PostgreSQL vs. MySQL: The Extended Support Tax on Staying Put

Market Position: Big Picture

PostgreSQL and MySQL are the two most-used open-source relational databases, and the comparison most other reports in this database series implicitly assume as the starting point: both are free to run, both have enormous ecosystems, and the real decision most organizations face isn't "which engine costs more," since neither has a license fee, it's which managed hosting path to take and what that path actually costs once real production requirements (high availability, version currency, backup retention) are included.

The finding that should change how any team running either database on a major cloud provider budgets: staying on an older major version carries a real, quantified cost that applies identically to both engines. AWS RDS's Extended Support program charges the same per-vCPU-hour rate whether the aging version is MySQL or PostgreSQL, and it's an easy cost to miss, since it doesn't show up until an organization is already past the standard support window, often the same organizations least inclined to disrupt a stable production database with a major-version upgrade.

Platform Ratings at a Glance

Both engines rate identically in user ratings, and reviewer sentiment tells a consistent story on both sides: praise for reliability, open-source flexibility, and community support, with PostgreSQL reviewers more often citing a steeper learning curve and performance tuning complexity, and MySQL reviewers more often citing difficulty with very large datasets and complex queries. Stack Overflow's 2025 Developer Survey shows PostgreSQL now leading MySQL in developer usage (55.6% to 40.5%), a real shift in the underlying popularity trend that ratings parity alone doesn't capture.

The Real Cost Isn't the Engine, It's the Hosting Path

Neither database has a core licensing fee. MySQL is available under the GPL for most deployments, with Oracle offering a separate, paid MySQL Enterprise Edition carrying additional management tools, security features, and commercial support, relevant primarily to organizations embedding MySQL in a distributed proprietary product, where GPL's copyleft terms would otherwise require releasing the combined work's source. PostgreSQL uses a permissive, BSD-style license with no equivalent commercial-tier fork; there's no PostgreSQL "Enterprise Edition" to license, only third-party support contracts layered on top of the free core.

For the overwhelming majority of buyers, neither licensing model is the real cost driver. Managed hosting is. On identical AWS RDS instance types, the two engines are now priced the same per vCPU-hour at the entry tier (both $0.016/hour on a db.t4g.micro), though some older-generation instance families and independent pricing analyses have shown gaps of up to roughly 10% between the two engines on specific configurations; confirm current rates for your exact instance family and region rather than assuming parity or a fixed gap.

Managed Hosting: A Wide Spread for the Same Open-Source Engine

The real cost variable isn't PostgreSQL versus MySQL, it's which hosting provider and configuration a team chooses. At a comparable mid-tier specification (roughly 16 GB RAM, 100 GB storage), independent 2026 pricing analysis shows managed PostgreSQL running from approximately $86/month (Selfhost.dev) to $475/month (AWS RDS with Multi-AZ and a read replica), a roughly 5.5x spread for materially similar underlying capability. Serverless and usage-based options (Neon, Supabase) add further variability, scaling from a few dollars a month for light, intermittent workloads to several hundred dollars for sustained, always-on usage.

MySQL's managed-hosting landscape follows a similar pattern with a smaller set of specialized providers: AWS RDS, Google Cloud SQL, Azure Database for MySQL, and PlanetScale (MySQL-compatible, with branching-based development workflows) cover most of the market, with PlanetScale's per-branch billing model in particular functioning as a real, easy-to-miss cost multiplier for teams running many development branches simultaneously.

Pricing (July 2026)

TCO Comparison: The Extended Support Surcharge, 4-vCPU Instance, 1 Year

This scenario models the real, quantified cost of running an out-of-support major version on AWS RDS, applicable identically to either engine.

This surcharge is additive, layered on top of whatever the base instance already costs, and it applies whether the aging version is MySQL 5.7 or PostgreSQL 11. It's a real, avoidable cost: staying current on major versions eliminates it entirely, but organizations running stable production databases specifically because they're reluctant to disrupt them with an upgrade are, structurally, the ones most likely to be paying it.

Costs the Pricing Table Misses

• AWS RDS Extended Support applies per vCPU, compounding with instance size. A 16-vCPU production instance pays 4x the surcharge shown above, roughly $2,336/month, for the identical "still on an old major version" situation.
• Multi-AZ and read replicas roughly double or triple the base instance cost. The $475/month AWS RDS PostgreSQL figure cited above reflects Multi-AZ plus a replica; a single-instance quote understates real production cost by a similar multiple.
• PlanetScale's per-branch billing is a real, scaling cost for active development teams. Each development branch can carry its own cost; teams with many concurrent feature branches should model this explicitly, not just the base plan price.
• MySQL Enterprise Edition's commercial license is a real consideration for embedded, distributed proprietary software. GPL's copyleft terms create real legal exposure for certain distribution models; confirm with counsel whether your use case requires the paid Enterprise tier rather than assuming Community Edition's free status extends to your product's redistribution.
• Storage and backup rates vary meaningfully by provider even at similar compute pricing. Neon's $0.35/GB storage rate is notably higher than several competitors; model storage growth explicitly rather than assuming compute is the dominant cost driver at scale.

The Decision Framework

Choose PostgreSQL if: Advanced SQL features, JSONB support, extensibility (PostGIS, pgvector, and similar extensions), and a permissive license with no commercial-tier licensing question matter to your use case. Developer usage trends now favor PostgreSQL, a real signal for hiring and community support depth going forward.

Choose MySQL if: Your team has deep existing MySQL expertise, simpler operational requirements are a priority, or you're building on a MySQL-native platform like PlanetScale. Confirm whether your distribution model triggers GPL's copyleft terms and requires an Enterprise Edition license before assuming the Community Edition is sufficient.

Everyone: Budget for major-version currency explicitly. The Extended Support surcharge documented in this report is a real, quantified, entirely avoidable cost that only exists because an upgrade didn't happen; treat version-currency planning as a cost-management line item, not just an engineering maintenance task.

The Bottom Line

PostgreSQL and MySQL are both credible, well-reviewed, truly free open-source relational databases, and the choice between them increasingly comes down to feature fit and team expertise rather than cost, since neither carries a core licensing fee and managed-hosting rates for both now largely converge on major cloud providers. The real cost story in this comparison isn't PostgreSQL versus MySQL at all, it's the wide spread across hosting providers for the same engine, and the specific, quantified, and entirely avoidable Extended Support surcharge that applies equally to both databases the moment an organization falls behind on major-version upgrades.