I watched a good chunk of the presentation. I'm immensely skeptical about the potential of playing games via streaming a video feed. The shine of the presentation did however do a good job of reminding me that there is an under-served market for this. Obviously, that market has all these caveats carved into it, but because I'm so negative about this stuff I found it noteworthy that I had positive visions associated with the concept.
OK -- all that said I'm still skeptical about this specifically because there was no talk about what it will cost the consumer.
If streaming your games isn't cheap, there is no market for it.
It's not that people "don't have access" to the non-streaming local hardware option, it's that they don't want to buy it. Or they have the hardware but don't want to buy every little game just to sample everything.
If you have the dough, you can play whatever games you want. Money obviates the need for buying a worse substitute. And if Stadia is too expensive for what you get, there would be no reason someone would pick it. The service would obviate itself. I know there was talk about doing big server side calculations with more GPUs than presumably a regular console would not match, but I'm saying that is/will be a small part of the selling point.
If Google wants the public and those GDC developers to think this will be successful (obv. the two are related), they should have given cost information even if they don't have the final price per month or whatever.
Assuming we had the cost information (and the "what you get" information, like the ownership question and what games are available), the next level of skepticism is whether this business can thrive. @Axel, you said you doubt Google expects it to blow up, but I'm still concerned about where that line is. I don't really know, but I feel there must be a point where not having enough users makes the business unprofitable. Cost per user might be too high, getting licensees onboard might be too hard, etc. if the company doesn't have enough scale or, in the case of things like servers, customer density.
I know Netflix works and is successful, but I think it is a mistake to use this as the fount of optimism. Ironically, the fact that games have such a high ratio of "hours of entertainment per dollar" means that substituting the local hardware model with "hey, we'll send you each hour of entertainment using servers and bandwidth" is inherently that much more expensive than movies.
How are the developers being compensated is another tier of skepticism. I've been seeing this tweet linked around: