to N1, N2, and N2D but not E2 or T2D.
What is the difference between preemptible and spot VMs?: Spot VMs are the successor to preemptible VMs with the same 60-91% discount but without the mandatory 24-hour maximum runtime. Both can be terminated with 30 seconds notice when GCE needs the capacity. Spot VMs also support custom pricing caps. Use either for fault-tolerant batch, CI/CD, and distributed computing workloads.
How are custom machine types priced?: Custom machine types are priced per vCPU and per GB of memory at slightly higher rates than equivalent predefined types. However, they can be more cost-effective when predefined types would waste resources. Extended memory (beyond 8 GB/vCPU) carries a premium surcharge.

Expert Tips

TIP

Spot VMs in GCP (replacing preemptible) have NO maximum runtime limit anymore — historically preemptible was capped at 24 hours, which made them unusable for long-running workloads. Spot VMs are evicted only when GCE needs the capacity back, which can be days or weeks for unpopular machine types.

TIP

GCP's automatic SUDs are the most underrated cost feature across all clouds. A standard N2 instance running all month gets ~30% off automatically — no commitment, no calculator needed. When comparing GCP to AWS/Azure for always-on workloads, always quote the SUDs-adjusted rate.

TIP

Sole-tenant nodes are billed per node-hour rather than per VM-hour, with the per-VM allocations being free within the node's capacity. For workloads with strict licensing (BYOL Windows Server, Oracle), sole-tenant can be dramatically cheaper than running a fleet of small VMs each with its own license cost.

Featured In

Spot and Preemptible Instances: Saving 60-90% Without Getting Burned

Related Learning Guides

Compute Engine Machine Types22 min read
Cost Optimization Guide24 min read

Was this tool helpful?

Disclaimer: This tool runs entirely in your browser. No data is sent to our servers. Always verify outputs before using them in production. AWS, Azure, and GCP are trademarks of their respective owners.