The Academic Oligopoly: How Five Publishers Drain Billions from Science

Getting your Trinity Audio player ready...

Scientific publishing operates through an extraordinary economic arrangement: governments and universities finance research, academics write the papers, other academics review them—usually without payment—and commercial publishers sell access to the resulting knowledge back to the same institutions.

The dominant companies provide real services, including editorial administration, digital infrastructure, archiving, indexing, typesetting, legal support, and research-integrity screening. The central problem is not that publishing has no cost. It is that control over prestigious journals allows a small group of companies to charge prices that are weakly connected to the marginal cost of distributing digital articles.

The academic publishing oligopoly converts publicly financed knowledge and largely uncompensated academic labor into proprietary products sold through subscriptions, article-processing charges, analytics, and licensing agreements.

Decentralized science, or DeSci, offers a possible alternative: open research records, transparent evaluation, distributed funding, and automated verification that assesses scientific work directly rather than relying primarily on journal brands.

Who Are the Five Major Academic Publishers?

The five companies most commonly associated with the concentration of commercial academic publishing are:

1. Elsevier, part of RELX
2. Springer Nature
3. Wiley
4. Taylor & Francis, part of Informa
5. SAGE Publications

Other important publishers include Wolters Kluwer, the American Chemical Society, IEEE, Oxford University Press, Cambridge University Press, MDPI, Frontiers, and numerous scholarly societies.

The phrase “five publishers” therefore describes a concentration of market power, particularly among prestigious and highly indexed journals. It should not be interpreted as meaning that five companies publish every—or necessarily most—scholarly articles worldwide.

A landmark 2015 study examining approximately 45 million papers found that five major commercial publishers accounted for more than half of papers published in 2013 within its Web of Science-based dataset. However, newer research using broader databases such as OpenAlex and Dimensions has found a much larger and growing “long tail” of independent publishers. The apparent degree of concentration changes substantially depending on which journals and regions a database includes.

The accurate conclusion is not that five publishers own all science. It is that they retain disproportionate control over many of the journals, reputation signals, and distribution systems that determine academic careers.

The Academic Publishing Business Model

The conventional publishing cycle contains several distinct transfers of value.

1. The Public Finances the Research

Governments, charities, universities, and research institutions fund laboratories, researcher salaries, equipment, software, travel, and data collection.

The publisher generally does not finance the underlying experiment or mathematical discovery. It receives a manuscript after much of the expensive intellectual work has already been completed.

2. Researchers Supply the Manuscripts

Researchers normally submit their articles without receiving royalties or publication payments.

In many fields, authors may instead be required to pay an article-processing charge, or APC, to make the final article openly accessible. The direction of payment is therefore the opposite of ordinary commercial authorship: the creator may pay the distributor.

Researchers accept this arrangement because publication in recognized journals affects:

• hiring;
• promotion and tenure;
• grant applications;
• institutional rankings;
• professional reputation;
• eligibility for awards and leadership positions.

A prestigious journal is not merely a publication venue. It functions as a privately controlled academic credential.

3. Academics Perform Peer Review

Publishers send submitted papers to specialists who examine the methods, arguments, evidence, novelty, and presentation.

Most reviewers receive no direct payment from the publisher. Their universities or research organizations frequently pay their salaries while the reviewing work is performed.

Peer review is therefore not literally costless to society. Its cost is commonly absorbed by universities, public funders, researchers, and the reviewers’ other responsibilities rather than appearing on the publisher’s payroll.

4. Editors Make Decisions

Some professional editors are salaried, particularly at major journals. However, many academic editors and editorial-board members also contribute substantial labor for little or no compensation.

The publisher coordinates this system, manages submissions, maintains the journal, performs production work, and distributes the final product. These services have value—but much of the scientific creation and evaluation remains externalized.

5. Universities Buy the Research Back

University libraries purchase subscriptions, database access, journal bundles, or institution-wide publishing agreements.

Consequently, a university may participate at several points:

• it pays researchers to conduct the study;
• it provides the laboratories and equipment;
• it pays scholars who write and review the paper;
• it may pay an APC;
• it purchases access to the resulting publication.

This is why the conventional model is sometimes described as a system in which universities repeatedly pay for the same research.

How Much Money Does Academic Publishing Generate?

The finances of major publishers demonstrate that scholarly communication is not a marginal business.

RELX reported that its Scientific, Technical & Medical division—commonly associated with Elsevier—generated approximately £2.7 billion in revenue during 2025. The division represented roughly 28% of RELX revenue and 31% of adjusted operating profit. RELX’s other businesses include legal, risk, and exhibitions operations, so the group’s total financial results should not be presented as if they came entirely from academic journals.

Springer Nature reported €1.93 billion in total revenue for 2025. Its Research division produced €486.4 million in adjusted operating profit, according to the company’s published results. Springer Nature also reported working with more than one million peer reviewers and over 180,000 editors, illustrating the enormous academic labor network surrounding commercial publication.

Wiley reported Research revenue exceeding $1 billion in its 2025 financial year. In the fourth quarter of fiscal 2026, Wiley’s Research division reported $296 million in revenue and an adjusted EBITDA margin of 37.7%, supported partly by growth in open-access publishing and AI licensing.

These figures do not prove that every journal contract is exploitative. They do demonstrate that scientific publishing, research analytics, and related services collectively generate billions of dollars and substantial operating profits.

Why Universities Cannot Simply Cancel Every Subscription

The obvious response would be for universities to refuse commercial contracts. In practice, several structural barriers protect the publishers.

Prestige Is Not Interchangeable

A library can replace one chemistry textbook with another. A researcher cannot necessarily replace a paper in a prestigious journal with a paper on an unknown website.

Journal brands influence how committees interpret research. Publishers therefore sell not only access but also certification, visibility, and status.

Every Article Is a Miniature Monopoly

Ordinary competing products are substitutable. A researcher who needs a particular article cannot substitute an unrelated article from a cheaper publisher.

Each publication is unique. Once an important paper is placed behind a paywall, access to that specific paper has no direct replacement.

Large Bundles Obscure Individual Prices

Publishers often negotiate packages containing hundreds or thousands of journals. These agreements can make high-demand journals contingent on purchasing less-used titles.

Confidentiality clauses and institution-specific negotiations may make direct price comparisons difficult.

Researchers and Libraries Face Different Incentives

Researchers want access to every relevant article and want their own papers placed in respected journals. Libraries must control expenditure.

A publisher negotiates with the library, but the consequences of cancellation are felt immediately by researchers. This division weakens the university’s bargaining position.

Did Open Access Solve the Problem?

Open access has improved the availability of scientific knowledge, but it has not automatically eliminated concentrated market power.

Under subscription publishing, the reader or library pays. Under many forms of gold open access, the author, university, or funder pays an APC.

The paywall disappears, but the publisher may retain control over:

• journal prestige;
• editorial selection;
• publication infrastructure;
• citation data;
• manuscript workflows;
• institutional contracts;
• research analytics.

This can replace pay to read with pay to publish.

Researchers at wealthy institutions may obtain APC support, while independent researchers and scholars from poorly funded institutions face exclusion. Waiver systems help, but they leave access dependent on publisher policies.

A more equitable model is diamond open access, in which neither readers nor authors pay fees. UNESCO describes diamond open access as shared infrastructure through which scholarly knowledge is treated as a digital public good.

Public policy is also moving toward immediate access. Since July 1, 2025, the US National Institutes of Health has required accepted manuscripts covered by its public-access policy to be made available in PubMed Central without an embargo on the official publication date.

These reforms reduce access barriers, but open access alone does not solve the deeper problem: academia still uses journal placement as a proxy for scientific quality.

The Real Product Is Institutional Trust

Publishers do not merely host PDF files. They package trust.

When a hiring committee sees a familiar journal title, it may assume that:

• experts examined the paper;
• basic integrity checks were performed;
• the result has some degree of importance;
• the authors followed disciplinary standards;
• the article will remain discoverable and citable.

These assumptions are imperfect. Peer review can miss errors, manipulation, irreproducible results, weak statistics, fabricated data, or incorrect proofs. Nevertheless, journals remain useful because institutions lack a universally accepted alternative trust system.

Replacing commercial publishing therefore requires more than placing manuscripts on a blockchain or repository. A credible replacement must reproduce—and improve—the functions of validation, attribution, discovery, preservation, correction, and reputation.

How DeSci Changes the Architecture of Publishing

Decentralized science applies decentralized networks, open protocols, programmable funding, and transparent governance to scientific activity.

A mature DeSci publishing system could separate functions that conventional journals bundle together.

Open Publication

A manuscript, dataset, code repository, proof, experimental protocol, and version history could be published openly from the beginning.

No company would need exclusive ownership of the article merely to coordinate evaluation.

Independent Verification

Review would become a distinct, attributable research output rather than an invisible service performed inside a journal’s private workflow.

Reviewers could publish signed, pseudonymous, or selectively disclosed assessments. Crossref already supports registering peer-review metadata and linking reviews to the works they evaluate, demonstrating that reviews can exist as identifiable scholarly objects.

Transparent Reputation

Instead of reducing quality to a journal title, an open system could record:

• who reviewed the work;
• which claims were checked;
• whether data and code were available;
• whether independent teams reproduced the result;
• whether errors were corrected;
• how later work depended on the result;
• which reviewers made reliable assessments over time.

Reputation would attach more directly to contributions rather than to publication brands.

Programmable Rewards

A decentralized protocol could compensate reviewers, reproducibility auditors, software maintainers, data curators, and researchers whose work later proves useful.

This is especially relevant to retroactive funding of scientific public goods: payment can be based partly on demonstrated output and subsequent value rather than solely on predictions made before the research exists.

Open Governance

Rules for ranking, moderation, conflicts of interest, appeals, and funding could be public and auditable.

This does not guarantee fairness. Token concentration, coordinated voting, identity fraud, and governance capture can reproduce old hierarchies in new forms. DeSci succeeds only when these risks are addressed explicitly.

Automated Verification as the Antidote to Journal Prestige

The strongest alternative to the publishing oligopoly is not merely decentralized storage. It is a verification system capable of analyzing the research itself.

Automated verification could perform several layers of screening.

Text and Citation Analysis

Systems can flag:

• suspicious textual overlap;
• unverifiable or irrelevant citations;
• citation rings;
• inconsistent references;
• retracted sources;
• unsupported factual claims.

Crossref’s Similarity Check already provides manuscript-comparison tools for detecting potential plagiarism.

Image and Data Integrity

Automated tools can search for duplicated, manipulated, reused, or synthetically generated images.

The International Association of Scientific, Technical and Medical Publishers maintains resources for automated image-integrity analysis, showing that machine-assisted screening is becoming part of mainstream research-integrity infrastructure.

Statistical Verification

Software can check:

• whether reported statistics are internally consistent;
• whether sample sizes match tables and figures;
• whether the stated method supports the conclusion;
• whether multiple-testing corrections were applied;
• whether data distributions contain suspicious patterns;
• whether the supplied code reproduces the reported output.

Software Reproducibility

For computational research, an automated environment can attempt to:

1. install the declared dependencies;
2. execute the submitted code;
3. regenerate tables and figures;
4. compare outputs with the paper;
5. record failures and environment details.

This turns reproducibility from an informal request into an executable test.

Formal Proof Checking

In mathematics and theoretical computer science, formal proof assistants can verify proofs expressed in machine-checkable languages.

Most mathematical research is not yet formalized, and automatic formalization remains difficult. Nevertheless, proof checking demonstrates an important principle: some parts of scientific validation can be made objective, repeatable, and independent of institutional prestige.

Dependency-Aware Evaluation

Research exists in a network. One theorem may support hundreds of later results; one software library may enable thousands of experiments.

An automated system can map citations, datasets, software dependencies, replications, corrections, and derivative discoveries. Funding can then recognize enabling work that conventional journal metrics overlook.

This is central to the proposed AI Internet Meritocracy model: scientific value should be evaluated from evidence, dependencies, reproducibility, and usefulness—not merely from the logo printed above the article.

Automated Verification Cannot Replace Scientific Judgment

Automation is an instrument, not an infallible referee.

An algorithm can determine whether code executes, but not necessarily whether the research question matters. It can flag unusual statistics, but unusual data are not automatically fraudulent. It can compare a manuscript with known literature, but genuinely novel work may initially appear isolated or implausible.

Automated systems also introduce new risks:

• biased training data;
• preference for conventional methodologies;
• manipulation of scoring criteria;
• false accusations;
• opacity in proprietary models;
• coordinated identity attacks;
• excessive dependence on measurable indicators;
• discrimination against work that cannot be easily formalized.

The appropriate model is therefore machine-assisted, transparent, contestable verification.

Automated checks should produce evidence and structured warnings. Human experts should interpret difficult cases, and authors should be able to inspect and challenge the system’s conclusions. The design must also address AI alignment and resistance to metric gaming.

What a Post-Journal Scientific System Could Look Like

A credible alternative could follow this sequence:

1. A researcher publishes a timestamped manuscript with its data, code, methods, funding disclosures, and version history.
2. Automated systems perform plagiarism, citation, statistical, image, code, and metadata checks.
3. Relevant specialists receive review requests based on demonstrated expertise rather than institutional title alone.
4. Reviews are recorded as citable contributions.
5. Authors publish revisions while retaining a complete history.
6. Replication teams report whether the result can be reproduced.
7. Evaluation models update the work’s confidence level as new evidence appears.
8. Funding is distributed to authors, reviewers, replicators, and infrastructure maintainers according to transparent rules.
9. Appeals and disputed judgments are handled through auditable governance.
10. Libraries finance shared infrastructure rather than repeatedly purchasing access to isolated journal collections.

The result would not be “science without peer review.” It would be peer review without exclusive dependence on journals.

Why the Transition Is Difficult

The technical components already exist in partial form: repositories, persistent identifiers, open metadata, executable notebooks, formal proof systems, plagiarism detection, image analysis, version control, decentralized ledgers, and online review platforms.

The harder challenge is institutional coordination.

Universities continue to use journal brands because hiring and grant committees need quick signals. Researchers continue to submit to prestigious journals because their careers depend on those committees. Libraries continue to subscribe because researchers need access.

This is a coordination trap. No participant can easily leave while the others remain.

DeSci must therefore produce a trust signal strong enough for universities, governments, philanthropists, and researchers to recognize. It cannot win merely by being more open. It must become more useful, more reliable, and more informative than the journal title it replaces.

A Better Economic Settlement for Science

Scientific publishing requires resources. Servers, preservation, editing, metadata, governance, fraud detection, and skilled review are not free.

The relevant question is who should control and finance these functions.

Under the conventional model, institutions repeatedly transfer money to publishers that control prestigious publication channels. Under a public-infrastructure model, the same institutions could finance interoperable repositories, open verification protocols, reviewer compensation, and reproducibility work.

Money would then support scientific functions directly:

• publishing;
• checking;
• preserving;
• correcting;
• reproducing;
• maintaining;
• discovering.

This would be more efficient than paying primarily for restricted access and inherited prestige.

Conclusion: Replace the Journal Brand with Verifiable Evidence

The academic publishing oligopoly persists because it controls more than articles. It controls certification, visibility, and career signals.

Paywalls are only the visible part of the problem. Replacing subscriptions with high publication charges does not necessarily change the underlying structure. Nor does placing unverified manuscripts on a blockchain.

The deeper solution is to separate scientific communication from proprietary prestige:

• publish research openly;
• expose the evidence and methods;
• make reviews attributable;
• automate checks that machines can perform reliably;
• preserve human judgment where interpretation is necessary;
• record corrections and replications;
• reward useful contributions directly;
• govern the infrastructure transparently.

The ultimate antidote to academic publishing oligopoly is not free access alone. It is an open scientific system in which trust follows verifiable evidence rather than the market power of a journal.

That is the potential of DeSci and automated verification: not the destruction of peer review, but its transformation from a closed publishing service into an open, auditable scientific public good.

Scientific knowledge should not be financed by the public, evaluated by the public, and then enclosed for resale to the public. It should remain part of the shared infrastructure on which future discovery depends.

Learn how World Science DAO proposes to finance independent scientists and open scientific infrastructure, or support the development of transparent alternatives to conventional research funding and publishing.