Last time they tried it for gaming - with Stadia - it failed miserable.

It makes some sense for things like office applications (though of late the trust in using American and American-owned service providers for that has plummeted, at least in Europe), but for gaming it has so far failed miserably.

Even the whole artificial bottleneck of data centers being the only place where the most advanced hardware can be found would only translate into gaming moving to be cloud-based if game makers did not react to the current PC components crush by reducing the hardware demand of their products and instead all fully bet on cloud-based gaming as the future, a stupid bet at multiple levels not only due to the risk that gamers would overwhelming chose the games of those game makers who dissented and fit their games to the available home PC hardware instead of the cloud-based one but also because it could give massive influence over the success or not of those games to the cloud gaming services suppliers who could just demand more money to run it in their systems at all or to run them at higher speeds.

In summary, the idea that gaming would shift towards cloud-based gaming instead of going in the direction of continuing supporting last gen hardware, isn't really supported by the results of past attempts at that or the comparative business outcomes for game maker from chosing one over the other.