# AI and the US university, 2025–2045: thinning, not shrinking — both answers hold asymmetrically

**Core answer.** The system bifurcates and thins rather than uniformly shrinking. Total enrollment drifts from ~19M toward ~14–15M by 2045 (Unverified: projection), but the sharper gap is between institution count and output: campuses close and consolidate while degrees conferred hold roughly steady, because displaced demand migrates to surviving flagships and licensed programs rather than vanishing. The median four-year campus loses 20–40% headcount over the window (Unverified); the selective apex and the licensed-professional segment grow or hold. The durable answer to "grow or shrink" is *both — a barbell: growth at the top and at the licensed tier, shrinkage at the commodity median, thinning not collapsing.*

**AI is an accelerant, not the cause.** Demographic contraction is the primary pressure; AI removes the option value of default enrollment and compresses from two directions at once — demand-side (screening decay of the generalist degree) and supply-side (automation of the instructional stack). US births fell from ~4.32M (2007) to ~3.59M (2023), a ~17% shortfall (CDC/NCHS; Unverified on exact figures), and high-school graduates peak around 2025 at ~3.9M/yr (WICHE, *Knocking at the College Door*). The traditional-age cliff hits 2029–2037, hardest in the Northeast, Midwest, and Plains; WICHE projects pipeline decline through 2037. Regional demand patterns lock in demographically by the late 2030s regardless of what AI does.

**Demand splits by major before it splits by institution.** The degree's *screening