Open source built the internet in the last couple of decades, yet its future is in jeopardy.

More than 99% of companies rely on OSS, 70%+ of production code comes from it, and a $100T+ digital economy is built on licenses designed for another era. Yet the maintainers carrying that infrastructure are barely funded and burning out, with 60% having quit or considered quitting, while only a tiny fraction of companies give back.

​Most of us have felt some version of this. But it's worth being precise about why the model that got us here no longer ensures the ecosystem can sustain itself — and why AI is bringing the question to a head now.

​Why the FOSS model no longer ensures sustainability

​The dependency-graph visibility paradox. Value propagates upward through software stacks; compensation does not. The deeper a library sits below the application layer, the more critical it is to everything above — and the less visible it is to anyone with a budget. That isn't incidental; it's a structural property of how abstraction works. The organizations benefiting most from foundational infrastructure are precisely the ones most insulated from perceiving it.

​The licenses predate the relationship they now govern. MIT, Apache, and GPL were built for communities with aligned incentives — peers sharing among researchers and hobbyists. They optimize for sharing, not for governing a power asymmetry between trillion-dollar enterprises and individual maintainers with no salary and no recourse. Companies free-ride because the license gives them every right to. The model was never designed for this relationship.

​VC incentives don't align with stewardship — and AI is compounding the gap. Venture capital is optimized for returns at exit, and funding your OSS dependencies produces no return at exit — it's pure externality from an LP's perspective. Even companies with genuine intentions face that structural pressure. And AI foundation models have now accelerated the extraction rate exponentially: training on open source corpora at scale, compressing decades of accumulated work into model weights, with no compensation mechanism proposed or enforceable. The clearest sign of the gap: when developers sued over models trained on millions of public repositories, the case became one of the most closely watched AI copyright disputes heading into 2026 — yet courts let the license claim proceed while largely dismissing the developers' demand for compensation. The extraction is visible. The recourse isn't.

​This is the start of a conversation, not a conclusion.

​We don't think anyone has fully solved this yet — ourselves included. The Prosperous Software Movement begins from a single conviction: that sustainability should be structural, not dependent on goodwill. Everything past that — what the mechanism is, who it serves, how it works — is exactly what we want to think through with the people who live this problem, rather than hand down as answers.

​So we're gathering the community to begin.

​Come to:

​- Talk through, in the open, why the current model falls short
​- Meet maintainers, developers, and the companies that depend on them
​- Help shape what a more sustainable model could look like — and pressure-test the ideas, including ours
​- This is one of the first gatherings of a community we're hoping to build. Bring your perspective — that's the point.

Join our Telegram group to stay in touch: https://t.me/prosperous_software
We are also seeking advisors to help build the movement.