AlloyDB Express is a configuration within AlloyDB for PostgreSQL that delivers rapid provisioning alongside scale-to-zero compute capabilities. Designed specifically for workflows involving AI coding agents and rapid prototyping, it enables you to spin up and query a PostgreSQL database in seconds.

It features pay-as-you-go pricing for compute and storage, along with a free tier every month. When an instance becomes idle, compute automatically scales down to zero so you are not charged for compute during inactive periods.

Compute:

• Compute is measured in “Compute units” (CUs). 1 Compute Unit One CU has a ratio of 1vCPU for every 4 GB of memory.
• Compute is billed per second. You aren't billed for compute while the instance is paused.

Storage:

• Storage is billed for the amount of data storage you use maximum up to 20 GB. Used Storage is billed even when instances is paused.

Free Tier:

• Free tier each month includes 50 compute units and 2 GB of storage per billing account for the first 12 months. 
• Usage beyond the free tier is billed on 1 CU or 2 CU instances respectively.

Discounts: 

• Standard enterprise discounts apply. CUDs are not available.

All other prices are the same as standard AlloyDB prices.

Billing starts from December 2026.

Compute resources for AlloyDB for PostgreSQL are organized as instances, and are priced per vCPU and GiB of memory. When you create an instance, you choose the number of CPUs and the amount of memory you want, up to 288 vCPUs and 2232 GiB of memory per node. A zonal primary instance uses one node. A high availability primary instance uses two nodes, one for the active primary and a second for the in-region standby. AlloyDB read pool instances can contain up to 20 nodes. CPU and memory pricing varies by location. Select your region in the dropdown on the pricing table.

AlloyDB supports multiple machine series. You can use different machine series across instances in your cluster, and switch between machine series with a simple, non-disruptive update operation. For more information about machine series and their regional availability, see Choose an AlloyDB machine type.

AlloyDB also offers committed use discounts (CUDs) for CPU and memory resources when you commit to continuously use databases for a one- or three-year term. The pricing table below illustrates effective pricing per vCPU and per GiB of RAM after CUDs are applied. For more information about these commitments, see Committed use discounts.

AlloyDB for PostgreSQL uses an intelligent regional storage system that automatically scales up and down. As a result, you only pay for the storage you use. Your database storage is shared between all of the instances deployed on your cluster. As a result, your storage costs don’t change when you horizontally scale your workload by adding read pool instances, or more read nodes into your existing read pools. Storage prices vary by the location of your database cluster. Select your region in the dropdown on the pricing table.

AlloyDB for PostgreSQL stores the data associated with your on-demand, automatic, and continuous backups in backup storage. When you take an on-demand backup, AlloyDB stores a full snapshot of your database at that time. When you configure your database cluster with automated and continuous backups, AlloyDB takes a full snapshot, and incremental backups periodically. A backup is billed from the time it is complete (designated by a "Ready" state) until it is outside of its retention period. AlloyDB’s continuous backup and recovery feature also stores transaction logs to enable recovery to any point in time within its recovery window. First seven days of transaction logs are retained at no extra charge, and logs retained after seven days are billed.

Data transfer into AlloyDB for PostgreSQL is free. Data transfer out of AlloyDB within the same region is also free. Pricing for data transfer out of AlloyDB resources depends on the destination of the traffic, and in some cases, whether a specialty networking service is involved.

Outbound traffic into another region

Outbound traffic prices below are based on source-destination pairs, where the source is an AlloyDB resource, and the destination is another AlloyDB resource or a Google Cloud service in another region.

Outbound traffic into the internet

Outbound traffic prices below are based on source-destination pairs, where the source is an AlloyDB instance, and the destination is any client that is not a Google product, such as using a local server to read data from AlloyDB.

Other networking charges

If you have chosen to use certain networking features, you may incur additional charges:

• For public IP, an idle static IP address charge is applied only when the AlloyDB instance is stopped. For more information, see External IP address pricing
• For Private Service Connect, an hourly charge per endpoint and data processing charge for both inbound and outbound traffic is applied. For more information, see Using a Private Service Connect endpoint (forwarding rule) to access a published service.

Extended support pricing applies to AlloyDB for PostgreSQL clusters with instances that are running major versions in extended support. Refer to the Database version policies documentation for the extended support timelines of AlloyDB major versions.

Extended support is priced per vCPU per hour and charged for every second that an instance is running. Extended support is charged in addition to regular instance pricing. Pricing depends on the region where your instance is located. Select your region in the dropdown on the pricing table.

Read pool instances are charged at the same rate as primary instances.

Primary instances configured for high availability, also called regional instances, will be billed for extended support for both the primary instance and the standby instance. Learn more about high availability.

Note: Committed use discounts (CUD) don't apply to extended support prices.

The table below outlines pricing by region: 

AlloyDB free trial clusters

AlloyDB free trial clusters let you evaluate how AlloyDB performs in real-world scenarios and try AlloyDB features and capabilities at no cost for 30 days. During the 30-day free trial period, you can store up to 1 TB of data in your free trial cluster, and deploy an 8 vCPU primary instance, along with an 8 vCPU, single node read pool instance. For more details, see AlloyDB free trial clusters. If you are a new Google Cloud customer, you are also eligible for the 90-day Google Cloud free trial that gives you $300 credits to use on Google Cloud products including AlloyDB. The AlloyDB free trial cluster is in addition to the $300 credits offered by the Google Cloud free trial.

You can upgrade from a free trial cluster to a paid cluster to continue using your resources beyond the 30-day free trial period and remove any free trial restrictions. If you choose to upgrade your free trial cluster to a paid cluster, you will be charged according to the pricing on this page.

AlloyDB AI

AlloyDB AI is available at no additional charge. Customers making calls to AI models will be billed through the model provider. Pricing for Gemini models is available here.