Abstract
The University of the People advertises itself as tuition-free, and the phrase is both accurate and incomplete. It is accurate because the institution charges nothing for instruction, course materials, or enrollment; it is incomplete because students still pay per-course assessment fees, and because “free” describes who is not billed rather than what the education costs to produce. This paper argues that the model’s low cost rests not on discounting the conventional university’s price but on deleting its three most expensive line items—the physical campus, licensed content, and a large salaried faculty—and replacing them with open educational resources, volunteer instruction with automated assessment, and peer-mediated learning. The decisive economic question is therefore not whether the model is cheap today, which it plainly is, but whether its unit economics survive growth and whether the cost it removes is genuinely eliminated or merely shifted onto students’ time, volunteers’ unpaid labor, and donors’ generosity. The paper reconstructs the conventional cost structure, specifies the inversion, models the behavior of marginal cost at scale, and subjects the model to five sustainability stress tests. It concludes that the model resolves cost more convincingly than any conventional reform but that its durability depends on a rising fee floor, a sustained supply of free academic labor, and a two-tier funding structure whose mission layer remains donor-dependent.
1. The Inversion Thesis
A discount lowers a price while leaving the thing priced unchanged; an inversion changes the thing. The standard responses to college cost have been discounts in this sense—tuition freezes, expanded aid, income-share agreements, three-year degrees—each of which adjusts who pays or how much, while leaving intact the campus, the library of licensed materials, and the salaried professoriate that together generate the cost. The University of the People (UoPeople) does something different. It removes those generators. There is no campus to heat, no licensed textbook to buy, and no full-time teaching faculty to pay; in their place stand open educational resources, a volunteer instructional corps, automated grading, and small asynchronous peer cohorts. The result is a degree a student can complete for low four-figure totals rather than five- or six-figure ones.
This is why the model is the cleanest available case for the cost dimension identified in White Paper 1. Where a conventional institution can only chip at cost because its cost drivers are load-bearing, UoPeople has built a structure in which those drivers are simply absent. The analytic task is to specify the inversion precisely, to determine where the removed cost actually went, and to ask whether a structure built on free labor and philanthropy can hold its shape as it grows from the scale of a college to the scale of a small nation’s worth of students.
2. Methods Note
This paper combines documentary cost analysis with comparative benchmarking. The conventional cost structure is reconstructed from the Delta Cost Project’s longitudinal analysis of federal IPEDS finance data, which decomposes institutional spending into functional categories (instruction, academic support, student services, institutional support, operations and maintenance, research, public service, and scholarships) and tracks their movement over time (Desrochers & Hurlburt, 2016). The UoPeople structure is reconstructed from the institution’s published fee schedules and catalog, from its founder’s public statements on sustainability, and from independent reporting and commentary, including critical treatments. Where the institution’s claims are self-reported—per-student break-even, volunteer counts, enrollment—the paper treats them as claims and flags them as such; the model’s financial transparency is itself one of the findings.
Two cautions govern the analysis. First, the comparison is asymmetric: UoPeople reports far less public financial detail than IPEDS requires of Title IV institutions, so several quantities must be inferred from fee structure and founder statements rather than read from audited statements. Second, the paper distinguishes throughout between cost to the student (price), cost to the institution (production cost), and cost to the system (including unpaid labor and time). Conflating these is the central error the paper exists to avoid, and the limitations section returns to it directly.
3. Where the Money Goes: The Conventional Cost Structure
To see what UoPeople removes, one must first see what a conventional university spends. The Delta Cost Project’s decomposition is the standard reference. Instruction—faculty salaries and departmental support—is typically the single largest functional category, yet it does not dominate the budget the way intuition suggests. Somewhere between 40 and 50 percent of general spending, exclusive of sponsored research and auxiliary enterprises, goes to overhead or indirect costs: academic support such as libraries and computing centers, student services such as counseling and career centers, institutional support such as legal and fiscal administration, and the operation and maintenance of buildings and grounds. The cost of a degree, in other words, is substantially the cost of everything around the teaching.
The trend lines matter as much as the snapshot. Over the long run, the share of spending devoted to the direct cost of instruction has held roughly flat or declined at most institutions, in part because they have contained instructional cost by shifting toward part-time and adjunct labor, while the non-instructional categories have grown faster. Per-student spending on student services more than doubled at private institutions between 1987 and 2013, and academic support and institutional support also rose markedly. This is the empirical face of the administrative-growth dimension from White Paper 1 and of Bowen’s revenue theory of cost: where revenue can be raised, spending expands to absorb it, and much of that expansion lands outside the classroom (Bowen, 1980; Desrochers & Hurlburt, 2016).
Two categories deserve emphasis because they are precisely what UoPeople deletes. The first is plant—buildings, utilities, maintenance, the residential and recreational infrastructure of campus life. The second is the salaried instructional and administrative workforce, the largest recurring expense at most institutions. A model that carries neither has, before it spends a dollar, removed the two heaviest weights in the conventional budget.
4. The Inversion: Deleting Line Items Rather Than Discounting Them
4.1 The campus, set to zero
UoPeople has no campus. Its only physical footprint is an administrative office in Pasadena; instruction, library access, advising, and assessment all occur online (University of the People, 2025). The entire operations-and-maintenance category—construction, utilities, grounds, security, residential and dining facilities—falls out of the budget. Because plant costs are largely fixed and grow with ambition rather than with teaching quality, removing them eliminates a cost layer that the conventional university cannot reduce without ceasing to be residential. This single deletion accounts for a large fraction of the model’s advantage and is, by construction, unrecoverable by any campus-based competitor.
4.2 Content, replaced by open educational resources
In place of licensed textbooks and proprietary courseware, UoPeople builds its curriculum on open educational resources and open-source technology (PBS NewsHour, 2025; Reshef, in FairPlanet, 2020). The economic property that matters here is that open content is a fixed, largely sunk cost rather than a recurring per-student one: once a course is assembled from open materials, additional students consume it at near-zero marginal cost, and the student’s textbook bill—often a four-figure annual expense at conventional institutions—disappears entirely. The institution’s deliberate choice to keep core materials text-based, with video optional, also lowers the bandwidth threshold for participation, a design decision with both pedagogical and access consequences taken up in White Papers 4 and 5.
4.3 Faculty, replaced by volunteers, automation, and peers
The largest inversion is in labor. UoPeople relies on a corps of volunteer instructors drawn from established universities—numbering in the thousands, with the president, provost, and vice provost themselves serving without pay (Reshef, in FairPlanet, 2020; the model was built with a few thousand volunteers from other universities and open-source technology). Routine assessment is automated: course assignments and much of the grading process are automated, and the institution runs with a very small paid staff—on the order of four people in admissions—leaning on technology to fill the gaps. Peer-to-peer learning, examined on its merits in White Paper 5, performs part of the instructional work that salaried faculty perform elsewhere, with students assessing and supporting one another within small cohorts.
The economic logic is straightforward and the economic risk is equally so. By converting the single largest recurring expense—instructional and administrative salaries—into a mix of volunteer time, automated process, and peer effort, the model drives its production cost toward the cost of the platform plus a thin layer of paid coordination. The risk is that two of these three substitutes are not infinitely elastic: volunteer goodwill is a donated resource that may not scale linearly with enrollment, and automated assessment has limits in fields requiring judgment. The model’s labor economics are its greatest strength and its most exposed assumption at once.
4.4 Tuition, replaced by the assessment fee
Finally, the model replaces tuition with a per-course assessment fee. The current schedule is a one-time application fee of sixty dollars, an assessment fee of $160 per undergraduate course and $400 to $450 per graduate course, with a small charge for transfer-credit evaluation. The resulting program totals are low by any conventional measure: an associate degree near $3,260, a bachelor’s near $6,460, and an MBA near $5,460. The institution frames these against conventional benchmarks aggressively, claiming that the bachelor’s costs roughly 93 percent less than the average in-state public tuition and that the MBA, against an average U.S. public MBA cost the institution cites at about $66,300, totals roughly $5,460. Even discounting the marketing framing, the order-of-magnitude difference is real.
The conceptual point is that the assessment fee is not tuition under another name. Tuition is a charge for instruction and enrollment; the assessment fee is a charge tied to evaluation, levied per completed course and waivable through scholarship. It functions less as a price for teaching than as a cost-recovery mechanism for the lean operations that remain after the three large line items are gone. Whether that mechanism can carry the institution’s full cost as the institution grows is the unit-economics question.
5. The Funding Architecture: Fees, Philanthropy, and the Scholarship Layer
The model’s revenue rests on two tiers that should not be confused. The first is fee revenue, which the founder maintains is sufficient on its own to cover core operations. By his account, the institution reached financial sustainability once enrollment passed roughly 15,000 students, at which point per-student assessment revenue covered operations independent of donations; the figure cited at the time was on the order of a hundred dollars per student per assessment (Reshef, in FairPlanet, 2020; Reshef, in Kritica Economica, 2020). If that claim holds, the core teaching-and-credentialing operation is self-funding from student fees, which would distinguish UoPeople sharply from conventional non-profits dependent on appropriations or endowment draw.
The second tier is philanthropy, and it does specific work: it funds scholarships and the mission-driven expansion that fee revenue does not. The institution has drawn support from foundations and technology firms, including Hewlett-Packard and Microsoft, and from major prizes such as the Yidan Prize (Financial Times, 2014; the institution has relied largely on philanthropic funding alongside an occasional large prize, such as the multimillion-dollar Yidan Prize, while counting principally on volunteers). Donor-funded scholarship pools—named funds for Afghan women, for refugees, and others—cover the assessment fees of students who cannot pay them (scholarship funds include an Afghan Women’s Scholarship Fund, a Catalyst Scholarship for Refugees, and named donor scholarships). A “pay it forward” mechanism lets paying students fund others’ assessments directly.
The architecture is therefore best read as a self-funding core wrapped in a donor-funded mission. This division is the key to the sustainability analysis: the cost dimension and the access dimension are financed differently, and a stress that threatens one tier need not threaten the other. The core is exposed to fee-revenue risk; the mission is exposed to donor risk.
6. Unit Economics and Marginal Cost at Scale
The model’s defining economic feature is the behavior of marginal cost. In a conventional university, adding students eventually requires adding sections, faculty, classroom space, and support staff, so marginal cost is meaningful and rises in steps. In the UoPeople design, the three largest cost components behave differently. Content, built from open resources, is a fixed cost spread over an unbounded student base; campus cost does not exist; and instruction is supplied substantially by volunteers and automation rather than purchased per section. The marginal cost of an additional student therefore approaches the cost of assessment, a thin slice of paid coordination, and incremental platform load—small relative to the assessment fee that student pays.
This is what allows the institution to claim a break-even at roughly 15,000 students and then to grow to a reported enrollment exceeding 170,000 without a proportional cost increase (University of the People, 2025; the institution’s paid administrative staff remains in the dozens rather than the thousands). The economics are those of a software platform more than those of a college: high fixed cost in building courses and systems, very low marginal cost in serving each additional learner, and a surplus over marginal cost on every fee-paying student that can cross-subsidize scholarship students.
Two qualifications keep this from being a story of costless scale. First, not every cost is fixed or volunteer-supplied. Academic governance, quality assurance, accreditation compliance, student support, and assessment integrity (including proctoring) carry costs that scale at least partly with enrollment, and these are precisely the costs that a regional accreditor expects to see funded adequately. Second, the volunteer input is a quantity, not a price: at 15,000 students a few thousand volunteers suffice, but whether free academic labor scales to hundreds of thousands of students without thinning per-student attention is an empirical question the model’s growth now poses directly. Near-zero marginal cost is real, but it is bounded by the scalability of the unpaid and lean inputs that produce it.
7. Sustainability Stress Tests
7.1 The rising fee floor
The most telling evidence about the model’s economics is the trajectory of its fees. The “tuition-free” label has been constant; the assessment fee has not. A decade ago the undergraduate charge was about $100 per course, putting a bachelor’s near $4,000 (Financial Times, 2014); it rose to $120, and now stands at $160, putting a bachelor’s near $6,460. The graduate fee has climbed further in proportional terms: the MBA assessment fee was introduced at $200 per course, for a total near $2,400 (Financial Times, 2016, on the launch of a roughly $2,400 MBA), and now stands at $450 per course for a total near $5,460. In a decade the undergraduate fee has risen by roughly 60 percent and the MBA total by more than double.
This pattern is the model’s own version of the dynamic White Paper 1 identified in conventional institutions: cost recovery rises toward what operations require. The fees remain low in absolute terms, but their direction is upward, and the gap between “tuition-free” as a slogan and a rising assessment charge as a reality is the first thing a sustainability analysis must register. There appears to be a floor below which the fee cannot fall and still fund the quality and compliance that accreditation demands, and that floor has been moving up.
7.2 Donor dependence in the mission tier
If the core is fee-funded, the mission is not. The refugee-education work, the named scholarship funds, and the cost of serving students who cannot pay even the assessment fee depend on philanthropy, and philanthropy is cyclical and competitive. A donor retrenchment would not, on the founder’s account, threaten the institution’s existence, but it would constrict the access mission that is the model’s distinctive feature. The two-tier structure thus contains a built-in tension: the part of the model that resolves cost is financially independent, while the part that resolves access is financially dependent—an asymmetry that bears directly on the access analysis in White Paper 4.
7.3 Volunteer-supply risk
The labor model assumes a durable supply of academics willing to teach without pay. That supply has held and grown, supported by the prestige and meaning of the mission, but it is a donated input subject to the same fatigue and competition as any volunteer resource. Should the supply tighten—because peer institutions build rival programs, because volunteer goodwill thins at scale, or because accreditation pressures push toward paid, credentialed oversight—the institution would face a choice between hiring (which breaks the unit economics) and stretching volunteers thinner (which risks quality). The model’s cost advantage is, at bottom, a bet that free academic labor remains abundant.
7.4 Financial-aid pressure as the mission deepens
The model’s success at reaching the poorest and most displaced learners creates a financial paradox: the more fully it serves those who cannot pay, the larger the share of enrollment that consumes scholarship rather than generating fee surplus. Cross-subsidy works only while paying students outnumber and outweigh aided students sufficiently. If the access mission grows faster than the paying base, the surplus that funds it shrinks, and the institution must either raise fees on payers (the floor again), expand donor funding (the dependence again), or cap the mission. This is the financial expression of the access-versus-completion tension and is among the model’s genuine long-run vulnerabilities.
7.5 Accreditation-driven cost
The move from national to regional accreditation, treated as a quality signal in White Paper 3, is also a cost event. Regional standards expect demonstrated investment in assessment, governance, faculty oversight, and student support—precisely the layers a lean model minimizes. Maintaining regional accreditation therefore exerts steady upward pressure on cost and helps explain the rising fee floor. The model’s legitimacy and its low cost pull in opposite directions at exactly this point, and managing that tension is central to its durability.
8. Limitations: Cost Eliminated Versus Cost Shifted
The paper’s central caution is that low price to the student is not the same as low cost to the system. Much of what the conventional university charges for, UoPeople does not eliminate so much as relocate. The cost of instruction is shifted onto volunteer faculty, who absorb it as donated labor, and onto peers, whose mutual teaching performs work that salaried instructors perform elsewhere. The cost of content is shifted onto the open-resource commons built and maintained by others. The cost of facilities and devices is shifted onto students, who supply their own computers, connectivity, and study space. The cost of scholarships is shifted onto donors. And a large, often invisible cost is shifted onto students’ time and self-direction: a low-touch, asynchronous model asks the learner to supply the structure, motivation, and persistence that a residential institution supplies through its staff and environment.
None of this makes the “tuition-free” claim false; it makes it partial. The honest statement is that UoPeople has achieved a dramatic reduction in cost to the student by redistributing the underlying costs across volunteers, donors, the open-content commons, and the students’ own time—not by making higher education costless to produce. Whether that redistribution is a fair bargain depends on whether the shifted burdens fall on those who can bear them, and on whether the time-and-self-direction cost translates into the attrition that White Paper 5 must examine.
Two further limitations constrain the analysis. First, the institution’s financial transparency is limited relative to Title IV peers; independent commentary has raised questions about historical filings, including a period in which a large share of expenses ran through “academic” and “outside” services and the institution’s relationship to a for-profit affiliate connected to its founder’s earlier ventures (EdTech Chronicle, 2022). These questions do not establish wrongdoing, but they mean that key figures—per-student cost, the precise split between fee and donor revenue—rest more on institutional statement than on audited public detail. Second, the comparison benchmarks the institution cites (a $66,300 average public MBA, a 93-percent saving) are presented in its own marketing and should be read as upper-bound framings rather than neutral measures; the order-of-magnitude advantage survives a conservative restatement, but the specific percentages do not carry independent authority.
9. Conclusion
On the cost dimension defined in White Paper 1, the University of the People scores higher than any conventional reform can, because it does not discount the conventional cost structure but removes its three heaviest components. The campus is gone, licensed content is replaced by open resources, and salaried faculty are replaced by volunteers, automation, and peers. The unit economics that result—high fixed cost, near-zero marginal cost, fee revenue exceeding marginal cost on each payer—give the model genuine scalability and a self-funding core, and they explain how it can serve six figures of students at four-figure per-degree prices.
The qualifications are equally real and define the agenda for the rest of the suite. The fee floor is rising, which suggests that “tuition-free” is a durable brand atop a slowly increasing real charge. The mission tier remains donor-dependent even if the core is not. The labor model bets on the continued abundance of free academic effort. And the low price to students is achieved by shifting cost onto volunteers, donors, and students’ own time rather than by eliminating it. The model resolves cost more convincingly than it resolves anything else—but it resolves cost to the student, not cost to the system, and the difference between those two is where the next papers do their work. White Paper 3 takes up whether the quality bought at this price survives the scale at which it is now delivered.
Notes
¹ “Tuition-free” is used here as the institution uses it: no charge for instruction, materials, or enrollment. It does not mean cost-free to the student, who pays per-course assessment fees, nor cost-free to produce. The paper treats the distinction between price to the student and cost to the system as the analytically decisive one.
² Fee figures are current as of the institution’s published schedules in late 2025 and early 2026 and are subject to change; the catalog states explicitly that fees are set at the institution’s discretion. The historical fee figures used to establish the upward trajectory are drawn from contemporaneous reporting at each date and should be read as approximate.
³ The break-even claim (sustainability at roughly 15,000 students; per-student assessment revenue covering operations) is the founder’s public statement and is not independently audited here. It is reported as a claim because the model’s economic story turns on it, and because independent verification would require financial detail the institution does not publish at IPEDS depth.
⁴ The volunteer count and the size of the paid staff are institutional self-reports and vary across sources and dates; the paper uses them to establish orders of magnitude (thousands of volunteers, dozens of paid administrators) rather than precise headcounts.
⁵ The discussion of historical filings and the for-profit affiliate (EdTech Chronicle, 2022) is included for completeness and to mark the limits of the public financial record. It raises questions of transparency, not findings of misconduct, and is treated accordingly.
References
Bowen, H. R. (1980). The costs of higher education: How much do colleges and universities spend per student and how much should they spend? Jossey-Bass.
Carey, K. (2015). The end of college: Creating the future of learning and the university of everywhere. Riverhead Books.
Christensen, C. M., & Eyring, H. J. (2011). The innovative university: Changing the DNA of higher education from the inside out. Jossey-Bass.
Desrochers, D. M., & Hurlburt, S. (2016). Trends in college spending: 2003–2013. Where does the money come from? Where does it go? What does it buy? Delta Cost Project, American Institutes for Research. https://www.air.org/resource/trends-college-spending-2003-2013
EdTech Chronicle. (2022). Reason for skepticism about University of the People. https://edtechchronicle.com/universityofthepeople/
EDUCAUSE Review. (2012). Technology and the broken higher education cost model: Insights from the Delta Cost Project. https://er.educause.edu/articles/2012/9/technology-and-the-broken-higher-education-cost-model-insights-from-the-delta-cost-project
FairPlanet. (2020, May 29). University of the People, changing the world, tuition free [Interview with Shai Reshef]. https://www.fairplanet.org/story/university-of-the-people-changing-the-world-one-student-at-a-time/
Financial Times. (2014, December 14). Online education: University of the People gains accreditation.
Financial Times. (2016, March 8). University of the People launches ultra low-cost MBA.
Higher Education Strategy Associates. (2025, May 2). Lean, global, and tuition-free: The University of the People model [Audio podcast episode]. https://higheredstrategy.com/lean-global-and-tuition-free-the-university-of-the-people-model/
Kritica Economica. (2020, June 26). A university for everyone? Interview with Mr. Shai Reshef. https://www.kriticaeconomica.com/a-university-for-everyone-interview-with-mr-shai-reshef/
PBS NewsHour. (2025, May 28). University of the People offers students a new and affordable college experience. https://www.pbs.org/newshour/show/university-of-the-people-offers-students-a-new-and-affordable-college-experience
Study International. (2026, January 7). Is UoPeople free? The answer is not so simple. https://studyinternational.com/news/uopeople-free/
University of the People. (2025). Tuition-free model, processing fees, and program cost pages. https://www.uopeople.edu/tuition-free/
University of the People. (2025). Graduate catalog: Processing fees, scholarships, and financial assistance. https://catalog.uopeople.edu/graduate-catalog-t1/processing-fees-scholarships-and-financial-assistance/
