**Claim:** Over the next 20 years AI bifurcates the US university system along a research-vs-teaching axis, not a uniform grow/shrink line. Research hubs (R1s, land-grants, academic medical centers) consolidate and grow — AI compresses discovery cycles, and their moat is scarce compute, proprietary datasets, and federally funded labs that AI amplifies rather than replaces. Teaching-centric institutions (small liberal-arts colleges, regional commuter schools) shrink as AI tutors erode the lecture model and credential-attendance coupling. Regionally, decline concentrates in teaching-dependent rural and rust-belt regionals, while growth concentrates where research funding and compute infrastructure density sit — coastal metros and SEC-state flagship publics riding Southern energy/datacenter buildouts. Geography proxies function: the sorting variable is research intensity, not conference affiliation.