On the Adaptations (Before the Exigencies) in Higher Education

On the Structural Pressures Ahead, Including AI

Last week, Oracle announced plans to cut thousands of jobs across the company, with some of the reductions explicitly targeting categories “that can be bolstered using AI.” This comes alongside Oracle raising $45-50 billion in new financing to fund AI data center expansion. Spend more on machines, spend less on people. The logic is not subtle.

Oracle is not an outlier. Entry-level hiring at the fifteen biggest tech firms fell 25 percent from 2023 to 2024. Amazon eliminated 14,000 corporate roles last year, citing AI-enabled “leaner structures.” Workday cut 1,750 jobs to “reallocate resources toward AI investments.” In the UK, tech graduate roles fell 46 percent in 2024 alone.

I mention all of this not because I think AI is going to wipe out the labor market wholesale. I don’t. But something is already happening at the bottom of the career ladder that higher education cannot afford to ignore: the entry-level jobs that degrees were supposed to unlock are thinning out, and the institutions that sell those degrees have not yet reckoned with what that means for their value proposition.

This is the downstream signal that connects to the upstream crisis I wrote about in my recent piece on the collapsing value proposition of higher education (https://kylesaunders.substack.com/p/the-collapsing-value-proposition). That essay was about structural, temporal, and epistemic unbundling that is happening at universities.

This one is about what happens when the labor market starts sending those signals back.

A Few Closures Are Already Here

In 2025, at least 16 nonprofit colleges announced closures in the United States (https://www.insidehighered.com/news/business/mergers-collaboration/2025/12/18/colleges-couldnt-survive-2025). That matched the prior year and continued a steady, grinding trend: over the past eight years, more than 100 colleges have closed or merged. A Federal Reserve study published in 2024 predicted 80 closures over the following five years, roughly 16 per year. We are tracking to that estimate almost exactly.

The closures are overwhelmingly concentrated among small, tuition-dependent private institutions. Limestone University in South Carolina shut down after failing to raise $6 million against a $12.6 million deficit. Martin University, a historically Black university in Indianapolis, “paused operations” in December and then let all staff go. These are not peripheral institutions. They served real students and real communities. But they were structurally vulnerable: small endowments, high tuition dependence, limited brand differentiation, and exposure to demographic decline. And that demographic picture is not improving. WICHE projects a 12-15 percent decline in high school graduates by the late 2030s, with sharper drops in the Midwest and Northeast (https://www.wiche.edu/resources/knocking-at-the-college-door-11th-edition/). That is no longer just a future scenario. It is a pipeline contraction already underway, and every institution in the “structurally vulnerable” category feels it first.

It is worth noting that enrollment has rebounded modestly since the pandemic lows. Flagship publics and well-positioned regionals are doing pretty well. Community colleges are recovering. So, the crisis is real but, just as the quote goes, its “future is already here — it’s just not very evenly distributed,” which is part of what makes it politically hard to address: the institutions in the most trouble are often the least visible, and the institutions with the loudest voices are often the least threatened.

Fitch Ratings labeled the entire higher education sector’s 2026 outlook “deteriorating” (https://www.highereddive.com/news/higher-education-faces-deteriorating-2026-outlook-fitch-says/807222), citing a fragile student pipeline, expense pressures, and public doubt about the value proposition. Moody’s and S&P Global issued similar assessments. When three credit rating agencies converge on “deteriorating,” the language is doing structural work, not rhetorical work.

Bryan Alexander, who has been meticulously tracking closures, cuts, and mergers for years now (https://bryanalexander.org/horizon-scanning/), captures the broader pattern well: the humanities bear the brunt of cuts, programs are eliminated for underenrollment, faculty lines are consolidated, and increasingly, even elite institutions are not immune as federal research funding disruptions and political pressure extend the pain upward.

The steady drip of 15-20 closures per year may not sound catastrophic. But it is the visible edge of a much larger contraction: programs cut, departments consolidated, faculty bought out, adjunct lines eliminated, and campuses quietly shrinking their academic footprint while marketing the same aspirational brand.

What Forced Restructuring Actually Looks Like

If the closures are the visible edge, West Virginia University shows what happens just behind it. And I don’t mention this to say “this is going to happen tomorrow,” instead, “this is what happens if institutions don’t adapt.”

In 2023, WVU initiated an accelerated academic program review tied to a persistent structural deficit: enrollment down roughly 10 percent since 2015, limited state support, pandemic revenue losses, and construction debt. The Board of Governors approved cuts to 28 majors and 143 faculty positions. Graduate mathematics was eliminated. World languages were cut. Large sections of the College of Education were restructured.

Faculty resistance was intense. Votes of no confidence. Protests. National media coverage. And understandably so. These were people’s careers, intellectual communities, and in many cases the only regional access point for certain fields of study. But the Board pushed through anyway, because the underlying constraint did not go away.

WVU had specific local conditions that accelerated its crisis. But the underlying fiscal and demographic pressures are not unique to West Virginia. They are playing out at different speeds across hundreds of institutions.

Exigency as the Only Lever?

Ah, there’s the word I used in the title: “exigency.” It’s one you hear on a campus now and again when things are getting dark. For you non-academics who have made it this far, it’s the term for when an organization (in this case a university) makes a formal declaration of severe financial distress that allows the institution to take necessary actions, such as restructuring programs or reducing staff, to stabilize its finances.

What WVU illustrates, and what the steady drip of closures and the Fitch “deteriorating” outlook confirm, is an uncomfortable pattern: in some portion of American higher education, fiscal exigency may be the only lever powerful enough to break entrenched veto points quickly. (What portion that is, well, we’ll talk about that in another essay, I think, and soon.)

Tenure and shared governance in academe create multiple veto points that make voluntary restructuring extraordinarily difficult. That is by design; those structures protect important things. But they also mean that once an institution crosses the threshold into acknowledged unsustainability, the political economy flips. Decisions centralize. Timelines compress. Trust erodes. Everything becomes existential.

This is not to say that slow adaptation is absent. Online expansion, adult-learner targeting, transfer pipelines, program triage, and experiments with shorter degrees are all underway at various institutions. But those incremental moves rarely reach the structural depth that the pressures demand. Institutions tend to adapt late, noisily, and after incurring significant damage. Not because the people inside them are foolish, but because the governance structures are built for stability, not transformation.

The Entry-Level Job Market Is Sending a Signal

While closures and restructuring represent the supply-side crisis, the demand side is sending its own signals. And they are not encouraging.

The entry-level job market for college graduates is in rough shape. NACE’s Job Outlook 2026 survey found employer sentiment at its most pessimistic since 2020. Entry-level postings in software development and data analysis have fallen sharply, just as an example.

The mechanism is straightforward: AI tools can now perform many of the tasks that once justified hiring a junior employee. Drafting memos, writing code, summarizing research, building presentations, analyzing data. These were the apprenticeship tasks, the things new graduates did to learn the job while contributing at the margin.

When those tasks get automated or dramatically accelerated, firms face a choice: hire the junior person anyway (for development purposes) or don’t (for efficiency purposes). In a tight-margin environment, efficiency tends to win. Especially when AI’s cost per task keeps falling.

AI is not the sole driver here. Post-pandemic correction, overhiring hangovers, high interest rates, and sector-specific retrenchment all play a role. But AI is accelerating a structural shift in what entry-level work looks like, and that shift is unlikely to reverse.

This creates what one analysis called “eating your own seed corn.” If companies stop hiring juniors now, in five years they may face a shortage of experienced professionals who understand the systems beneath the AI abstraction layer. But that is a long-run problem, and firms optimize for the short run.

For higher education, the immediate implication is this: the degree-as-readiness signal is weakening not just because of credential proliferation or employer skepticism, but because the jobs that degrees pointed toward are structurally changing. If entry-level positions in knowledge work are thinning out, the promise “get a degree, get a job” becomes harder to make with a straight face. Healthcare, government, and trades remain robust hiring sectors, and college still pays off on average. But the sectors where the degree premium was most visible to middle-class families are exactly the ones under the most pressure.

And students know it. Their behavior is already shifting: more transfers, more credential stacking, more “what’s the fastest path to something that pays.” The aspirational model of four years of exploration is giving way to a more transactional calculation, one that universities are not well positioned to win on price. You can see it in program-level enrollment patterns: nursing, computer science, and applied health fields are growing, while humanities, general social sciences, and undifferentiated business programs are contracting. Students are not abandoning college. They are repricing it, program by program.

The disruption is not limited to the labor market for graduates, either. AI is also reshaping how universities themselves produce knowledge. Alexander Kustov recently made a compelling case…

Academics Need to Wake Up on AI

…that AI can already perform recognizable components of social science research at a level that should unsettle most professors, and that the commercial journal system may come under serious pressure from the resulting flood of low-cost manuscript production

Academics Need to Wake Up on AI, Part II

If that’s even partially right (NB: it’s more than partially right, imho), the internal epistemic machinery of universities is under the same kind of pressure as their external signaling function. More than a few institutions are being squeezed from both sides.

The Missing Receipt

One of the reasons the value proposition erosion is so hard to address is that the accountability infrastructure barely exists. Or as Doan Winkel put it: “Trust is doing the job that data should do.”

Winkel, writing on his Substack “How to Teach With AI,”

College’s Biggest Scam Is the Missing Receipt

had his students map the information flows in the “graduate employment system” and found that every exchange point is either blocked or filtered. Employers cannot evaluate readiness beyond credentials. Universities lack real post-graduation outcome visibility. Parents see big milestones but not actual development. And no one has agreed-upon metrics for graduate readiness.

When the bundle was intact and the labor market reliably absorbed graduates, that trust worked well enough. Families paid. Employers hired. Institutions collected tuition. Outcomes were assumed.

Now the bundle is fragmenting and the labor market is sending noisier signals, and the system has no good mechanism for accountability. Federal College Scorecard data exists but remains coarse. Program-level outcomes are inconsistent. Employer feedback to universities is essentially nonexistent.

When the bundle breaks, people demand the receipt. And most institutions have to do a better job than they are of producing one.

Why Institutions Resist Change (Even When They Know They Should)

So the closures are mounting. The labor market is shifting. The accountability gap is wide open. And most institutions are doing... not much.

The structural constraints are real. Accreditation, financial aid rules, rankings, alumni expectations, and faculty governance all create institutional lock-in. But those constraints are not the whole story.

Institutions also resist change for psychological and social reasons that run deeper than governance rules. I’ve written about this dynamic at length: how groups screen new ideas for social safety before evaluating them on merit (https://kylesaunders.substack.com/p/how-groups-kill-new-possibly-good), and how AI in particular triggers identity threat and motivated reasoning in academic settings (https://kylesaunders.substack.com/p/why-higher-educations-ai-backlash). The short version: AI doesn’t just challenge how students complete assignments. It challenges how higher education understands itself, how it certifies competence, and how it justifies its authority. That’s an identity-level threat, and identity-level threats get managed socially before they get evaluated analytically.

That combination of structural lock-in and psychological resistance explains why so many institutions are stuck in a posture that everyone privately recognizes as unsustainable: preserve the brand of “real college” while quietly managing its dissolution. Layer flexibility onto inherited structures. Avoid the confrontation. Hope the numbers hold.

Three Paths Forward (None of Them Are Easy)

So how will institutions actually adapt over the next decade or so? I see three broad trajectories, not as predictions but as gravitational pulls that the current pressures point toward. None is costless, and none guarantees success, because universities are increasingly competing inside a dynamic, negative-sum environment. Some institutions will drift toward one of these models gradually. Some will try the hybrid option for as long as they can until they can’t. And when I say “human differentiation” below I mean mentorship, intellectual formation, judgment, and disciplinary training. If those drift into vibes and co-curricular theater, they fail.

The Credential Factory. Strip the bundle to its minimum: fast, cheap, labor-market-aligned credentials. Competency-based programs, stackable certificates, employer partnerships, accelerated timelines. There is even a growing debate about whether the bachelor’s degree itself should be compressed to three years, a structure already common in the UK and parts of Europe, which would cut both cost and time-to-credential. This is where Western Governors University already operates, and where many community colleges are heading over time. It serves a real need. But it risks abandoning intellectual formation entirely and becoming a throughput machine. Without serious intellectual formation, it is just branding.

The Elite Mentorship Institution. Double down on the scarce human goods: mentorship, research apprenticeship, intellectual formation, elite socialization, network access. Charge accordingly. Serve a small, well-funded population. This is where the Ivies and strong liberal arts colleges already sit. It can work beautifully for the students who access it. The risk, over time, is that it becomes a luxury good for the wealthy, with democratic access shrinking to a few scholarship recipients.

The Hybrid Knowledge Platform. Try to combine elements of both: AI-augmented content delivery at scale for foundational material, paired with high-touch human mentorship, experiential learning, and supervised performance for the scarce, valuable stuff. Partner with employers for real-time feedback. Invest in outcome transparency. Redesign assessment around demonstration, not artifacts. This third path is the most promising and the hardest to execute. It requires institutions to honestly distinguish between what is scalable and what is scarce, price them differently, and stop pretending the old bundle still exists. No institution has yet pulled this off convincingly at scale, which is itself diagnostic of how hard the transition is.

Most institutions are not currently structured for any of these futures, and the transition toward them will be slow and uneven. As I laid out above: institutions don’t just resist change structurally. They resist it socially and psychologically, through the group dynamics that kill new ideas before they get a fair hearing. That resistance is far from a moral failing; it is just a predictable outcome of how human cognition and institutional incentives interact. But it does mean that the institutions most in need of transformation are often the least capable of initiating it voluntarily.

The Democratic Stakes

In my head, I keep coming back to a line I used in my previous essay (https://kylesaunders.substack.com/p/the-collapsing-value-proposition): the postwar democratic settlement assumed that expanding higher education expanded democratic inclusion.

That settlement was always mixed. Higher education expanded opportunity, but it also expanded credential sorting and status closure. What is changing now is that its integrative promise is getting harder to sell.

If universities drift toward bifurcation, splitting into elite legitimacy factories and mass credential processors, educational stratification becomes political stratification. Epistemic stratification becomes democratic fragility. That trajectory is not inevitable, but the pressures are pointing in that direction, and the forces working against it are weaker than they used to be.

The weakening entry-level job market makes this more acute, not less. If the labor market no longer reliably absorbs degree holders into stable careers, and if the institutions that issue those degrees cannot demonstrate outcomes, then the entire social contract around higher education weakens.

So, that is not just a campus problem—it can also become a democratic legitimacy problem.

When the Exigency Comes

WVU is what happens when the exigency arrives at one institution. The trickle of closures and cuts happening across the sector are what it looks like when exigency arrives piecemeal, institution by institution, program by program. And every institution is fighting to not be WVU. The numbers suggest some will.

The weakening entry-level job market is the demand-side pressure that makes the supply-side crisis harder to ignore. AI is not the sole cause, but it is an accelerant, and it is unlikely to become less capable or less accessible over time. These pressures are not going to resolve in a year or two. They are structural, and they will unfold over the next decade.

The missing receipt problem means institutions cannot easily prove their value even when it exists. And the accountability vacuum means the system can continue under-delivering for longer than it should. But not forever.

On top of that, universities are also now embedded in a broader, uncontrolled experiment in epistemic governance (https://kylesaunders.substack.com/p/we-live-inside-the-experiment-now). Pedagogy, research, credentialing, and institutional authority are all being tested simultaneously by technology, demographics, politics, and platforms. No one is steering this process, and there is no stable equilibrium to fall back on. Higher education is adapting in real time, under contested legitimacy, while the ground keeps moving.

Exigency is a symptom when it happens. The real challenge is whether universities can redesign their value propositions before negative-sum constraint forces the issue. Whether they can be honest about what they offer, what they can no longer sustain, and what they exist to do in a world that no longer defers automatically.

If they cannot, the system will change anyway. Just not on its own terms.

---

If you enjoyed (hated) this piece, please share it with someone who you think would appreciate (hate) it. Thanks. :)

---

Related pieces from Sacred Cow BBQ:

• “The Collapsing Value Proposition of Higher Education (and What We Can Do About It)” — https://kylesaunders.substack.com/p/the-collapsing-value-proposition

• “Why Higher Education’s AI Backlash Reveals Some of Its Deepest Cracks” — https://kylesaunders.substack.com/p/why-higher-educations-ai-backlash

• “How Groups Kill New (Possibly Good) Ideas in ‘The Room’” — https://kylesaunders.substack.com/p/how-groups-kill-new-possibly-good

• “AI Won’t Fix Our Epistemic Problems, It Just Reveals Them” — https://kylesaunders.substack.com/p/ai-wont-fix-our-epistemic-problems

• “Trust, Critique, and the Problem of Knowing Together” — https://kylesaunders.substack.com/p/trust-critique-and-the-problem-of

• “We Live Inside the Experiment Now” — https://kylesaunders.substack.com/p/we-live-inside-the-experiment-now

• “Trust, Incommensurability, and the Limits of Intersubjectivity” — https://kylesaunders.substack.com/p/trust-incommensurability-and-the

Worth reading from others:

• Alexander Kustov, “Academics Need to Wake Up on AI”

• Alexander Kustov, “Academics Need to Wake Up on AI, Part 2”

• Doan Winkel, “College’s Biggest Scam Is the Missing Receipt”

• Bryan Alexander’s campus closure/merger tracking — https://bryanalexander.org/horizon-scanning/

---

Data Sources

• Oracle layoffs (March 2026): Bloomberg — https://www.bloomberg.com/news/articles/2026-03-05/oracle-layoffs-to-impact-thousands-in-ai-cash-crunch; Reuters — https://www.reuters.com/business/oracle-plans-thousands-job-cuts-data-center-costs-rise-bloomberg-news-reports-2026-03-05/

• Entry-level hiring decline: SignalFire via IEEE Spectrum — https://spectrum.ieee.org/ai-effect-entry-level-jobs

• AI-attributed layoffs 2025: Challenger, Gray & Christmas via AIMultiple — https://research.aimultiple.com/ai-job-loss/

• Entry-level job market analysis: HR Dive (Jan 2026) — https://www.hrdive.com/news/AI-entry-level-jobs-talent-pipeline/809413/

• Entry-level crisis report: Rezi.ai (Jan 2026) — https://www.rezi.ai/posts/entry-level-jobs-and-ai-2026-report

• AI and career ladder: CNBC (Sept 2025) — https://www.cnbc.com/2025/09/07/ai-entry-level-jobs-hiring-careers.html

• College closures 2025: Inside Higher Ed — https://www.insidehighered.com/news/business/mergers-collaboration/2025/12/18/colleges-couldnt-survive-2025

• College closures tracking: BestColleges — https://www.bestcolleges.com/research/closed-colleges-list-statistics-major-closures/

• 2026 closures projected: The College Fix — https://www.thecollegefix.com/16-colleges-closed-in-2025-and-more-could-shut-down-in-2026/

• Bryan Alexander closure/merger updates — https://bryanalexander.org/horizon-scanning/campus-cuts-closures-mergers-and-layoffs-for-winter-2025-2026/

• Fitch Ratings 2026 outlook: Higher Ed Dive — https://www.highereddive.com/news/higher-education-faces-deteriorating-2026-outlook-fitch-says/807222

• Federal Reserve closure projections: Kelchen, Ritter & Webber (2024), Federal Reserve Bank of Philadelphia

• WVU Academic Transformation: AP — https://apnews.com/article/west-virginia-university-academic-faculty-cuts-245527c044cc2cfe80bcbe8c2eda7e98

• NACE Job Outlook 2026: National Association of Colleges and Employers

• WICHE demographic projections — https://www.wiche.edu/resources/knocking-at-the-college-door-11th-edition/