Introducing Always-On pricing: automatic savings for Databricks Lakebase

2026-05-27 · Source: Databricks · Field: Technology & Digital — Cloud Computing & IT Infrastructure, Data Science & Analytics, Artificial Intelligence & Machine Learning · Depth: Intermediate, short

Summary

Databricks has introduced Always-On pricing for its Lakebase operational database, offering serverless flexibility with a predictable, lower price for baseline usage. This new model allows users to set compute scaling parameters and automatically receive a 25% lower price on their baseline capacity after 24 hours of continuous use, without long-term commitments. It eliminates the traditional tradeoff between "Serverless" products, which offer elasticity but higher per-hour rates, and "Provisioned" products, which provide lower unit costs but lack agility and require payment for peak usage continuously. Always-On pricing ensures the database still autoscales for demand spikes, with usage above the baseline billed at standard rates. This innovation also applies to High Availability instances and can be combined with a 50% promotional discount through January 31, 2027.

Key takeaway

For Directors of AI/ML or Data Engineers managing established production databases, you can now optimize costs without sacrificing elasticity. By enabling Always-On pricing on Databricks Lakebase, you secure a 25% lower rate for your baseline compute capacity automatically, while still benefiting from autoscaling for unpredictable peaks. This eliminates the need for manual rightsizing or long-term commitments, allowing you to reallocate budget and focus on innovation. Consider enabling this for consistent workloads, but retain scale-to-zero for new or intermittent projects.

Key insights

Databricks Lakebase's Always-On pricing offers serverless elasticity with a predictable, lower baseline cost, removing traditional database provisioning tradeoffs.

Principles

• Baseline usage can be optimized without commitment.
• Elasticity and cost predictability are not mutually exclusive.
• Automated billing ensures best rates for workloads.

Method

Disable scale-to-zero and set an autoscaling range on a Lakebase Postgres Autoscaling project. After 24 hours of continuous use, the minimum capacity is billed at the lower Always-On rate, while spikes autoscale at standard rates.

In practice

• Apply to production workloads with consistent activity floors.
• Combine with a 50% promotional discount through January 31, 2027.
• Revert to standard autoscaling for intermittent loads.

Topics

• Databricks Lakebase
• Always-On Pricing
• Cloud Cost Optimization
• Serverless Databases
• Postgres
• Autoscaling

Best for: CTO, VP of Engineering/Data, Entrepreneur, Data Engineer, IT Professional, Director of AI/ML

Related on AIssential

• Unlock seamless and cost-effective marketing campaigns with Lakebase — Lakebase merges transactional databases with data lake economics for scalable, cost-effective marketing data serving.
• The Evolution of Data Engineering: How Serverless Compute is Transforming Notebooks, Lakeflow Jobs, and Spark Declarative Pipelines — Databricks serverless compute automates infrastructure management, offering significant cost savings and performance improvements.
• Databricks nears $5bn equity raise at $134bn valuation — Databricks secured $7B in funding at a $134B valuation to expand its AI-driven data platform offerings.
• How We Reduced a Large Spark ETL Job Runtime by 80%: A Practical Databricks Optimization Guide — Systematic Spark ETL optimization can drastically reduce runtime and costs without changing business logic or cluster size.
• A practical guide for platform teams managing shared AI deployments — Rate limiting protects model health; quota reservations ensure critical workload priority in shared AI deployments.
• High Performance Rate Limiting at Databricks — Optimistic rate limiting with in-memory sharded counters and asynchronous reporting significantly improves scalability and latency.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Databricks.