tl;dr: Demographics, discounting, debt, international student volatility, and AI-driven impacts are breaking the unit economics of hundreds of private nonprofit colleges.

For decades, U.S. colleges expanded into a market that kept growing, kept borrowing, and rarely questioned the price. That market has changed.

• Demand is smaller
• Buyers are more skeptical
• Alternatives are real
• Every student carries a college counsellor and professor in their pocket

The result isn’t a dip.
It’s a structural correction and it is falling hardest on small, tuition-dependent nonprofit colleges.

Source: Integrated Postsecondary Education Data System (IPEDS)

So, a college starts to look fragile if it has:

• <2,000 students
• Endowment < ~$150M
• Tuition-dependent (>70% revenue)
• Discount rate >50%
• Non-elite brand

This is the trend that carries through the 48 colleges that have closed and 40 that have merged from 2020-2025 (identified by Best Colleges reporting Feb 2026). According to their estimates, about 53k students were impacted by these closures.

Here is how we think this collapse extends to the next 500 private nonprofit colleges by 2030:

Big pic: enrollment already peaked in 2010

Total U.S. higher ed enrollment (degree-granting institutions)

Source: National Center for Education Statistics, National Student Clearinghouse

What matters:

• Enrollment is down from peak
• Recovery is uneven
• Growth is shifting toward:
  • public institutions
  • community colleges
  • lower-cost pathways

The system didn’t just shrink—it reallocated demand away from small private nonprofits

Headwind 1: Demographics don’t negotiate

The number of college-age students is declining.

Projected change in high school graduates (U.S.)

• Peak: ~2025
• Decline: ~13% by 2041
• Northeast decline: ~17% (areas like NY seeing -27% declines)

Source: Western Interstate Commission for Higher Education (WICHE)

Inconvenient Implication: Colleges are not just competing harder; they’re competing over fewer students and the future predictions from WICHE only see this worsening ober the coming decade.

Headwind 2: Tuition discounting is trending up

Private colleges aren’t charging what they list.

Average tuition discount rates (private nonprofits)

Source: National Association of College and University Business Officers (NACUBO)

This means that for every dollar of undergraduate tuition list, they give ~56 cents in grant aid to first-time undergraduates and ~51 cents to all undergraduates who received institutional grant aid. And this trend has been moving this way since 2015, according to NACUBO study of Private colleges.

Implication: Pricing power is gone as a tool for tuition revenue. Sticker price ≠ revenue, and admissions must enroll more students just to stay flat (while fixed costs of buildings, staff, and programs increase)

Headwind 3: The loan buffer is weakening

The system relied on debt to bridge affordability.

Federal student loan stress (2025)

• Total outstanding: ~$1.58 trillion
• Borrowers delinquent: ~34% (~6M people)
• Borrowers in default: ~5.3M

Source: U.S. Department of Education

Implication: The financing model is under pressure.

• Borrowing is no longer invisible
• Families are more cautious

Headwind 4: International students = declining revenue

International students have been a quiet tuition subsidy for decades.

Total international students in the U.S.

New international students (leading indicator)

Source: Open Doors (IIE)

What the data show: COVID was rough on international students, the growth pipeline is policy-sensitive and new student flow is declining again.

Additional signals:

• ~45% of colleges reported declines in new international students during early Trump years
• ~52% cited U.S. climate as deterrent

Source: Institute of International Education

Implication: International students are not a stable growth engine and have been trending down as a high-margin dependency.

Headwind 5: Alternatives are gaining share

Students are choosing different paths, especially as articles about AI job disruption for college grads increase.

Source: Young Graduates Face the Grimmest Job Market in Years, NYT 3/24/2026  

Enrollment shifts (recent trend)

• Community colleges: +3.0% (2025)
• Public 4-year: +1.4%
• Private nonprofit 4-year: -1.6%

Source: National Student Clearinghouse

Implication: Demand is moving toward lower cost, clearer outcomes, and flexible formats.

Headwind 6: AI is now a college advisor

Every high school grad has their own personalized GPT College Advisor—a shift powered by emdashes. According to College Board surveys up to 85% of students are using AI in their work. 

What AI changes

1) Information asymmetry disappears

Students can instantly compare cost, debt, salary, and job outcomes.

2) Weak positioning gets exposed

AI surfaces graduation rates, employment outcomes, and financial risk of student debt. 

3) The prestige bias

When students ask for “Best small liberal arts colleges for [Major]” or “Affordable private colleges in [Region],” the AI consistently surfaces the top 50 schools from national rankings. AI pushes decisions toward reputation, economic value, and speed to outcome.

Implication: AI turns college selection from story-driven → data-driven

And many institutions were not built for that.

The college collapse pattern

This is how institutions fail:

1. Freshman enrollment dips (remember only ~75% freshman make it seniors) 
2. Discounting rises
3. Revenue falls
4. Cuts begin
5. Experience degrades (or oops, we cut admissions and they stopped recruiting students)
6. Retention slips
7. AI exposes outcomes (articles like this get noted as a risk of college collapse)
8. Yield declines

Then it is a choose-your-own-adventure:

• Merger (which actually is a great choice if done properly)
• Asset sale
• closure

A clear canary for any college is the freshman enrollment rates because that dictates the total student body for 4+ years. If a student is worried a school might collapse, it impacts their choice, and hello self-fulfilling prophecy (que sound of silence).
Who is safe in the new ed landscape?

Winners

• Elite privates (brand + endowment)
• Large publics (scale + outcomes)
• Low-cost / flexible providers

Squeezed middle: small, tuition-dependent nonprofits

The uncomfortable truth is that there are too many small private colleges for current market demand. This is not about:

• education quality
• faculty strength
• mission

It is about:

• product-market fit
• pricing power
• capital structure
• AI-mediated decision-making

This does not mean every small college fails. It means the sector sorts harder. The likely winners are already visible: elite privates with strong brands and endowments, large publics with scale and better price-value optics, and lower-cost or shorter-form providers aligned with growing demand for flexibility.

The likely losers are small tuition-dependent nonprofits caught in the middle—too expensive to be obvious bargains, not prestigious enough to be irrational purchases, and now far easier to compare in an AI-mediated decision environment.

Get To The Point

The clock was already ticking and current freshman enrollment numbers are showing where trends are going. Demographics weakened. Low enrollment rolled over. Discount rates hit record highs. Loan stress reappeared. Lower-cost alternatives kept gaining traction.

Merge like it matters

Decisions to merge mean more when done earlier. It turns out keeping a failing marriage together “for the kids” isn’t healthy for anyone.

1. Acting early preserves choice and continuity—students get cleaner teach-out pathways, more credit transfer certainty, and less disruption to their degrees.
2. Faculty and staff have time to transition with dignity, negotiate roles, and retain programs rather than face abrupt cuts under financial distress.
3. Institutions with early mergers have greater leverage instead of desperation, leading to better terms, preserved identity, and far less value destruction.

Adjust for the AI College Advisor

Nonprofit colleges need to solve for the Visibility Cliff in AI. New perspective students use AI to help inform decisions that have a Prestige Bias. Admissions offices and marketing need to take their AI Brand Footprint very seriously, very soon as they plan recruiting.

On the front end, it makes the buyer more informed, more comparative, and less patient with hand-wavy value propositions. On the back end, it is quietly hollowing out the labor market outcomes that made the price feel rational in the first place.

The traditional college value proposition rested on three things: knowledge, network, and credentials. AI impacts a solid two out of three and probably doesn’t weigh them the way real humans do.

Small tuition-dependent nonprofits are caught in the middle of both forces simultaneously. They lack the brand power to survive on prestige alone. They lack the scale to compete on price. And they are now selling a product whose ROI is harder to defend to a buyer who just asked an AI to run the numbers.

The only question left is which institutions adapt and which ones the market decides for them.

Sources

1. National Student Clearinghouse Research Center, Current Term Enrollment Estimates. nscresearchcenter.org
2. National Center for Education Statistics (NCES) / IPEDS. nces.ed.gov/ipeds
3. Western Interstate Commission for Higher Education (WICHE), Knocking at the College Door. knocking.wiche.edu
4. National Association of College and University Business Officers (NACUBO), 2024 Tuition Discounting Study. nacubo.org
5. U.S. Department of Education, Federal Student Aid Data Center. studentaid.gov
6. Institute of International Education, Open Doors Report. opendoorsdata.org
7. BestColleges Research. bestcolleges.com/research
8. Inside Higher Ed / Generation Lab AI Survey. insidehighered.com

Update (April 2026): The AI College Recommendation Bias Report

Our original analysis identified AI as one of the forces reshaping higher education enrollment. We’ve now tested this directly: The AI College Recommendation Bias Report reveals how ChatGPT, Claude, Gemini, Grok, and Perplexity systematically steer students toward elite schools.

Key findings from testing 5 AI models across 15 student scenarios:

• 43% of all AI recommendations point to US News Top 25 schools
• AI recommends schools at 3-4x the national average tuition
• HBCUs and HSIs are invisible unless students specifically ask about them
• Stanford appears in 8/15 prompts for every model tested
• The 3,975+ institutions NOT in the Top 25 are functionally invisible to AI

For colleges already facing the demographic and economic pressures described above, AI invisibility is an accelerant. If nearly half of prospective students are using AI to build their consideration set, and AI only recommends ~25 schools, the enrollment pipeline for everyone else just got narrower.

Read the full AI College Recommendation Bias Report →