GCP Cost Calculator

About the GCP Cost Calculator

Google Cloud Platform (GCP) is the cloud of choice for data analytics, machine learning, and organizations already using Google Workspace. Its pricing model differs from AWS and Azure in several key ways — including per-second billing, sustained use discounts (automatic), and committed use discounts for predictable workloads. This estimator gives you a ballpark monthly cost before you dive into the official GCP Pricing Calculator.

Compute Engine Pricing

GCP charges separately for vCPUs and RAM, allowing custom machine types. General-purpose (N2) pricing (us-central1): ~$0.031/vCPU-hour and ~$0.0042/GB-hour. GCP automatically applies sustained use discounts — if you use a VM for 50%+ of a month, you get up to 30% off automatically, no commitment required.

Committed Use Discounts (CUDs)

Committing to 1 year: up to 37% off on-demand pricing. Committing to 3 years: up to 55% off. CUDs apply to vCPU and memory resources (not GPUs or local SSD). Unlike AWS Reserved Instances, GCP CUDs are flexible — they can be applied across any VM of the committed resource type in a region.

Preemptible and Spot VMs

GCP's Spot VMs (formerly Preemptible) offer up to 91% discount over on-demand. They can be stopped by Google at any time with a 30-second warning. Ideal for batch processing, rendering, and fault-tolerant applications.

Storage Pricing

Persistent Disk SSD: ~$0.17/GB/month | Standard HDD: ~$0.04/GB/month | Cloud Storage (nearline): ~$0.01/GB/month | BigQuery storage: ~$0.02/GB/month | Cloud SQL (SSD): varies by DB engine

Egress Costs

Inbound to GCP: Free. Egress to internet: First 1 GB/month free, then ~$0.08–0.23/GB depending on destination. Egress to other GCP regions: $0.01/GB (same continent) to $0.08/GB (inter-continental). Using Cloud CDN significantly reduces egress costs for static content delivery.

GCP-Unique Services Worth Budgeting

BigQuery: $5/TB queried (on-demand), $0.02/GB stored | Cloud Run: Pay only during request processing | Vertex AI: Per-node-hour for training and prediction | Google Kubernetes Engine: Free cluster management, pay only for nodes