Cloud infrastructure and developer platforms are increasingly embracing usage-based pricing models, moving away from traditional subscription fees. This shift, championed by companies like serverless Postgres provider Neon and deployment platform Railway, aims to offer greater cost efficiency and flexibility to developers and businesses. The model ensures users pay only for the resources they consume, aligning costs directly with actual usage.
Neon, recently acquired by Databricks in May 2025, significantly updated its pricing in August 2025 to a fully usage-based structure for its serverless PostgreSQL offering. This model bills customers based on compute unit-hours (CU-hours), storage (GB-months), and active branches, with a $5 monthly minimum for paid plans. Neon's approach, which includes features like scale-to-zero for inactive databases, is particularly suited for dynamic, AI-native workloads where resource demands fluctuate.
Similarly, Railway operates on a hybrid pricing model that combines a base subscription fee with charges for actual resource consumption, including RAM, CPU, network egress, and volume storage. The platform's plans, such as Hobby and Pro, include a credit towards usage, with additional consumption billed at specific rates. Railway emphasizes that this pay-per-use system prevents overspending on idle resources, offering substantial savings for its users.
The growing adoption of usage-based pricing reflects a broader industry trend towards transparency and value alignment in cloud services. This model contrasts sharply with traditional enterprise software licensing, which some developers, like Jake in a recent tweet, have criticized as "SSO/Enterprise Tax Ransomware!" The sentiment underscores a desire for pricing structures that eliminate hidden costs and provide clear, proportional billing. This approach allows businesses, from startups to large enterprises, to scale their infrastructure dynamically without being locked into rigid, high-cost contracts.