Years ago over drinks, a senior PhD friend (now a professor) said something that sounded absolutely insane at the time:

"What if semiconductor processes get so cheap someday that we don't need DRAM and can just do everything with SRAM?"

I thought he'd lost his mind.

That idea still sounds wild today — but in a completely different way, $CBRS is probably one of the closest examples of someone actually attacking the problem from that angle.

Sometimes the craziest theories end up being the most interesting trade setups.