“Relays are not designed…”
There’s no reason you couldn’t specialized in storing data for the long haul. If there is demand, services will spring up. We don’t have an architecture scaling problem at the moment, we have demand problem but its not a nostr issue rather than a unique experience issue which can and will be resolved. There’s a lot of value in what nostr provides from censorship perspective, walled garden breakdown and not needing to be responsible for takedowns. 

I’d say everything is looking up given the “hobbyist” nature of this protocol. 

If scaling becomes an issue; the market will solve it. 

I can’t agree with much of anything you’ve said.