How Oracle’s Grip Is Killing MySQL and Why the ‘Toy’ SQLite Is Rising

Oracle announced a March 2026 layoff round that cut tens of thousands of jobs, with the MySQL team being the first to go. The open‑source community petitioned for independent governance of MySQL Community Edition, but Oracle replied with a terse "No".

Percona engineer Julia Vural’s deep statistics show active MySQL contributors dropping from 135 in Q4 2017 to just 75 in Q3 2025 – a 44% decline over eight years.

The 2025 Stack Overflow developer survey highlights PostgreSQL’s 10% year‑over‑year growth, securing the top spots for "most loved" and "most wanted" databases for three consecutive years, while MySQL has slipped to a "still used but no longer exciting" status.

DB‑Engines ranking curves confirm MySQL’s downward trend, and a Reddit post warns that MySQL’s popularity collapse will accelerate in 2026.

Community demands such as native vector‑search support and a transparent roadmap remain unmet; new features are prioritized for Oracle Cloud (OCI), and pull‑request acceptance appears arbitrary.

MariaDB is positioned as the "natural successor," with Ubuntu 25.04 discussions about making MariaDB the default MySQL variant and growing articles urging developers to stop using MySQL in 2026.

SQLite, despite being mocked as a "toy" for two decades, is actually the world’s most deployed database. It ships in every iPhone, Android device, Mac, Windows 10+, Chrome and Firefox browsers, and even the Airbus A350 flight‑control system – the only database engine certified to DO‑178B aviation standards.

SQLite’s official site claims over one trillion active installations, likely surpassing the combined deployments of all other databases.

Historically, SQLite has been criticized for single‑writer concurrency, lack of replication, and no network access, leading many developers to avoid it in production.

Since 2025‑2026, a suite of edge‑computing tools has lifted SQLite from a single‑machine database to a production‑grade component:

LibSQL (Turso team’s SQLite fork) adds native replication, server mode, concurrent writes, and built‑in vector search.

Turso (hosted LibSQL platform) provides embedded replicas across 30+ edge nodes.

Cloudflare D1 runs SQLite in more than 300 cities.

Litestream continuously backs up SQLite files to S3.

A developer benchmark shows local SQLite reads taking ~0.01 ms per record, Turso’s embedded replica ~0.02 ms, while PostgreSQL in the same region requires 1–3 ms – two orders of magnitude slower because SQLite avoids network round‑trips and connection‑pool overhead.

Reddit users in 2026 recommend starting personal projects with SQLite, arguing that most developers will never need to switch to PostgreSQL.

Hardware and deployment models have evolved, redefining what constitutes the "right tool." While PostgreSQL remains optimal for high‑concurrency writes, distributed transactions, and row‑level security, SQLite combined with Turso or Cloudflare D1 offers near‑zero operational cost for read‑heavy web apps such as blogs, e‑commerce catalogs, SaaS back‑ends, and content sites.

In summary, after 16 years under Oracle, MySQL has shifted from an open‑source star to a mature but sluggish enterprise product, with the community drifting away. Meanwhile, SQLite, long dismissed as a toy, has been elevated to a production‑grade platform by a vibrant ecosystem, illustrating that open‑source project vitality depends more on governance and community support than on code quality alone.