>The fact is that Stadia makes it possible to play games on devices where it previously was impossible is pretty compelling. (even if it's at 720p).

It makes it possible to play modern AAA games on devices where it previously was impossible.

That's a dramatic difference. There are plenty of incredible gaming experiences you can have on low-powered hardware, without needing a clunky streaming-as-a-service infrastructure to back them up. They just won't be the latest FIFA/Elder Scrolls/Call of Battlefield: Modern Duty.

Honestly, I tend to think that having environments that don't have full fat AAA games available is good for the gaming market as a whole. It opens up markets for indie developers and new formats. I'm thinking of the Wii and Switch for comparison-- if they had been built with the same spec hardware as their competitors, they would have been flooded with ports of the big multi-platform franchises, rather than the unique experiences they're known for.

Sonme of my favourite gaming experiences include playing Zangband on an already-obsolete Pentium II in 2003, and Cave Story on a pre-Atom netbook starring a 900MHz Celeron-M.

I think this is truly the crux of the product's potential. Triple A FPS games are never going to hit it big on a streaming service, especially in any sort of competitive circle.

The real opportunity is for games where latency isn't really important. Beautiful turn based games, cinematic choose-your-own adventures.

The second path is for games to be developed with appropriate tolerances & latency-forgiving mechanics. Take a modern platformer like Super Meat Boy or Celeste, where you are given a window of frames in which you can jump/react before failure. Even modern games like Mario Odyssey accommodate input latency of wireless controllers/HDMI displays. Similar accomodations can and will be made for games targeting streaming services.