Tech News
← Back to articles

$50 PlanetScale Metal Is GA for Postgres

read original related products more articles

PlanetScale Postgres is the fastest way to run Postgres in the cloud. Plans start at just $5 per month.

$50 PlanetScale Metal is GA for Postgres

By Richard Crowley | December 15, 2025

Today we’re making PlanetScale Metal for Postgres available in smaller sizes and at much lower price points, all the way down to the new M-10 for as little as $50 per month. We’ve lowered the floor from 16GiB of RAM with four sizes all the way to 1GiB and paired these with eight storage capacities ranging from 10GB to 1.2TB.

These new sizes are powered by the same blazingly fast, locally attached NVMe drives that customers like Cash App, Cursor, and Intercom use to decrease latency, increase reliability, and decrease costs, too.

This release is the first step towards decoupling CPU and RAM from storage capacity, while maintaining all the benefits of PlanetScale Metal. Each of these new CPU and RAM sizes can choose from at least five storage capacities, all of which still use locally attached NVMe drives. Customers can spec their PlanetScale Metal database to perfectly match their workload while still enjoying the lowest possible latency, the fewest possible failure modes, and online resizing.

Decoupling CPU and RAM from storage capacity means you can get as much as 300GB of storage per GiB of RAM, almost four times the highest density AWS offers natively. Or you can max out CPU and RAM on minimal storage to serve small, high-traffic workloads. The choice is finally yours.

Since we launched PlanetScale Metal, customers have asked loudly for two things:

A lower starting price, which we’re reducing today from $589 per month to $50 per month. Flexibility to buy more storage without also buying more CPU and RAM.

Today we’re proud to deliver on both requests. PlanetScale Metal is now available for Postgres Databases in AWS regions on both Intel and ARM CPUs with more I/O capacity than you can possibly use. Support for GCP is in the works and Vitess will follow soon.

... continue reading