He's probably right honestly, just not 100% sure on the timeframe due to infrastructure.
In the same way that digital started to overtake physical and how we came to largely accept account-bound games, people would eventually be fine with a wee bit of lag. And cloud gaming makes sense as the next step in terms of distribution, since if you have a fast enough internet connection that you can play the games fine, it's the same as digital sans the download and need for worry about disk space or ageing hardware.

There's nothing wrong with the concept in itself, but I am pretty weary of how big companies will implement the system, since it can be abused much more easily than digital sales from the sellers side.

I don't think it's inevitable (we've seen plenty of technologies that seemed inevitable, and didn't take. Like 3D TVs, VR sets, 3D games, AR smartphones, etc.) but I do think it's possible. And many people are betting on it to happen.

I do feel they would need Technology, Product and Market figured out for it to work.

Technologically:

1. The HW is not there yet. The way people stream games today is build a bunch of $10K gaming rigs and let people connect to the to play (1 person = 1 rig). This cannot work. It's wasteful (most of the computing power is not used), it's not scalable (you need twice the amount of machine to handle twice the amount of people) and it's abusable (1 person can "take up" a machine 24/7). It's not going to work, the customers would need to pay too much, or will get too little for what they paid for.
   A viable solution would be to create virtual machines (like on Amazon/Google/Microsoft cloud) which would used shared CPU, Memory & GPU resources. And preferably dynamically allocate resources per game (e.g. a FPS will get x10 times the resources of a Point&Click game).
2. The internet/streaming technology is not there yet. Bandwidth-wise it's there in majority of 1-st world countries, but lag-wise and stability-wise, we have a way to go.

Product-wise:

1. The product needs to have a "killer app" (or be a "killer app") so people will flock to it. Netflix isn't the market leader because renting a movie online is easier than in a DVD store. Netflix rules the market because it allows hu8ndreds of Movies & TV shows, all for a small monthly fee, and does it all conveniently (search, subtitle, dubbing, suggestions, auto-next-episode and most importantly: perfect streaming of any platform).
2. The product needs to give value, and be cheap enough for people to say "wow, I really don't need a gaming computer now". If you'll need to pay $40 per month, to only be able to play Ubisoft games - it's never going to work. It'll be a niche product at best. They need to offer 100 AAA games released this year (plus another 900 older games) for people to buy into the new gaming method.

Market-wise:
People need to be willing to jump over to this new way of gaming. Most people are accustomed to certain things and don't like changes. So even if everything else works perfectly, it will still take time for them to switch (especially in the beginning, where not everything will be there, and not everything will be perfect.). So there must be a large-enough group of early adopters, that will like the product and recommend it to their friends, for it to actually work.

I think the concerns people here have are unfounded, this technology cannot be forced upon us.
If Microsoft decide to force everyone use some inferior cloud-based product to be able to play their games - sure some people are fanatics and go for it (because they HAVE TO play a certain game, or Microsoft fans), but most will say "fuck it" and just play a different game. And it won't be anything more than a gimmick.
Only when someone will do it right, will he become the Steam (or the Netflix) of the game streaming world.