I think it provides potential for them. At least I see it as the main point, where it can differentiate. How it's used in practice is of course up to developers.
What's listed is indeed not a good example of how it can provide unique features. Most games are too focused on pushing graphics, instead of pushing complex simulation.
It's possible, but complex simulations have a lot of room to grow before they stop fitting on desktops. Go from one to ten threads and you still fit on a mid-high end CPU.
That's the point. If they are pushing it to the cloud, they can at least use it for something desktops can't handle. If they have a cluster, let them make some game with AI that can utilize it, instead of making it a replacement for something desktops can already do well.
What I mean is, they could vastly expand it and still not need a cloud. They would have to vastly vastly expand it, which is possible but much less likely.
And for AI you can already handle that in the cloud if you actually want to. The innovations you can get on Stadia are mostly tied to rendering capabilities.
> The innovations you can get on Stadia are mostly tied to rendering capabilities.
Which is my point exactly. They should have focused on innovations where cloud really can differentiate. Instead they run for "push pixels from the server, because your PC isn't strong enough" idea, which will be obsolete tomorrow, since PC hardware is getting better at pushing pixels every year.
What's listed is indeed not a good example of how it can provide unique features. Most games are too focused on pushing graphics, instead of pushing complex simulation.