Open Library of Humanities

THE LONG READ Infrastructure Landlords: The Rentier Capitalism of Commercial Academic Publishers

Posted by Posted by OLH on 26 March 2026

THE LONG READ

Infrastructure Landlords: The Rentier Capitalism of Commercial Academic Publishers

This piece has been written by a friend of the OLH who wishes to remain anonymous.

Illustrated pyramid of capitalist society, with laborers at the bottom supporting ruling classes.

Figure. Pyramid of the Capitalist System (1911), attributed to Nedeljkovich, Brashich, and Kuharich. Originally published in The Industrial Worker (Industrial Workers of the World). Public domain.

If you want to understand where the commercial parts of scholarly communications may be heading, you need to look beyond policy documents, conference panels, or public-facing strategy statements. You should look at what large commercial actors say when speaking to investors. Earnings calls are one of the places where that language becomes especially revealing: less concerned with sector ideals than with growth, market opportunity, competitive position, and what will ultimately generate value for shareholders. For this reason, it can be worthwhile to review earnings calls and investor presentations, as these are often overlooked when discussing OA policy and sectoral movements.

Before getting into the company-by-company detail, it is important to understand what an earnings call is. These calls are part presentation and part Q&A session, where a company’s senior management discusses its recent financial results with analysts. They are usually accompanied by a presentation and a results statement (found on the company website). They tend to cover headline figures such as revenue, profit, costs, debt, and outlook, but they also reveal something more useful: what management thinks is driving performance, where it sees growth, what problems it is trying to solve, and which markets or products it considers strategically important.

This is why they can be worth reading, even for people who are not investors; they offer a direct view into how major commercial players describe the academic and research market when speaking to the financial community. These calls demonstrate where companies believe they have pricing power and which technologies or services they want to expand (or deprecate). How they talk about AI, analytics, licensing, open access, and journal strategy can also be revealing and informs us where they may be seeking tighter customer lock-in or higher-margin recurring revenue. In other words, earnings calls help make visible the commercial logic shaping parts of the scholarly communications system.

At the same time, these calls need to be read critically. Companies usually highlight strengths and downplay weaknesses, and will be motivated to frame results to reassure markets. It is also important to keep in mind that some claims, especially forward-looking ones, are management’s framing (or hoping) rather than independently verified facts. The analysis that follows below treats these calls and presentations as useful evidence of strategy and priorities, but not as neutral truth.

Note: this is not a financial analysis and not financial advice. It is a sector-focused reading of, and commentary upon, company reporting aimed at understanding what these firms say about their performance and direction, and what that may mean for libraries, universities, and scholarly communications more broadly. The author is not a financial expert. 

The comparison below looks at the earnings calls and associated data from five major commercial players in and around scholarly communications: standalone publishers Springer Nature and Wiley; publishing businesses embedded within larger information groups Informa (Taylor & Francis) and RELX (Elsevier/STM); and Clarivate, which sits somewhat differently as a research, analytics, and infrastructure vendor. These are, of course, not perfectly like-for-like businesses, and Clarivate in particular is not directly comparable to the publishers. However, all five shape the environment in which libraries, researchers, and universities operate, whether through reading and publishing content, metrics, workflow tools, licensing, or metadata.

This commentary will use figures and statements taken from earnings calls transcripts and presentations listed below. It first summarises how each company reports its performance, then looks across the group for shared themes, differences, and signs of broader sectoral movement.

Glossary block

  • EBITDA - Earnings before interest, tax, depreciation and amortisation.
    • EBITDA assesses business performance by excluding (uncontrollable/non-operating-related) costs such as taxes and interest.
    • Good EBITDA margin depends on the sector. As a quick and dirty rule of thumb: 15%-20% for more ‘traditional digital publishing’ would be good, as opposed to higher margins (30%-40%) for more technology, data and infrastructure-based companies. This can (and is), however, be debated.
  • AOP - Adjusted operating profit.
    • Profit from operations using the company’s adjusted measure, which removes some of the one-off or non-core items).

All numbers here are in the original currency. They have not been converted to avoid accuracy issues arising from currency fluctuations.

   Springer   Nature*  Informa  RELX  Wiley  Clarivate
 Adjusted   EBITDA  €714.1M (+5%)  $87.3M (+11%)  £3,846M (+3%)  $105M (+12%)   $1,002M (-5%)
 Adjusted   EBITDA   margin  37%  17.9%  40.1%  25.7%  40.8%
 Revenue  €1,926.4M (+6%)  $486.8M (-0.7%)  £9,590M (+7%)  $410M (+0%)  $2,455M (-4%)

The following are numbers from relevant divisions specifically - where text is in quotation marks (“ ”) the number was taken from a description or transcript (with no written equivalent in a presentation or report that could be found) and may not be exact.

   Springer Nature (research)  Informa (Taylor & Francis)  RELX(STM)  Wiley (research)
 Revenue  €1,517M (+7%)  £671M (+4%*)  £2,714M (+5%)  $274M (+1%)
 AOP  €486M (+10%)  £245.7M (-3%)  £1,035M (+7%)  -
 Subscription renewal rate  “Near 100%”  “Strong subscription renewals”  -  82% so far, 99% customer retention expected.
 Submission growth  “>30%”  “>20%”  “>20%”  +26%
 Article output  12%  “145K articles published on T&F online”  “>10%”  +11%
 OA output  c. 25%  “double-digit growth in Open Research volumes”  -  +24%

*Excluding non-recurring data contracts. If these were included, it would be -2%.

By nature, these calls go beyond simply presenting statistics and figures. They must also address where this reported growth comes from, what areas are their profit drivers and their strategies for future development.

Springer’s main strategies are to scale OA and expand AI across editorial and publishing workflows. Notably, they employ multiple AI-driven or AI-enhanced systems, many of which aim to keep either submissions or users within the Springer ecosystem. Springer executives also stated that transformative agreements typically bring in higher revenue than legacy subscriptions; “we typically see that TAs are performing better than subscription revenues. So subscription revenues in Springer portfolio was typically 1 to 2% because that's library budget growth. And TAs tend to grow 1 to 2% better than that.” (Springer-Nature call transcript, p.13)

Taylor & Francis employs a blended model of subscriptions, open research, archives, and data licensing, and increasingly frames content as AI-ready knowledge infrastructure. Their interests, in line with wider sector trends, appear to lie in monetising data and specialist knowledge beyond the traditional library package. Similar to RELX and Springer, Taylor & Francis indicates that over half of its income comes from subscriptions, rather than transactional sales.

RELX’s STM growth is driven by growth in publication volume; they state that 80% of its revenue comes from subscriptions. Its main strategic theme is Elsevier’s combination of journals, data, knowledge graphs, analytics, and workflow systems for universities, funders, governments, and corporate R&D. It is building a layered research infrastructure business around content, metadata, analytics, and decision support.

Wiley is making AI central to its story: it highlighted content licensing deals, AI Gateway for researchers and institutions, corporate subscription customers, and a five-year OpenEvidence agreement using journal and review content. Wiley also expects $70M revenue from its Advanced Journal Portfolio of 27 journals, at about $2.6M per journal (this will be one of its best-performing journal portfolios).

Clarivate, as a data and infrastructure vendor without a publishing division, has slightly different strategies. We see a strategy that prioritises AI-enhanced infrastructure, subscription-led revenue models, and capturing as much as possible within proprietary software. 97% of their A&G revenue comes from proprietary solutions, “providing a durable foundation for innovation” (Clarivate presentation, p.11). They also note that part of their revenue change was due to the cessation of book sales this summer, though they expect this will expand their profit margins in the long term.

Overall, these reports show there is growth in the sector, but it is uneven. Springer Nature and RELX reported strong underlying revenue and profit growth, with research publishing, analytics, and workflow tools doing most of the work. Informa’s wider group showed growth, but Taylor & Francis was softer. Wiley seems to be doing well, with its learning segment being the weakest element. Clarivate is the most mixed of these, but is trying to recover through its focus on subscription-based revenue. As a whole, publishers and information vendors are leaning hard into AI, analytics, subscription, and embedded (research) workflow products, and reducing both print and transactional activity. Whilst this is good for their margins, it is likely to deepen dependence and increase pricing pressures for libraries.

One key theme emerging from these calls is AI as a product and an efficiency tool. Wiley is the most explicit about licensing content into AI products and tools. It cites $42M AI revenue for the year to date (YTD) and mentions that over 80% of its journals have migrated to their ‘research exchange’ platform. This platform reportedly transforms their content into AI-ready data, which is described as “the foundation that makes everything we're doing in gateway, licensing and subscription knowledge feeds possible” (Wiley call transcript, p.4). Whilst Wiley’s research division seems strategically ambitious, they have not yet proven whether AI income can become sustainable, recurring revenue rather than a series of opportunistic licensing deals.

Whilst Taylor & Francis discuss data licensing and describe their content as a licensable machine-readable asset, they state, “we will do deals on the right terms if we can get them. We're not in a rush to do deals on the wrong terms as you'd expect” (Informa call transcript, p.18) indicating they may be expecting a better market for deals to emerge in the future. This is made explicit in the following statement:

“[W]e will continue to look for deals [in the LLM licensing market], [but there are ongoing negotiations] for follow-on deals, either in the space that we have so far monetized or in other areas like journals that we haven't yet monetized. We'll report those to the market as we do them rather than speculate about what we might do in the future, but it's something we're targeting and looking at on the right terms with the right protections in place to try and drive extra revenue” (p. 15)

Clarivate is (unsurprisingly) framing AI as part of its value-creation and solution strategy, focusing on proprietary AI software. AI is treated less as a dramatic new business line than as an enhancer of stickiness and pricing power across embedded institutional products. However, the evidence and examples given in the materials are more strategic than detailed.

Interestingly, when asked about (proprietary) data licensing for LLMs - Claude CoWork being given as a specific example, RELX essentially said no, broad licensing of its core proprietary data is not part of the strategy. They see their content and data assets (proprietary databases and datasets; legal, business, healthcare etc.) as the foundation of the business and want to keep using them inside their own products, including future AI-enabled tools, rather than letting others build on top of them via external integrations. It says small, non-core copyright or licensing deals may happen, but these are minor and not strategically important. This would fit with predictions that general LLMs such as ChatGPT and Claude may be unsustainable in the long term, with a move toward smaller, specialist models instead [1] [2].

In a broader landscape of AI implementation and strategy, this is not surprising. More concerning is the other recurring theme within these calls: deliberate convergence. Each firm, from its own position, is moving toward owning more of the environment in which research is produced, evaluated, and used — not just the content within it.

The story here isn't that publishers profit from research. It is that they are quietly occupying the infrastructure around it. What is being created is a set of increasingly integrated, proprietary systems built around content, data, and workflows, creating a deeper lock-in risk for universities and libraries. And once an organisation relies on a small number of companies - or worse: a single company - for reading and publishing, discovery infrastructure, clinical databases, repository software and more, switching becomes harder, and procurement shifts from single products to ecosystem dependence. Universities are not just buying access to content, but are slowly becoming tenants in commercially governed scholarly infrastructures.

These companies are repositioning themselves as infrastructure landlords. Like landlords, they control access to a space people depend on, set the terms of occupancy, and benefit when leaving is costly. As described before, once a university relies on the same companies for reading and publishing, open-access agreements, discovery infrastructure, metrics, research intelligence and other tools -  procurement stops being about buying discrete content and services, and starts to resemble a tenancy within a privately governed estate. The question stops being what a product costs and becomes whether you can realistically leave.

This observation isn’t new. Posada & Chen make this argument in their 2018 paper “Inequality in Knowledge Production: The Integration of Academic Infrastructure by Big Publishers” [3], where they describe the rent-seeking behaviours of big publishers and critically observe that “publishers are able to co-opt and monetize Open Access to a greatest extent, with the move towards infrastructure and Open Access thus being simultaneous and complementary processes” (p.16). More recently, in Publishing Beyond the Market (2025) [4], Samuel Moore also argues that large publishers increasingly seek to own and monetise the entire knowledge-production life cycle, moving from content providers to service providers within centrally controlled ecosystems. In this model, the goal isn’t to publish (or even own) research outputs, but to fence researchers into “walled gardens”, automate and influence their behaviour, and then generate more data to build further services and deepen lock-in. Roger Schonfeld’s term “supercontinents” (cited by Moore) explains this well: seamless, enclosing environments in which more and more research is brought together under a small number of commercial owners. This seamlessness is not a neutral choice; it is a business strategy.

This shift is accelerated by conditions of financial austerity. Libraries are forced to be pragmatic: budgets are tight, and spending decisions must be justified by financial efficiency, easy to implement without disproportionate staffing investment, and provide demonstrable value to both the institution and individual researchers. Under those conditions, investment in integrated commercial systems can look rational. If one vendor offers smooth workflows, easier reporting and administration, and promises a convenient way to measure impact, it may seem sensible to purchase the service or software - even if this means further consolidation. But what looks efficient in terms of budget can be corrosive to both the organisation and the sector in the long term. Scarcity pushes us toward convenience, and convenience can become the route by which dependency deepens. As the earnings calls themselves state, the integration of workflows and data are deliberate product strategies, designed to generate sustainable, recurring revenue which ‘increases shareholder value’. The convenience purchased by libraries is the dependency that publishers are selling.

This is where the serials crisis meets the infrastructure problem. Moore notes that scholarly publishing has long exhibited monopoly-like behaviour and that journals are not “substitutable” in the way ordinary market goods are (p. 15-16). Libraries cannot simply replace a title with a cheaper equivalent. That logic drove the growth of bundled subscription deals and rising journal expenditure. But the new problem does not replace the old one. It builds on it. Historic control over journals is now being used as the foundation for wider control over discovery, assessment, workflow, analytics, and AI-mediated uses of research. The article remains important, but it is no longer the only strategic asset.

The stakes aren’t purely financial either as these infrastructures aren’t just descriptive; they directly impact researchers at the individual level. Metrics, indexing databases and other data products shape hiring, tenure, and career trajectories (see also: Sarah Lamdan’s Data Cartels (2023), also cited in Moore (2025)). Coupled with the drive to capture more user and research data, this means commercial infrastructures increasingly shape academic behaviour. These systems do not just distribute scholarship. They help determine what becomes visible, legible, and valuable within universities. The issue is not simply cost, but control over the conditions of scholarly life.

This should also sharpen our thinking about open access, as we’re observing in real-time that openness, in its mainstream institutional form, has often been made compatible with infrastructural dependence. The recent dominant strategy has been to (re)negotiate TAs, and whilst these deals have had value in that they have driven both the submission and publication of OA content, they have not resolved the structural issue underlying it. They have deepened it. 

With TAs being negotiated with the same few publishers who were part responsible for creating the problem TAs aim to resolve, they further entrench those publishers as the primary route to compliant open access. Springer-Nature’s executives themselves noted in their earnings call that TAs tend to outperform legacy subscriptions, evidencing that Gold OA is being folded into strategies that promote platform retention and data monetisation. Gold OA has changed the route by which money flows into systems, without considering what systems this money empowers or the interests of those who govern it.

This is what Posada & Chen (2018) identified as well: that publishers can co-opt and monetise open access, with the move towards infrastructure and OA being simultaneous and complementary, rather than competing processes. A paper can be openly available while the surrounding systems of submission, evaluation, discovery, analytics, and licensing become more concentrated and more proprietary. Openness at the level of the output does not translate into openness at the level of the system. The earnings data here confirms that dynamic is not only continuing but accelerating.

This does not mean that open access mandates or even transformative agreements have been worthless. It means they are no longer sufficient. Structural reform requires attending to who owns and governs the infrastructure through which open research flows.

For libraries, the question is what follows from recognising this. One answer is that procurement has to be understood politically, not just administratively or financially. “Can we afford this” is not enough. We need to be asking what dependencies does this purchase create, what forms of governance does it lock in, what alternatives does it foreclose, and who gains leverage from it over time. That logic is already emerging in some library collection strategies which explicitly pair financial resilience with open research, ethical supply, and a commitment to reduce reliance on problematic vendors. Open infrastructure, then, cannot be treated as a luxury for better-funded times. We start to see this in library content and collection strategies (such as those of the University of York and Sheffield libraries), and in RLUK’s newly published shared principles for evaluating and acquiring open-access monograph models. This shows that different models are possible when libraries, consortia and publishers are understood as collaborators in building a more federated, equitable and publicly accountable scholarly communications landscape, rather than as customers and vendors locked into purely transactional relations.

The danger of landlordism is that dependence eventually starts to look like efficiency. Under budget pressure, integrated commercial systems can seem like the most practical response to complexity. But it is clear that the commercial actors already have a destination in mind for where the sector could be heading: toward deeper ownership of the infrastructure around research, not just its outputs. If universities continue to treat that only as a question of procurement efficiency, they will keep losing ground while believing they are acting prudently. The task now is not only to negotiate harder, but to defend the possibility of scholarly infrastructure that is interoperable, collectively governed, and not permanently rented back to the academy.