I don't believe dropping to 100‑man teams will be enough.
A big problem is that we now have such big budgets that they need approval higher up, and worse, you end up looking like a fool when your project fails. That is still the case with 100-man teams.
This is not necessarily a terrible situation to be in if it were a startup situation where you could count on having 90% failures and the winner making the whole bucket worth it and people knew and accepted that.
But outside Sand Hill Road nobody gets that, so the strategy is to play it safe, safe, and safe, and then you end up with boring games.
It also doesn't help that it used to be that gamers would reliably update every couple of years, so graphics cards and processes would get faster, and you could make a generic shoot‑'em‑up game that was significantly better than the previous shoot‑'em‑up just because you now had more to work with.
Computers have actually increased in price, so much so that no one buys to update their computers. And at the same time, you need permission from an extremely high level to make games, so you can only make games that are very safe.
Meanwhile, the golden age of indie games is out there. It's never been easier to make your creative game and it's never been easier to find an audience for it.
As a result, there are more games than ever on Steam, and those fun games can last a lot longer because they don't have the marketing budget, so there's no big, immediate launch-month splat window. People will find them on social media, watch their favourite streamers play them etc.
Until we get to a point where games are created by studios who accept that 90% of them will fail, we are going to have interesting games from non‑studios.
That leaves a hole in the options for game development, though. Where are we with games that can be developed by 50 people but not by four? Those games do not have a home right no; They are too risky and uninteresting for the big guys, and they are out of scope for the independent.
The entire game-development business is being torn apart and remodeled by AI development right now. It is very likely that the big companies may never be able to make a game with 500 people.
Many games that were previously too large and required too many people to be feasible for indie developers are now very feasible for them. It's also quite possible that exclusives, such as for PlayStation, will very quickly disappear due to widespread piracy aided by AI powered decompiling. The future is extremely hard to predict at present, even more so than for non-gaming segments of the industry.