If memory is now upstream of so many companies, then maybe a higher regulatory involvement would be appropriate (like power companies).
Specifically, and it could fix all of this, capping the max % of their supply they're allowed to sell into the AI space.
The arguments for infrastructure being managed more tightly by the government is because it's a natural monopoly that everything depends on. Memory isn't a natural monopoly, if we found ourselves in a spot where we all depend on 3 companies then that's not a law of nature, we need more companies.
Memory absolutely fits both these criteria.
As for the net public good weighing, at the end of the day the question would be if it would be more economically efficient to mass produce memory with a managed price structure (i.e. US agrees to buy X amount at $Y regularly).
Which turns into whether having an anchor buyer allows the producer to economically create (and continually run) more production capacity than they would otherwise.
It feels like for semiconductors, the planet as a whole would benefit from maximizing production to floor cost.
Granted, AI infra mania current times is an anomaly, as was internet build-out.
I put "food" and "airplane frames" in the "worth it for state intervention" category personally. Memory is obviously not food, the question is whether it is airplane frames at this point in time. Maybe? I'd give it a solid 55/45 maybe. But we shouldn't sponsor korean/taiwanese companies, so what's the plan for an American memory company?