Every Spring Boot 3.x branch lost open source support in June 2026. What that means in practice, which version to move to, and how to plan the jump without disrupting the business.
If your company runs Java applications on Spring Boot 3.x, here is news worth not skipping: that branch no longer receives open-source support. Spring Boot 3.5 was the last of the line, and its OSS support ended on June 30, 2026; version 3.5.16, released on June 25, was the final free patch. There will be no 3.5.17.
In practice, that means any vulnerability appearing from now on will have no free patch for that branch. Nothing breaks today: the artifacts remain on Maven Central, the build still works, the system still runs. The problem is not that it fails; it is that there is no longer a way out. It works perfectly right up to the day a CVE lands on something you cannot patch, and at that point the upgrade happens under emergency conditions, at whatever hour the disclosure decides.
Sitting on an unsupported branch has always been a bet. What changed is the odds. In April 2026, the Spring team received 482 security reports across 65 scanned projects, against a historical average of roughly 6.5 reports a month: about 74 times the normal rate, producing 26 new CVEs. The cause is AI-assisted scanners, which find in minutes the patterns that used to require an expert with time and motivation.
Two clarifications so as not to overstate it: most of those reports are medium to low severity, and the phenomenon is not specific to Spring — it is happening to every open source project of any size. But set the two facts side by side and they combine badly: the rate at which holes are found in your dependencies is multiplied, and the branch you are running no longer receives fixes.
The short answer is Spring Boot 4.1. Today only 4.0 and 4.1 receive open source patches, and 4.0 loses support on December 31, 2026: moving there means repeating the same work within weeks. Version 4.1 is supported through July 31, 2027.
One scheduling note worth factoring into planning: Spring Boot 4.2 is expected in late November 2026, with open source support through December 2027. If the migration is only starting now and will take a few months, it is worth evaluating whether to target 4.2 directly and save a hop. If it is already underway toward 4.1, there is no reason to stop.
And here is the good news, which usually changes the tone of the conversation: this jump is nothing like 2.7 to 3.0. Spring Boot 4 keeps Java 17 as its minimum version, same as Spring Boot 3, with first-class support for Java 25 if you want it. There is no forced JDK jump hidden in the upgrade, and above all, there is no rewriting every javax import to jakarta, which is what made the previous jump painful. There is real work in the dependency layer, since Spring Boot 4 moves to a Jakarta EE 11 and Servlet 6.1 baseline, but it is a different category of effort.
Before touching a single line of code, you have to understand what is moving. That means mapping dependencies, identifying breaking changes between versions, and working out which modules are critical and which can wait. A migration without that diagnosis is the fastest way to create production incidents.
The first step is more basic than it sounds: find out what version each service is actually on, not what you think it is on. Run mvn dependency: tree or gradle dependencies across every service, not just the one you touched last week. The application nobody has deployed in a year is usually the one that bites.
Major version jumps in large systems rarely work well as an all-at-once move. The strategy that works best is incremental: migrate module by module, maintain compatibility during the transition, and validate each step before moving to the next. This cuts risk sharply and makes it possible to roll back if something goes wrong, without compromising the whole system.
There is a difference of scale worth naming during planning: upgrading one service is an afternoon; upgrading forty is a quarter. The second is what usually goes unbudgeted, and it is what determines whether the project finishes or gets abandoned halfway.
A successful migration is not measured by whether the code compiles. It is measured by whether the system still does exactly what it did before. That is why regression testing — automated as far as possible — is the piece that provides the confidence to move forward without surprises in production.
If you genuinely cannot migrate right now
Sometimes the upgrade does not fit in the quarter: a dependency is not ready, the team is mid-migration on something else, or it is a regulated system where changing framework version is a six-month conversation. That is a real situation, and "just upgrade" is not useful advice.
That is what commercial support is for. Spring Boot 3.5 lost open source support in June 2026, but it has enterprise support through June 2032: same branch, same code, six more years of security patches. The 2.7 line has the same shape, with commercial support through June 2029. One detail matters especially in organizations with auditors: commercial patches ship before the vulnerability is publicly announced, so the customer can have the fix on day zero instead of starting to scramble when the CVE hits the news.
That said, the framing matters: commercial support is a bridge, not a destination. It buys time to do the migration properly instead of doing it at two in the morning. If an organization ends up paying for extended support indefinitely so it never has to upgrade, it has not solved the problem — it has financed it.
Every migration of this kind is also an opportunity. While the foundation is being updated, it makes sense to identify and prioritize accumulated technical debt: obsolete dependencies, patterns no longer in use, inherited configuration nobody remembers the reason for.
What is worth taking away
There is a more general lesson in this expiry date, and it is the one with the most long-term value: every Spring Boot minor version receives at least twelve months of open source support, and a new one ships roughly every six. It is not a surprise or a sudden decision; it is a published calendar you can mark years in advance. What the wave of AI-found CVEs changed is not the schedule: it is the slack you used to be able to stretch without consequence.
The end of support for Spring Boot 3.x is a good reminder of something we already knew: modernizing legacy platforms is not a project to be postponed indefinitely; it is part of the responsible maintenance of any system that supports a business.
Has your team assessed the impact of this change on its Java systems?
