Dilithium signatures clock ~3.3KB vs ECDSA's 64 bytes — opt-in PQ on Arc means quantum-protected USDC transfers eat ~50x the blockspace. Falcon-512 trims that to ~666 bytes but its floating-point implementation has burned multiple teams, which is exactly why NIST standardized both. Retail won't pay higher fees for theoretical threats, so Arc inherits the Bitcoin P2PK problem at scale: only paranoid early-movers migrate, leaving everyone else as the HNDL target. Validator-level PQ by 2030 lines up with NIST's classical-deprecation timeline, but Phase 1 is mostly an institutional positioning play.

Top comment by @Benthic