S17.2 · Education, Training & Human Capital
Degree-granting institutions and outsourced campus services, a $700B-$1.2T market where AI erodes the seat-time case for tuition.
Higher education and postsecondary comprises degree-granting institutions and the outsourced services that run parts of their campuses. Precedence Research estimates the global market at $1,042B (2025), though credible estimates range $700B-$1.2T depending on whether public subsidy is included, and the US Title IV-eligible institutional revenue proxy alone runs roughly $700B (NCES/IPEDS). The thing to watch is not enrollment, which keeps growing outside the US. It is tuition's justification: AI is stripping cost out of grading, tutoring support and instructional design, and with it the seat-time logic tuition has been priced on.
The demographic enrollment cliff holds US enrollment growth to low-single-digit. The blended global forward estimate — roughly 11% CAGR to $3,025B by 2035 — is a volume story almost entirely: non-US emerging-market enrollment, not price. Revenue sits in the US, China and the EU; enrollment growth sits in India, Southeast Asia and Africa, and the two maps barely overlap. The institutional base is fragmented across thousands of colleges and universities, with for-profit and proprietary chains holding only single-digit combined share. Concentration appears one layer out from the institutions, in online program management (OPM) providers such as 2U and Pearson and in outsourced campus-services vendors such as Sodexo and Aramark.
Accreditors and government financial-aid systems sit upstream of institutions; students, employers, OPM providers and outsourced-services vendors sit downstream. Durability rests on multi-year cohort persistence — a four-to-six-year enrollment cycle — plus alumni relationships and brand equity that outlast any single student's enrollment. Capital intensity is high and sticky: campus and research infrastructure cannot be shed quickly. Two regulatory gates decide whether an institution operates at all — accreditation and, in the US, Title IV financial-aid eligibility under the gainful-employment rule. Losing either is existential, not a competitive disadvantage.
OPM and online-delivery partnerships are enrollment defense — no new money, protected money — against competitors offering more flexible delivery. Microcredential partnerships with corporate learning open an employer-payer channel that sits alongside tuition rather than substituting for it: genuinely new dollars. Professional certification is the adversarial adjacency — standalone credentials compete directly with degrees as a signal to employers.
Agentic AI is collapsing the cost of grading, tutoring support, and instructional-design and content-authoring work — the labor-intensive functions behind course delivery that used to absorb significant faculty and staff time. The cost collapse itself is not the threat; the threat is what it does to pricing logic, because tuition priced on seat-time rather than demonstrated learning loses its justification when seat-time no longer maps to cost. What stays defensible is not the instructional content. Accreditation, Title IV eligibility and the signaling value of the credential rest on institutional standing, which AI cannot replicate.
A genuinely new boundary opens as AI-verified competency credentials mature: assessment technology that certifies what a learner actually knows, independent of where or how they learned it, blurs the line between a degree and a standalone certification. OPM partnerships and AI tutoring tools are already in production use; the credential-value disruption plays out over 5-10 years, not the next product cycle.