Estimate Cloud Spanner costs for regional and multi-region instances with processing units, storage, and capacity recommendations.
Last verified: May 2026
Single region. Min 100 PU. $0.90/node/hr ($657/node/mo).
= 1 node (1 node = 1000 PU). Increments of 100 up to 1000 PU, then 1000.
$0.30/GB/month
$0.30/GB/month
Output will appear here...The Cloud Spanner Cost Estimator models monthly costs for Google Cloud Spanner instances based on configuration type (regional or multi-region), processing unit count, and storage volume. Spanner charges by node-hour (or processing unit-hour) and GB-month of storage, with higher rates for multi-region configurations that provide global strong consistency. The calculator helps you understand Spanner's unique pricing structure where one node equals 1,000 processing units, and includes recommendations for minimum instance sizes based on your throughput and storage requirements. It also shows how Committed Use Discounts of up to 20% can reduce costs for steady-state workloads.
A processing unit is the smallest compute increment in Spanner. One node equals 1,000 processing units. You can provision in increments of 100 processing units (minimum) for small workloads, or in full nodes for larger ones. Each processing unit provides a baseline of reads, writes, and storage capacity.
Multi-region configurations replicate data across three or more regions with synchronous writes, providing 99.999% availability and global strong consistency. This requires significantly more infrastructure than a regional deployment, resulting in approximately 3x the cost per node-hour compared to regional configurations.
Google offers Committed Use Discounts (CUDs) of up to 20% for 1-year commitments and higher for 3-year commitments on Spanner compute. Unlike on-demand pricing, you commit to a minimum capacity for the term. The estimator calculates savings so you can evaluate the commitment tradeoff.
Your fintech team needs a globally consistent ledger across US and EU regions. Initial estimate: 3-node multi-region Spanner = ~$10,000/month. Alternative architectures considered: Cloud SQL with cross-region replicas (lacks strong consistency, would need application-level reconciliation = engineering cost), CockroachDB on GKE (operational burden, similar compute cost). The $10K/month for Spanner saves an estimated 4 engineer-months per year of distributed systems work. After 1 year of running, you commit to a 3-year CUD and reduce cost to ~$7,000/month.
The estimator computes Spanner cost as: processing units × per-PU-hour rate (regional vs multi-region rates differ ~3x) × hours, plus storage GB × per-GB-month rate. Committed Use Discounts reduce the PU rate by up to 20% (1-year) or higher (3-year). The recommendation engine maps your throughput requirements to minimum PU count using Google's documented baselines (10K reads/sec per node, 2K writes/sec per node).
Spanner has a hard rule: you need 1,000 processing units (1 node) per 10 TB of storage. As you grow data, you can't keep the same compute — Spanner forces an upgrade. Budget for compute scaling along with storage growth, not just storage.
Multi-region Spanner's 3x cost premium is real but it's the only Google database that delivers global strong consistency at sub-100ms latency. If your app legitimately needs that (financial ledgers, multi-region inventory, distributed locks), Spanner is dramatically cheaper than building this on top of Cloud SQL or Firestore. If you don't need it, the premium is wasted.
Spanner's per-PU minimum (100 PU = ~$50/month + storage) makes it expensive for small workloads. Below ~$1,000/month equivalent traffic, Cloud SQL or Firestore is almost always more cost-effective. Spanner's value is at the high end of consistency and scale, not as a default OLTP database.
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.