The End of MySQL as We Knew It

For years, MySQL has powered blogs, forums, startups, and some of the largest web platforms in the world. It was simple, fast enough, and open source in spirit and in license. Today, that certainty is gone, and a lot of it has to do with AI.

MySQL’s decline can be tracked on platforms like Hacker News, Reddit, and X, where developers are not debating features but questioning whether the project still has a future at all.

The loudest warning came from Otto Kekäläinen, former chief executive of the MariaDB Foundation and now a software development manager at AWS. In a long blog post, Kekäläinen argued that MySQL is no longer a real open source project and that users should plan to leave. 

He pointed to a sharp fall in public GitHub commits during 2025. He described that in the development process, most real work happens behind closed doors. Engineers don’t use public bug trackers, and external contributors see their patches acknowledged and then silently rewritten by Oracle staff. 

When Kekäläinen ran the RDS MySQL and MariaDB teams at AWS, he says his engineers disliked contributing to MySQL because feedback was rare and authorship often vanished.

On DB-Engines Ranking, Oracle and MySQL still lead, but their score has been declining severely over the last year. 

The Real Reason

This is the part of the story that matters most. MySQL has a general public licence, but modern open source is not just about the license. It is about who controls the code, who reviews changes, and who decides the roadmap.

On that front, MariaDB looks like the project MySQL once was. Development happens in the open. Bugs are discussed publicly; anyone can submit a patch and get it reviewed. That contrast explains why so many Linux distributions, including Debian and Fedora, switched their default database from MySQL to MariaDB more than a decade ago. Wikipedia made the same choice.

Yet the case against MySQL is not as clean as Kekäläinen’s essay suggests.

On Hacker News, the top commenter wrote, “I’m not going to stop using it because it’s not truly open source. I’m going to stop using it because there’s better databases out there.” That line captures a quiet shift. 

For many engineers, the question is no longer Oracle versus community. It is MySQL versus PostgreSQL. PostgreSQL has gained features MySQL still lacks, from row-level security to richer indexing and better extensions. 

Some have already started writing obituaries for MySQL, as vibe coding platforms like Lovable and others rely heavily on Supabase, which runs Postgres, which has been on the rise when compared to MySQL.

“Yes, I think MySQL is dying, sadly,” Aaron Francis, software developer and co-founder of Try Hard Studios, wrote on X. This is similar to others announcing that they are shifting away from MySQL for another database.

Still, Dying is a Strong Word

However, MySQL remains number two on DB Engines, above Microsoft SQL Server and far above MariaDB and SQLite. That ranking does not measure real deployments, but it does reflect how deeply MySQL is still embedded in documentation, tutorials, and job postings. 

Some developers noted that many sites and books use “MySQL” as a generic term for the entire family of compatible databases, including forks. Forks, such as MariaDB, form the open source part of MySQL. 

Usage inertia is powerful. Many teams stay because they started with MySQL, and switching is hard. Also, simplicity remains MySQL’s strongest asset.

The deeper concern is not whether MySQL still works. It does. The concern is where it is going.

Kekäläinen’s technical case is unsettling. He explains that after MySQL 8.0 shipped in 2018, there was no real major release for six years. MySQL 8.1 in 2023 was a short-lived preview. The first long-term release, MySQL 8.4 LTS, arrived only in 2024 and added a few visible features.

Some changes made things worse. When Oracle switched the default ALTER TABLE method to run in place in MySQL 8.0.29, corner cases caused crashes and corruption for many users. Fixes took almost a year. Oracle also changed its release model, introducing behaviour changes in minor releases that users had long assumed were safe upgrades.

Performance complaints followed. Mark Callaghan, one of the most respected figures in MySQL tuning, published benchmarks showing write-heavy workloads running about 15% slower in MySQL 9.5 than in 8.0. That is not catastrophic, but it is the wrong direction.

There’s a Deeper Reason

At the same time, Oracle’s strategic focus became clearer. Vector search, analytics acceleration, and new features appeared in HeatWave, Oracle’s cloud-only MySQL service, not in the open source server. 

Tobias Petry, founder of MySQL Explain, wrote on X, “It’s alarming to see Oracle’s effort for the open source MySQL decline year by year. I couldn’t even name something really interesting they’ve added in the past two years.” When asked about HeatWave, he replied that it is “their closed source hosting offering they probably work a lot on and the reason MySQL doesn’t get much love.”

Workforce cuts made the picture darker. In September 2025, Oracle reduced staff across divisions, including MySQL. This has led to recent maintenance releases containing fewer bug fixes than before.

Despite this, some developers do not see a decline. Several write on Hacker News that MariaDB is “not necessarily better than MySQL” and that, for their use case, with 100 billion rows and tight space limits, MySQL still wins. 

The honest answer is yes. MySQL still leads in replication tooling, diagnostics, and some enterprise features. That is why cloud services like Amazon RDS continue to offer it and why clients insist on compatibility. And migration remains painful. 

MySQL might not be dying, at least not yet. But it is losing its role as the default future choice.

On Twitter, Kristaps Bezbailis, a software developer, described ditching MySQL for Supabase for faster deployment and simpler management. He admitted it almost broke his system and introduced seconds of latency before tuning brought it under 50 milliseconds. “It might totally bomb,” he wrote, “but time will tell.”

That line captures the mood. Teams are experimenting. They are no longer assuming MySQL is the safe long-term bet.

Oracle has every right to build closed services and monetise its assets. HeatWave may be an excellent product. But as open-source history shows, projects controlled by a single vendor age differently. They optimise for revenue, not for ecosystems. 

Over time, communities thin out. Forks become the place where energy moves.

Phil Eaton, a software engineer at Oracle and co-founder of TigerBeetle, suggested a way out. He wrote that he wants “Github, Facebook, PlanetScale, Percona folks to come together and fork MySQL into some new independent foundation.” 

That would mirror what happened when MariaDB was created after the Sun acquisition.

Until then, the market is already voting.