But why is this a desirable outcome? It sounds like failure to me. Or at least like retirement.
For an analogy with sports: Good athletes do not retire as early as possible to become coachs. They keep winning medals for as long as they can. When they cannot win anymore, they retire and become "facilitators" (after a significant pay downgrade).
Athletes are seen perform, but not ICs at a company: nobody pays tickets to see John Smith, the TensorFlow master, at work at Google (and when they pay, like for "webinars", is to become like them, not to enjoy seeing them at work).
ICs and managers can move from domain A to domain Z, but you don't see someone playing (or coaching) football for 4 years and then when the stock is vested or whatever equivalent event indicating it is more profitable to move on, start coaching volleyball.
Also, managers, apart from some companies that have the double ladder--IC and leadership--, managers are paid more than ICs. That does not happen in sports, where the top players are certainly paid more than coaches.
Software engineering and software engineering management are both brain work that require a similar level of intelligence. It's the same bell curve if you zoom out a bit.
Also in some sports the coaches make more than the athletes, and in others the athletes make more than the coaches. You can't generalise.
"Why is this desirable?" is the wrong question to ask, IMO. It is not always desirable, as your analogy shows.But the higher value tasks in many business are often not achievable by a single person, like in individual sports. In those cases, it is very valuable to have a good facilitator, and no, they don't take pay cuts doing so.
It's not that the most successful quarterback gets to become the coach. It's the running back who has insight into the strategy, the long term goals adn inter-personal dyanmics that ends up as a good manager.