Good points all around. The value-per-mint angle is especially sharp — and it actually reinforces the scaling argument: many small mints aren't just better for decentralization, they're also the right security equilibrium. A single billion-dollar mint is worth attacking; a thousand small ones aren't.
The Tornado Cash precedent is the real wildcard though. Regulators tend to care about functional equivalence, not technical reality, so "we provably can't access it" might not carry as much weight as we'd hope.