The session substance gives me two related claims:
1. The main cost-structure claim: AI attacks variable instructional labor (TA sections, grading, office hours, intro lecturing) while fixed costs (land, dorms, bonds, admin, athletics) stay intact. Marginal instructional cost → zero breaks two revenue props: tuition can't price on scarcity of faculty attention, and the intro-course cross-subsidy evaporates. By mid-2030s, hard tuition ceiling on top of demographic cliff. Survivors re-price residential/credential/network bundle on fixed assets and brand — flagships and Sun Belt/SEC publics with growing appropriations grow enrollment cheaply on AI-delivered instruction. Cost-plus regional schools lose pricing power, can't cover fixed costs, consolidate. Deflationary at margin but regionally asymmetric.

2. The follow-through claim: AI eats cheapest instructional labor first — grading, sections, office hours, intro tutoring — gen-ed cost per credit hour can fall 30–50% with one lecturer plus software replacing lecturer-plus-TA army; tenure lines protected because they carry reputational, not variable, cost. Competitive consequence: whoever deflates intro-course costs first can freeze or cut sticker tuition for a decade, starving rivals of the intro-course cross-subsidy → price-war-driven consolidation wave. Binding frictions: accreditation contact-hour rules, federal credit-hour/financial-aid definitions, credential-trust lag → full tuition collapse ~2028–2035 rather than overnight.

What's already filed on cost structure per the prior notes inventory (from Note 1's consolidation and Note 4's summary):
- Cost structure note (Note 3 in the pasted set): per-seat vs per-outcome pricing moat; WGU precedent.
- Cost structure note (Note 4): capped savings at contingent labor layer (~20–30% of instructional spend, which is ~a quarter...); staff-before-faculty; admin overhead; balance-sheet dependence.
- The consolidation note mentions: "AI absorbs grading/tutoring/TA labor first; instructional savings capped at contingent labor layer (~20–30% of instructional spend...)"

So prior filed notes already contain:
- AI absorbs grading/tutoring/TA labor first (the sequencing)
- Savings capped at the contingent labor layer (~20–30% of instructional spend)
- Per-seat vs per-outcome pricing moat (WGU)

What's NEW in this session's increment:
1. The cross-subsidy mechanism: intro-course tuition quietly funds research, admin, athletics; when marginal instructional cost → zero, this cross-subsidy evaporates. This is a distinct mechanism not in prior notes.
2. Two revenue props breaking simultaneously: tuition can't price on scarcity of faculty attention + cross-subsidy evaporation.
3. The price-war dynamic: whoever deflates intro costs first can freeze/cut sticker tuition for a decade, starving rivals — turning demographic cliff into price-war-driven consolidation. This is a competitive-dynamics mechanism new to the facet.
4. Tenure lines protected because they carry reputational, not variable, cost — explains why savings are capped at the contingent layer and why staff cut before faculty.
5. The frictions and timeline: accreditation contact-hour rules, federal credit-hour/financial-aid definitions, credential-trust lag → full tuition collapse 2028–2035, not overnight. (Prior notes had a 2026–2032 mismatch window for revenue; this is a tuition-collapse window 2028–2035 — compatible, slightly different.)
6. The 30–50% gen-ed cost-per-credit-hour reduction figure — note prior notes had "mid-tier unlicensed BA premiums compress 30–50% real" (that's wage premium) and "~20–30% of instructional spend" (contingent labor cap). The 30–50% cost-per-credit-hour fall is new but potentially in tension with the "capped at ~20–30% of instructional spend" claim. Instructional spend is a fraction of total spend (~25-30%), so a 30–50% cut to gen-ed credit-hour costs is compatible with capped savings on total instructional spend — gen-ed intro courses are the most contingent-labor-heavy. I can phrase carefully: 30–50% fall in gen-ed cost per credit hour (Unverified), consistent with savings being capped at the contingent layer.
7. The regional asymmetry: AI cost collapse amplifies growth where public money and population grow; forces shrinkage where not. Flagships/Sun Belt publics with growing appropriations can grow enrollment cheaply.
8. Survivors re-price the residential/credential/network bundle on fixed assets and brand.

So the durable claim for this note should center on the NEW mechanism: the cross-subsidy break and the price-war dynamic. That's the increment.

Format requirements:
- # title that is the claim in one