# Databricks: Always-On Pricing
_Databricks introduces Always-On pricing for Lakehouse Postgres, offering a 25% discount on baseline capacity without long-term commitments._
**Published:** 2026-05-28
**Source:** https://www.startuphub.ai/ai-news/technology/2026/databricks-always-on-pricing
---
Databricks is aiming to simplify cloud database costs with its new Always-On pricing for Databricks Lakebase. The company is removing the traditional tradeoff between serverless flexibility and provisioned cost predictability.

Cloud DB Cost DilemmaDriver

serverless flexibility vs. provisioned cost predictability tradeoffFrom the article 2 mentionsDatabricks is aiming to simplify cloud database costs with its new Always-On pricing for Databricks Lakebase.Disable Scale-to-ZeroDriverFrom the article 2 mentionsUsers can activate Always-On pricing by disabling the scale-to-zero feature and defining an autoscaling range for their Lakebase Postgres instance.Define Autoscaling RangeDriverFrom the articleUsers can activate Always-On pricing by disabling the scale-to-zero feature and defining an autoscaling range for their Lakebase Postgres instance.24hr Continuous UseDriverFrom the articleOnce the minimum capacity is established as the baseline, it automatically qualifies for the lower Always-On rate after 24 hours of continuous use.Databricks Lakehouse PostgresCoremanaged Postgres database for Databricks LakehouseFrom the articleThis move follows other enhancements to Databricks Lakehouse, such as native Postgres synchronization, as detailed in Databricks Lakehouse Gets Postgres Boost on Azure.introducesAlways-On PricingContextnew pricing model for Databricks Lakehouse PostgresFrom the article 7 mentionsThe new Always-On pricing offers a 25% reduction on baseline compute capacity for its managed Postgres database.25% Baseline DiscountEffectoffers a 25% reduction on baseline compute capacityFrom the article 5 mentionsDatabricks is also offering an additional 50% promotional discount on top of Always-On pricing until January 31, 2027.Eliminate Manual OptimizationOutcomeFrom the articleThis aims to eliminate the need for manual optimization or lengthy commitments to achieve cost savings.No Long-Term CommitmentsOutcomewithout long-term commitments for cost savingsFrom the articleThis aims to eliminate the need for manual optimization or lengthy commitments to achieve cost savings.
The new [Always-On pricing](https://www.databricks.com/blog/introducing-always-pricing-automatic-savings-databricks-lakebase) offers a 25% reduction on baseline compute capacity for its managed Postgres database. This aims to eliminate the need for manual optimization or lengthy commitments to achieve cost savings.

## How Always-On Pricing Works

Users can activate Always-On pricing by disabling the scale-to-zero feature and defining an autoscaling range for their Lakebase Postgres instance. Once the minimum capacity is established as the baseline, it automatically qualifies for the lower Always-On rate after 24 hours of continuous use.

Any usage spikes above this baseline will continue to be billed at standard autoscaling rates, ensuring elasticity is maintained. This approach is designed for established production workloads with a consistent baseline activity.

For new or intermittent workloads, scale-to-zero remains the default to avoid overpaying for unused capacity. This feature is particularly beneficial for [cloud database cost optimization](/ai-news/technology/2026/databricks-lakebase-tackles-marketing-costs).

## Eliminating the Serverless vs. Provisioned Dilemma

Historically, managed PostgreSQL services have forced users into a difficult choice: pay a premium for serverless elasticity or commit to provisioned instances for a lower unit price, often leading to over-provisioning. Databricks claims its approach bridges this gap.

This move follows other enhancements to Databricks Lakehouse, such as native Postgres synchronization, as detailed in [Databricks Lakehouse Gets Postgres Boost on Azure](/ai-news/technology/2026/databricks-syncs-postgres-to-lakehouse-natively). The company emphasizes commitment-free pricing and a unified product experience.

High Availability replicas and larger instances, which do not scale to zero by design, automatically benefit from the lower Always-On rate without any configuration changes.

Databricks is also offering an additional 50% promotional discount on top of Always-On pricing until January 31, 2027.
---
Original analysis from [startuphub.ai](https://www.startuphub.ai), the #1 AI startup directory.