On ships, you’re making my argument better than I did. Munitions are the post-2022 symptom; shipbuilding is the fifty-year disease. The US built 0.03% of the world’s commercial tonnage last year - 19th place, behind the Netherlands and Vietnam. One Chinese state shipbuilder (CSSC) delivered more commercial tonnage in 2024 than every American yard combined has built since WWII.
That’s not a procurement cycle you fix with an appropriations bill. Navies ride on commercial shipbuilding bases - shared yards, workforce, suppliers - and the American base didn’t die of neglect. It was traded away. The same reserve-currency arrangement that let America import goods and export dollars made domestic shipbuilding (and much else besides) permanently uncompetitive. Financial dominance and empty slipways are the same transaction, viewed from opposite ends. I’ve written that side up here (free): https://www.patreon.com/SenseMakingLens/posts/invisible-defeat-164099950
And your credit/willingness distinction is the crux of the whole thing. Willingness holds while the costs are exported and invisible - that was the genius of the old system. It collapses once the bills arrive at home, visibly, in treasure and blood together. Ukraine is instructive: the West has consistently chosen treasure over blood, and even the treasure is now politically contested. A Taiwan fight demands both, simultaneously, at Pacific distances. Men will bleed for positions they hold as sacred - Thermopylae was a position. But Washington’s stake in Taiwan is instrumental, not sacred: it’s valued for what it does to China’s sea lanes, not for what it is. Instrumental commitments get repriced when the bill arrives; sacred ones don’t. That’s the real question hanging over the whole thread - how durable is the commitment once holding the position gets expensive, in both currencies?
SMIC is already running 5nm-class pilot lines for Huawei on DUV multi-patterning. Yields are poor and the machines lag ASML badly, but the direction only points one way: every year China needs TSMC less.
That’s a closing risk window, not a permanent casus belli. (The Shenzhen EUV prototype is real but isn’t making chips yet - claims beyond that are ahead of the evidence.)
Second, America is duplicating the asset at home. As thoroughly discussed in another Forumosa thread, TSMC Arizona is in production and expanding. Every wafer that can be made in Phoenix shrinks what an American voter is actually buying when asked to bleed for Hsinchu.
Third, the killer: a war fought over cheap semiconductors is the one guaranteed way to make semiconductors not cheap. Fabs don’t survive great-power wars - TSMC’s capacity is a hostage of peace, not a prize anyone can win. So the honest pitch is “fight a war to keep chips cheap; first casualty, cheap chips.” Publics have been sold worse, but not easily and not for long once prices move.
Which is why the positional motive is the durable one - and it’s exactly the one nobody in Washington can say out loud.
Not in the least. Am becoming less and less enamored with AI. It often is uber-PC, spouts mainstream propaganda positions, and on complex topics, needs proper prompting more often than it is worth.