The user has filed many notes across a session about whether US universities grow or shrink due to AI, with regional dependence. The final instruction: revise the notes as a set — keep every claim that still holds, drop repetition, sharpen wording. Output ONE Markdown note that is the current best statement of what is established. Start with a # title, then short paragraphs, then "What would change this: …". Keep sources or Unverified: tags. Don't mention tools, sessions, or that I'm revising.

Let me inventory all the durable claims across the notes:

**Note 1 (full synthesis, 2025):** By 2045 US higher ed bifurcates rather than shrinks: top ~100 grow, middle hollows into teaching-only, bottom third of four-year schools merge/close. Regional decline tracks demographic map more than AI. AI is sorter/accelerant, not primary cause. Unbundles campus-credential product at mid-tier. Research flagships, land-grants, Sun Belt/SEC growth-tier gain pricing power (population inflow, athletics-attention economies, AI research hubs, workforce pipelines). "Region works through tier."

**Note 2 (duplicate/truncated of Note 1)** — plus demographic details: births fell ~17% from 2007 (~4.32M) to 2023 (~3.59M); HS grads peak ~2025 (~3.9M/yr). Two-front demand squeeze: fewer 18-year-olds AND degree signal weakened (skills directly verifiable). Hardest hit: Northeast/Midwest small privates, open-enrollment regional publics. Most resilient: SEC/Sun Belt flagships + super-selectives (prestige-elastic, not cohort-elastic). Adaptations: admit-standard relaxation, international recruitment contingent on visa policy, credential repackaging. Watch-tell: endowment repurposing from prestige spending to need-based discounting = fighting the cliff.

**Note 3 (meta-planning)** — nothing to file.

**Notes 4/5 (program mix × region, the most evolved mechanism):** Program mix (licensed vs unlicensed share of credit hours) is dominant predictor of institutional direction; region governs whether the winning mix is reachable. Licensed expansion (nursing, engineering, allied health) takes 7–10 years (clinical placements, faculty pipelines, facilities, state approvals) — longer than declining-region enrollment runway → decline locks in. SEC/Sunbelt publics: +10–25% seats through ~2040 as legislatures fund them as economic development. Northeast small privates without licensed core: −30–50% by 2040 regardless of governance quality. Surviving Northeast exceptions: licensed-core institutions with hospital-system or land-grant ties. Formulation: "program mix picks the winners, regional demography picks which losers can still switch." Falsification: if Northeast/Rust Belt pivot attempts succeed at rates comparable to Sunbelt licensed launches, pivot-lag mechanism wrong — decline is governance-fixable, not structural.

**Note 8 (revenue shock):** AI's tuition shock is a revenue event before a headcount event. 2026–2032 mismatch window: seats priced above ~$15k net with no licensed-credential anchor empty fastest; lands first in Northeast/Midwest regional privates. Mechanism: buyers defect, not enrollees vanish. Anchors: degree wage premium (BLS) and AI hollowing entry-level roles (Stanford Digital Economy Lab "Canaries in the Coal Mine" — that's a real 2025 paper about entry-level employment declines in AI-exposed occupations for young workers). Demand migrates to "protected degree + AI skill stack." Online-only and workforce-funded providers ship in months; traditional campus needs 3–7 years through faculty governance and accreditation. Revenue falls while seats hold. Note 8 was truncated mid-sentence ("Compounding and sorting. The shock compounds rather than...").

**Note 9 (instruction staffing cost claim):** By 2035, AI substitution for grading/tutoring/TA labor cuts instructional staffing cost per student credit hour by ≥25% from 2025 levels at US four-year institutions outside top-50; published real tuition at teaching-focused regionals flat or falling as savings competed away. Falsified if: instructional labor share of operating budgets stable/rising in 2035 IPEDS data, or real published tuition at teaching-focused regionals rising above 2%/yr through 2035.

**Note 10/11 (cost structure, savings competed away):** AI instruction savings are competed away, not banked. Oversupplied market (demographic cliff 2025–2038) → cost cuts passed through as tuition discounting → tuition-dependent middle schools gain no margin, only price war. Gains accrue to students and AI-native low-cost operators; endowment-rich elites hold sticker price. AI compresses the middle from both sides. Tuition, not faculty headcount, is