Chromium is Open Source and embeddable, so my guess is a big part of why their headless variant isn't open source is simply that there isn't really much to it.

From a naive point of view it's easy. But it has to scale and launching a new process for every request works only for a pet project.

So a crawler based on a headless browser that consumes little memory and runs for weeks is a major achivement.

Why doesn't that scale? You might be underestimating Google's resources.