Have you done[published very transparently and with equal emphasis to other technobabble hype] any analysis?
You should be asking Ethereum that question.
You are the one proclaiming it cannot scale because of this theory. Without knowing the increase in average verification costs, this is just FUD.
I don't think that's even possible because their stated goal seems to be a general purpose "world computer". To analyze anything like this, they would probably have to identify not one, but all of their target markets (killer apps) and compare the costs of doing it centralized vs price on Ethereum chain, then decide if there's any benefit from doing it decentralized at a likely higher markup.
The fact that they don't know A) who their target market actually is, or B) how much it can scale at all, makes that a very difficult analysis.
If you really think about it, assuming Ethereum doesn't burn to the ground, with it's limited load capacity you could likely only end up with one small industry using it. One industry would likely find much higher value in it than others and dwarf them in the ability to pay. At that point, there's no reason to have a general purpose "world computer" and Ethereum would probably turn into a transitional technology as a new chain is created specifically for that industry. Or maybe 50 different Ethereum chains to run the lottery and ticket purchases in 50 different states. I'm just slapping myself in the face trying to figure out how this thing is supposed to work at scale on a commercial level.