Most of the time, you're right: one is better off going with a solution put together by some other group who made it their core expertise. But if you keep bumping into limits, and have the resources to throw at the problem, you can develop a compelling solution in-house.
It's a bit like the tradeoff (for enterprises) between buying a turn-key solution and going with an opensource project. The latter is often more expensive (once you factor in the cost involved in figuring everything out) BUT will end up much more taylored to your needs—especially if your usage involves controbuting patches to the project.
I literally just now got the joke: the boring company.
Because they bore tunnels.