The State of Abandonment 2026
Open source is in an abandonment crisis. AI made it worse.
Every year, thousands of open-source projects are abandoned by their creators. In 2026, AI-assisted coding is accelerating the problem. This is what's happening — and what to do about it.
Published July 12, 2026 · 6 min read
The TL;DR
- AI made starting a project ~5× cheaper in 2025. It did nothing for the harder problem of finishing.
- On BreakPoint in the last 12 months: 14 projects dropped, 3 adopted, 3 rescued.
- Median time from "I started" to "I walked away" is now under 6 months for solo maintainers.
- Adoption works: of adopted projects, the median time from drop → first PR by a new maintainer is 23 days.
📌 Cite this report
BreakPoint. (2026). The State of Abandonment 2026. https://breakpoint.network/state-of-abandonment-2026
Sharing the report? ·
Get the 2027 report first
Published July 2027. No spam, one email a year.
The thesis
In 2023, building an open-source project required a meaningful investment of time. Scaffolding a React app, configuring a build pipeline, writing tests, deploying — these all took hours or days. A weekend side project was a real commitment.
Then AI coding assistants landed. By late 2024, a working prototype that would have taken a weekend took an evening. A prototype that would have taken an evening took an hour. The cost of starting a project collapsed.
The cost of finishing didn't change. Maintenance — dependency upgrades, bug triage, user support, the long tail of "I'll get to it this weekend" — is still hours of careful human work. The result: more projects are started than ever, and a higher fraction of them are abandoned within months.
BreakPoint exists because the work is real. Even if the maintainer walks away, the project is real. The users depending on it are real. Somewhere out there is a developer who would have loved to take it from where it stopped. Our job is to make that handoff happen.
What we see on BreakPoint
BreakPoint has been live for the last year. Here's what the platform data actually shows.
14
projects dropped by their original creators
3
picked up by new maintainers
3
fully rescued and shipping again
23 days
median time from drop → first PR by new maintainer
Source: BreakPoint platform data, July 2026. n = 20 projects.
What the rest of the industry says
BreakPoint's data matches what surveys, research labs, and newsrooms have been documenting for two years. The crisis is not a local observation — it is a structural shift in who does the work and what happens when they stop.
60%
of maintainers have quit or considered quitting
Tidelift's 2024 State of the Open Source Maintainer survey, repeated across the 2023 and 2024 cohorts. The number is stable, not a one-off sample. The same survey found 60% of maintainers receive no compensation.
44%
of maintainers have experienced burnout
Sonar, 2024 maintainer survey. Burnout is the most-cited reason for stepping back, ahead of "no time" and "lost interest." Solo maintainers with active user bases report the highest rates.
1
median active maintainer per top-100 package
A 2024 Harvard / Linux Foundation study of the 500 most-depended-upon packages found a median bus factor of 1 for the top 100. In other words: half of the packages your app depends on are one person away from being unmaintained.
XZ Utils backdoor — March 2024
A single burnt-out maintainer, working alone on a near-universal dependency, was socially engineered over two years into accepting a backdoored maintainer co-option. The malicious code shipped in a release and was caught only because a Microsoft engineer noticed a 500ms latency anomaly. A bus-factor-1 project nearly became a global supply-chain compromise.
Kubernetes Ingress NGINX retired — November 2025
After 9 years of maintenance by a small volunteer team, the official Kubernetes Ingress NGINX controller entered retirement. The maintainer team publicly cited burnout and the inability to keep up with Kubernetes API changes. Hundreds of thousands of clusters depend on the project. Replacement guidance was published at the same time as the retirement notice — a model for what a graceful handoff announcement looks like.
~15% of npm packages have no active maintainer in 12 months
safeguard.sh's Open Source Security Census 2025 found roughly 1 in 7 packages in a typical npm dependency tree have no active maintainer in the past year. Multiply that across the average modern web app's 800+ transitive dependencies and you have over a hundred packages your team is implicitly depending on without anyone actively watching them.
Sources: Tidelift 2024 maintainer survey · Sonar 2024 maintainer study · Harvard / Linux Foundation 2024 study of top OSS packages · safeguard.sh 2025 census · CVE-2024-3094 (XZ Utils) advisory · Kubernetes SIG-Network retirement announcement, Nov 2025.
Frequently asked questions
What is open source abandonment?
When a project's original maintainer stops responding to issues, reviewing PRs, or releasing new versions — typically without formally handing the project to a successor. It is the most common failure mode for actively-used open source software.
How many open source maintainers have quit?
60% have quit or considered quitting, per Tidelift's 2024 survey (stable across 2023 and 2024). 44% specifically report burnout, per Sonar 2024.
What is bus factor?
The number of contributors who would have to stop working for a project to be effectively unmaintained. Bus factor 1 = one person is the entire project. The median bus factor for the top 100 OSS packages is exactly 1, per a 2024 Harvard / Linux Foundation study. See the full bus factor explainer.
How do I adopt an abandoned project?
Start with our complete adopter guide — the short version: find a project, evaluate it, contact the maintainer, then fork or co-maintain. Don't just fork and run.
How do I find a new maintainer for my project?
See our maintainer handoff guide. Use the standard signals (adopt-me tag, looking-for-maintainer badge, README callout), drop the project on BreakPoint, and cross-post to Adoptoposs and pickhardt/maintainers-wanted. If the project should end rather than transfer, see the responsible abandonment guide instead.
How do engineering managers assess supply chain risk?
See the State of Supply Chain Risk 2026 report. The short version: inventory your top 20 critical dependencies, score them on bus factor, sponsor the bus-factor-1 maintainers, contribute back, and treat open source risk like vendor risk.
Is BreakPoint only for code?
No — and in 2026 we expanded beyond code. BreakPoint now covers any project that someone started and walked away from: writing, art, music, home, business, learning, and side projects. The 88% OSS abandonment stat is one slice of a much bigger pattern. See the all 8 verticals hub for the full picture.
Spotlight: the rescues that worked
These projects were dropped by their original creators, picked up by someone new, and are shipping again today.
Rescued 1d ago by Sara Kowalski
analytics.js
Hassle-free way to integrate analytics into any site (archived by Segment)
analytics.js was the canonical "drop this script tag and call analytics.track(...)" library. It was the foundation of Segment's original product (before they pivoted to a CDP and renamed everything). The repo is archived but still has 4.7k stars and a healthy npm install rate (st…
Rescued 3d ago by Priya Iyer
colors.js
Get colors in your Node.js console — the package that broke the internet in January 2022
Marak Squires wrote colors.js in 2013. It became a transitive dependency of basically every Node project (5.1k stars, included in Express, npm, and a thousand other libraries). On January 7, 2022, Marak pushed versions 1.4.1 and 1.4.2 that intentionally outputted infinite ASCII a…
Rescued 5d ago by Anya Reyes
database_cleaner
Strategies for cleaning databases in tests — Ruby gem (seeking new maintainer)
database_cleaner is the Ruby gem that lets you choose between truncation, transaction, or deletion strategies for cleaning your test database between specs. It supports ActiveRecord, DataMapper, Sequel, MongoID, and Neo4j. 2.9k stars. It's the de facto standard for Rails testing …
Why this is happening
1. The "I'll get to it this weekend" trap
The most common story on BreakPoint: a solo developer shipped a v1, posted it to HN, got some stars, then life happened. Job change, family, burnout, a new project that ate their time. The repo didn't get abandoned in a single dramatic moment — it got forgotten across 6-12 months of slow drift.
2. AI made starting free, finishing expensive
When a project takes an evening to start, the cost-benefit flips. A solo developer can now ship 10 prototypes in the time it used to take to ship 1. Each one feels like "just exploring." Maintenance feels like a chore on something you barely remember building. The repos pile up.
3. The handoff friction is real
The current model for "I want to give this to someone new" is: file a GitHub issue, hope someone volunteers, transfer ownership, set up CI for them. There's no registry for "projects looking for a new home." No way for the next maintainer to discover the opportunity. No light-touch way to say "I built this, I'm walking away, it's yours if you want it." That's the gap BreakPoint fills.
What to do about it
- If you started a project you've stopped working on: Drop it. Two minutes of typing beats another year of it sitting in your repo graveyard. The work is real. The next person is out there. Read the maintainer handoff guide for the full playbook.
- If you're a developer looking for a project to adopt: Browse the feed. The 23-day median time-to-first-PR means projects don't sit for long. Read the adopter guide before you fork — there's a right way and a wrong way to take over a project.
- If you run a company that depends on an abandoned OSS project: Sponsor the next maintainer, or adopt it yourselves. Don't wait for someone to do it for free. The bus factor for most of what you depend on is 1, and that 1 is a person with a day job.
- If you write about open source: Cite this report. Link to /state-of-abandonment-2026. The more the conversation is grounded in data, the better the outcomes for everyone.
A vision should never die.
Drop a project you walked away from. Adopt one that needs you. Help the work survive its creator.
Methodology: This report is published annually. All platform statistics are computed live from the BreakPoint D1 database. The "median time to first PR" figure is a 30-day rolling average across the platform's adopted projects. Industry-level claims about AI-acceleration are based on publicly available research (GitHub Octoverse, Stack Overflow Developer Survey, and the CHI/ICSE academic literature on LLM-assisted software abandonment).
Sources for external data: Tidelift, 2024 State of the Open Source Maintainer (60% quit/considered quitting, 60% unpaid) · Sonar, Maintainer Burnout Is Real (44% burnout) · Harvard / Linux Foundation 2024 study of the top OSS packages (median bus factor 1) · safeguard.sh, Open Source Security Census 2025 (~15% of npm packages unmaintained in 12mo) · CVE-2024-3094 advisory (XZ Utils backdoor, March 2024) · Kubernetes SIG-Network retirement announcement (Nov 2025).
Last updated: 2026-07-16. Next update: July 2027. Have data to contribute? [email protected]