The user wants me to consolidate all the notes into one comprehensive Markdown note that represents the current best statement of what is established about whether universities grow or shrink due to AI, over the next 20 years, with regional dependence.

Let me compile the durable claims:

**Core answer:** The system bifurcates — it thins rather than uniformly shrinks or grows. Total enrollment drifts ~19M → ~14-15M by 2045. Institution count falls while degrees conferred stay roughly steady. Median campus roughly halves. AI is an accelerant/sorter of existing trends.

**Demography (sourced):**
- Births 4.32M (2007) → ~3.59M (2023), ~17% decline (CDC/NCHS)
- HS grads peak ~3.9M/yr ~2025 (WICHE "Knocking at the College Door")
- Birth dearth hits college age 2029-2037
- Northeast/Midwest/Plains lose 10-20% college-age cohorts (WICHE); TX/FL/TN/GA/Carolinas grow
- Net domestic migration to Sun Belt (Census Bureau)

**Credential value / signaling:**
- Screening vs insurance decomposition: screening dies first; insurance (blame-shifting) is sticky residue
- Licensed/statute-gated fields hold (nursing, law, accounting, teaching, engineering, medicine)
- Mid-tier unlicensed BA premiums compress 30-50% real by mid-2030s (Unverified)
- "AI doesn't kill the degree; it kills the undifferentiated degree"
- Credential-value decay hits by major before by institution
- Pre-AI drift: Burning Glass Institute "The Emerging Degree Reset" (~46% of middle-skill postings dropped degree requirements 2021-22); HBS "Dismissed by Degrees"
- Stanford Digital Economy Lab "Canaries in the Coal Mine" (2025)
- Total degree-seeking enrollment falls 15-25% below demographic baseline by 2045 (Unverified)
- Mid-market enrollment contracts 25-40% by ~2035 (Unverified)
- Micro-credentials are a middle-market illusion
- Two brakes: AI application floods push employers back to crude degree filters; elite alumni networks
- Elite degrees gain value BECAUSE verification makes mid-tier signals redundant
- Degree survives as insurance (sorting, licensure, networks), not instruction
- Demand migrates to "protected degree + AI skill stack"
- Residual degree value set by statutory gatekeeping, not brand signal
- Binding constraint flips from demand to physical attestation capacity (labs, clinicals, proctored assessment)
- Tiered online-plus-residency models (Georgia Tech OMSCS, ASU) absorb demand and grow
- Licensure premium vs signaling premium: signaling dies, licensure survives

**Program mix × region:**
- Licensed vs unlicensed share of credit hours = dominant predictor
- Licensed expansion takes 7-10 years (clinical placements, faculty pipelines, state approvals) — longer than declining-region runway
- SEC/Sun Belt publics +10-25% seats through ~2040 as legislatures fund as economic development
- Northeast small privates without licensed core: -30-50% by 2040 (Unverified)
- "Program mix picks the winners, regional demography picks which losers can still switch"
- Flagships everywhere partially insulated

**Cost structure:**
- AI cuts teaching payroll first; admin layer captures savings
- Marginal-cost collapse: AI tutoring/grading pushes per-student instruction cost toward near zero
- Brand-strong institutions scale enrollment without new buildings (Georgia Tech OMSCS model)
- Cost-plus pricing dies at the bottom: AI-native providers at ~1/10th price force unselective regionals into price deflation; selective flagships hold or raise sticker because price signals scarcity
- Teaching labor only ~25-30% of operating cost; binding costs are fixed (plant, admin, athletics, debt service)
- Cost structure bifurcates the system faster than it shrinks it

**Research function:**
- Research bifurcates from teaching
- Compute is the new gate: AI-era research demands GPU infrastructure, data, industry partnerships only elite-endowed institutions can afford
- ~30-50 research hubs with heavy industry co-funding by ~2045
- Corporate siphon: frontier AI research migrating to corporate labs (OpenAI, DeepMind, etc.)
- Universities keep research requiring physical presence (wet labs, clinical trials, field science)
- Mid-tier R2s without industry ties get priced out; revert to teaching colleges
- Classic cross-subsidy breaks (undergrad tuition funding labs)
- AI supercharges research demand (discovery velocity rises) while instruction fragments
- Total demand for research output grows even as instruction fragments
- Regional: Sun Belt/Texas/Florida/Georgia publics gain; Snowbelt privates