# AI bifurcates US university enrollment by region — 'grow or shrink' resolves as 'both'

**Claim (filed 2025, wind-down of the bifurcation line):** Over the 2025–2040 horizon, AI erodes the *signaling value* of the generic bachelor's degree faster than it erodes demand for *AI-adjacent skills*. Because those two effects land on different institutions, US enrollment demand does not grow or shrink uniformly — it splits geographically:

- **Decline side:** Small private colleges and tuition-dependent regional publics in the Northeast, Midwest, and rural America — institutions whose core product *is the credential itself*. When employers can assess skills directly (increasingly with AI tools), the degree bought as a signal loses its premium first, and these schools shrink structurally. This predates AI but AI accelerates it.

- **Growth side:** Large public flagships and SEC-corridor schools in the Sunbelt growth states (Texas, Florida, Georgia, Tennessee, Alabama, the Carolinas). Three reinforcing drivers: sustained population in-migration, employer relocation to low-cost/pro-business states, and student cultures that already treat college pragmatically — as a skill-and-network purchase rather than a credential. These institutions can credibly teach frontier AI skills and keep filling seats even as the generic degree's value falls.

- **Net:** National enrollment roughly flat-to-down (~10–15% below peak), but the map redraws. The regional asymmetry is the answer to the core question: growth concentrated in SEC/Sunbelt publics, decline concentrated in legacy credential-producing institutions elsewhere. The US university system doesn't shrink — half of it does.

**Test by 2030:** If Sunbelt public flagships hold or exceed 2024 enrollment while Northeast/Midwest small privates fall another 15%+ from an already weak base, the bifurcation thesis is confirmed; a uniform national decline (or uniform growth) would falsify it.