Why open-sourcing a product is a competitive act

A large firm releases, free and permissively licensed, the exact component a smaller rival sells for a living. The announcement is about openness and community. The effect is that the rival’s product now has a price ceiling of zero, and the large firm still charges for the hosting underneath it.

What everyone sees

An open release is read as generosity or as a bid for developer goodwill, both of which are usually genuine motives held by the people involved. They are not the strategic function. The release changes the price at which that layer can be sold by anyone, and the firm choosing which layer to commoditise is the firm that earns elsewhere.

What is actually happening

Lerner and Tirole’s analysis of the economics of open source identifies strategic complementarity as a principal corporate motive: firms contribute to and release code that raises demand for a complementary product they do sell, since a cheaper complement expands the market for the thing being charged for. Fitzgerald’s account of the transformation of open source describes the same logic operating at ecosystem scale, with commercial participation concentrated where the released layer is adjacent to a proprietary revenue point. This is the mechanism behind the pattern visible in the 2020s AI stack, where model weights are published by firms selling compute, tooling or distribution, and withheld by firms selling access to the model itself. The licence choice is a statement about where the margin is.

Why it stays hidden

The move hides because everything said about it is true. The code is genuinely free, the community benefit is genuinely real, and the engineers involved genuinely want it released. None of that is inconsistent with the release having been approved because it lowers a competitor’s ceiling, and the strategic memo is not published alongside the repository.