AASM – Springer Nature Partnership
Congratulations to the American Academy of Sleep Medicine for their new publishing agreement with Springer Nature. AASM’s flagship The Journal of Clinical Sleep Medicine will move from self-publishing starting January 2026. C&E was honored to help AASM run a competitive RFP process and broker the agreement with Springer Nature.
Consolidation
David Crotty’s new piece in The Scholarly Kitchen looks at ongoing consolidation in the scholarly journals market. Spoiler alert: the largest publishers get larger. 57% of the articles indexed in the Web of Science are published by just 5 publishers – up from 55% a year ago.
Resilience
1
Given the US administration’s growing penchant for tariffs and its drastic cuts to research grants, combined with its privileging of AI companies over content owners, one would expect that predominately European publishing companies focused on research content (and selling to academia) would be just about the worst business to be in for the first half of 2025. It turns out to be quite the contrary!
Springer Nature reported strong H1 results, notching 6% underlying revenue growth across the portfolio. The press release notes that:
Research, the company’s largest segment, reported revenue of €731 million. The segment contributed most significantly to operating performance, with underlying revenue growth of 7%. This was driven by the Journals portfolio and the continued momentum in open access (OA) globally. The number of published articles rose by approximately 10% across the whole portfolio and around 25% in Full Open Access (FOA) journals.
It is perhaps worth noting that, as James Butcher has pointed out, Springer Nature’s Scientific Reports journal has already published more articles in 2025 than it did in all of 2024, and is on pace for a record 45,000 articles this year, easily shattering PLOS One’s record for the most articles published by a single journal in a calendar year (31,319 articles in 2013).
For people who want to dig deeper into this data, here is the accompanying report, investor presentation, and the transcript from the investor call.
Wolters Kluwer similarly posted 5% organic growth, although, unlike Springer Nature, the division in which their journals are located (Health) posted a more modest 2% increase in revenues.
Health Learning, Research & Practice (42% of divisional revenues) recorded 1% organic revenue growth (HY 2024: 4%). Excluding print, organic growth would have been 5% (HY 2024: 5%). As expected, our medical research unit recorded slower organic growth against a prior period which had benefitted from reaching full-scale distribution for the New England Journal of Medicine.
That paragraph caught our eye for a couple of reasons. First, it is notable that slowing (because they have reached market saturation) institutional subscription sales of a single title might rise to the level of a note in an H1 investor report of a €6 billion company (even when that journal is the New England Journal of Medicine). Second, we don’t know what to make of the comment that “excluding print, organic growth would have been 5%” (instead of only 2%). Wolters Kluwer mostly publishes clinical medical journals. A lot of those journals continue to have strong print advertising revenues. If they excluded print subscriptions, presumably they would exclude all revenue associated with print subscriptions, including print advertising. Additionally, we would assume that most, if not all, print subscribers also have online access and hence are additive to online subscriptions. In a hypothetical situation where they excluded print, revenue would surely be reduced, as print subscribers would likely already have digital and the added revenue from print would just go away. Presumably they just mean that digital revenues are growing faster than print subscription revenues (which are likely declining).
When the strong H1 results (and full-year guidance) of Wolters Kluwer and Springer Nature are added to those of Elsevier and Taylor & Francis (reported on in last month’s issue of The Brief), it paints a remarkably cheerful economic picture for an industry that seems beset by so many headwinds. (Oxford University Press’s H1 results, which we also mentioned last month, were somewhat less heartening.) That said, tariffs are not currently affecting scholarly literature. Budget cuts at federal agencies will most immediately affect the OA revenues of publishers. However, those cuts have not yet had a significant impact on publication output (which obviously lags the research itself); and the impact is likely to be mitigated by implementation of the Nelson Memo’s zero-embargo policies. These will result in pushing more content into paid-OA publication channels.
Even the looming specter of caps on article processing charges (APCs) might pose less of a threat to publisher revenues than initially thought. Such caps are likely to impact only a small number of papers published in high-profile journals and will be mitigated in many cases by transformative agreements (where institutional, not agency, funds are used to pay for OA publication fees). Moreover, the percentage of US federally funded papers published by most publishers is small (see figure below). With the exceptions of the American Chemical Society (ACS) and Oxford University Press, federally funded research accounts for less than 10% of the annual published output of the largest publishers.
While the outlook remains sunny for scientific and scholarly publishing this year, rougher weather may be ahead. US institutional subscription renewals pose perhaps the greatest vulnerability for publishers. There are a lot of universities in the US, and their budgets will be impacted by research budget cuts and cuts to overheads sought by the current administration on federal grants. This is alluded to in comments in Springer Nature’s investor call by the publisher’s CEO, Frank Vrancken Peeters. Peeters, in response to a question from industry analyst Sami Kassab, noted that 2025 subscription renewals have been largely wrapped up (by which we assume he meant renewal orders placed in 2025 for the 2026 subscription year) but that the picture for 2026 renewals (meaning, we assume again, the 2027 subscription year) are less clear:
We are a resilient business — with about 60% of contracted revenues in research. And as we just mentioned, you know, more or less all our renewals have been done in line with expectation. And I think the good news also is, that our 2025 renewal season which we’ve just completed, actually, have [sic] quite a lot of active renewals in the US. And we expect that amount of renewals in 2026 to be quite significantly smaller in the US. I think on top of that, keep in mind that — let’s say — publication costs and also library university budgets typically are less than 1% of total spend either of the research and development funds or academic spending. And what we have seen is that the relationship between whether US universities get more or less funding doesn’t necessarily immediately result — unfortunately — in more money for library budget. But at the same time, also necessarily an equal drop in the — let’s say — funding for libraries. (Springer Nature investor call transcript)
Peeters is correct in that library budgets are a small portion of overall university spending and that they do not move in lockstep with the overall budget. Nonetheless, increased university budget pressures will likely be felt in libraries and reflected in their collection decisions. Most of the larger publishers have multiyear contracts, insulating them somewhat in the 2027 subscription cycle. Nonetheless, we anticipate the next few renewal cycles to test resilience at many publishing houses.
Briefly Noted
McGraw Hill raised $415 million in its recent IPO, valuing the publisher at $3.25 billion. While now publicly traded, Platinum Equity continues to hold close to 85% of total outstanding shares according to Reuters.
A Proceedings of the National Academy of Sciences USA (PNAS) paper concludes that the number of fake papers from paper mills has been doubling every 1.5 years. While worrisome, this still only represents a small proportion (less than 1% by one estimate) of the literature, but coverage of the article in The New York Timesand The Economist will only further fuel the decrease in public trust of scientific research. Interesting reactions to the paper came from PLOS (the subject of the study), Nature (whose reporters quickly identified the study’s anonymized retraction-prone editors at PLOS), and McGill University biologist Michael Hendricks, who suggests that “Paper mill / LLM [large language model] sludge in predatory/fake/dump journals that are meant to game metrics” go largely unread by humans, and thus pose less of a threat to scientific progress than the much smaller number of papers from “people who fabricate data or evidence with the express purpose of influencing their field.”
A new Executive Order from the US White House not only puts political appointees (rather than scientists) into the deciding role for who receives grant awards, it also allows for grants to be de-funded at any point after the award is made. As Carl Bergstrom points out, this runs counter to the time scales at which science works. With the government able to pull funding at any time with no cause, he writes, “This makes it impossible to engage in anything other than the shortest-term, most incremental research. … Far-reaching, innovative, creative, world-changing research will no longer take place in the United States.” Further, since PhD training often takes around five years, the instability of grant funding makes training impossible, as support for graduate students cannot be guaranteed all the way through to their degrees.
In perhaps the most direct (attempted) intervention yet between the current US administration and a scholarly journal, US Health Secretary Robert F. Kennedy Jr. called for a retraction of a large-scale, long-term study that concluded that exposure to aluminum salts in vaccines did not offer increased health risks. The journal, Annals of Internal Medicine, has declined Kennedy’s request and will not retract the article.
The American College of Obstetricians and Gynecologists (ACOG) has announced that it will no longer accept federal funds. “Recent changes in federal funding laws and regulations significantly impact ACOG’s program goals, policy positions, and ability to provide timely and evidence-based guidance and recommendations for care.” ACOG received around $990,000 in funding from the US Department of Health and Human Services this year.
While the US Senate Appropriations Committee has pushed back on cutting grant indirect costs in its draft appropriations bill for the National Institutes of Health (NIH), language is included commending the NIH for seeking to limit APCs, and directs the NIH to “work with the scientific community to inform development of an APC allowable charge limit and, as part of that process, to: account for different publishing models, particularly U.S. based publishers focused on rigorous peer review and quality checks; support a robust American scientific research and publishing enterprise amid unprecedented global competition” and “[establish] parameters to guard against potential abuses, including payment of APCs to journals that don’t prioritize research quality and integrity.”
Unpaywall announced a major update to their classification data, which is used by most of the major bibliometric databases in the market for determining whether articles are OA, and, if so, to which type of OA they belong. The announcement states that various changes have been implemented in the data provided, including the addition of more than 18,000 fully OA journals from sources such as Open Journal Systems (OJS), Scientific Electronic Library Online (SciELO), and Japan Science Technology Information Aggregator, Electronic (J-STAGE). Our (very) preliminary look at the new numbers suggests that while the overall trends seen in the market hold true, the quantities and percentages of Gold OA articles now appear significantly higher, Green OA slightly higher, and everything else (Bronze OA, Hybrid OA, and Subscription articles) lower. With a reclassification (away from Hybrid/Gold OA) of freely available Elsevier articles under an Elsevier User License rather than Creative Commons licenses, Springer Nature is now the largest OA publisher cumulatively, whereas previous data held Elsevier in this position.
Brigham Young University librarian Rick Anderson offers a view on why he sees the Subscribe-to-Open (S2O) model for OA publishing as “doomed.” Rick is pessimistic and suggests that most universities will not be willing to convert budgets allocated for “access to content” to voluntary spending that supports broad open models, particularly when universities have “far more needs to meet than they have resources available.” Lauren Kane, CEO at BioOne, offers a more optimistic counterpoint, noting that S2O subscriptions are “not voluntary payments” and that most include “exclusive, subscriber-only benefits” that incentivize paid subscriptions. Meanwhile the Royal Society announced plans to shift all of its journals to OA via the S2O model in 2026.
A new report from ITHAKA S+R looks at “The Current State of Academic E-Book Business Models,” and while there’s much of interest in the report, the main takeaway seems to be that a lot of the models in place are not working particularly well.
The lack of nomenclature standards for OA continues to confound, as a new publisher has announced itself as operating “exclusively under a platinum open access publishing model” despite being based on book publishing charges.
“Websites in general will evolve to serve primarily as data sources for bots that feed LLMs, rather than destinations for consumers.” So says MailChimp Director Ellen Mamedov, as the company, along with many others, begins to adapt its websites to serve AI bots, rather than readers. While LLM searches don’t yet drive sales and only account for a tiny proportion of traffic, their rapid growth presages the rise of zero-click search. A recent survey suggests that 80% of consumers already resolve 40% of their search queries without clicking a single link.
Are peer reviewers really more likely to approve articles that cite their own work? Maybe, but with an awful lot of caveats. The preprinted study in question suggests that this is the case, but the study group used was an atypical set of journals, all using the F1000 platform, where peer review is radically different from that used in most publications. On these platforms, peer reviewers self-select and make decisions to approve or reject articles, rather than offering advice to journal editors who make those decisions. Even when it’s not being done as an unethical way to increase the number of citations to their own papers, it’s probably not surprising to find that people who spend their time doing these sorts of reviews often favor papers that cite their own work.
Bad news out of India for Sci-Hub: In the (very) slow-moving case brought against the illicit network by publishers, the Delhi High Court has issued an order requiring authorities to block access to Sci-Hub, LibGen, and other related websites by users in India. The order is due to a finding that Alexandra Elbakyan, founder of Sci-Hub, violated the court’s order not to upload any new content to the sites, which, according to one analyst, was seen “as wilful disobedience, qualifying as contempt under Indian law. Her counsel even requested discharge from the matter, citing lack of communication from Elbakyan, which the court interpreted as a deliberate refusal to engage with legal proceedings.” The next hearing in the case is scheduled for December 1.
With funding agencies taking measures to tame an overwhelming flood of AI-generated grant applications and Wikipedia facing an “AI-slop” onslaught, preprint servers have now become the latest victims of the overzealous use of LLMs to generate massive amounts of scientific-sounding but largely inconsequential text. Mark Hahnel from Digital Science suggests there are two types of researchers creating this spam — “true believers who genuinely think they’re onto something revolutionary,” and cynics who are playing a quantity game to satisfy publish-or-perish requirements. As Hahnel notes, “this approach fundamentally misunderstands both how scientific breakthroughs happen and what LLMs actually do.”
Reddit announced that it will block access by the Internet Archive’s Wayback Machine because AI companies are using it to circumvent their blocks on content scraping.
Cory Doctorow offers an extended essay on Goodhart’s Law (“When a measure becomes a target, it ceases to be a good measure”), how that affected Google search rankings and evaluation of students in the US, and how it is similarly likely to impact the use of AI in education. Speaking of measurements becoming targets, Nature reports that ResearchGate has an increasing credibility problem as users are gaming its metrics.
Given Google’s long history of terminating projects that fail to earn significant profits, the long-term future of Google Scholar remains in doubt.
Competition is heating up in the “point-of-care tools for physicians” market, as Doximity has acquired Pathway Medical.
Can someone please license the works of Edward Tufte for AI training? The chart junk examples in this Washington Post article on the visual crimes of the AI industry are appalling.
***
LLMs are essentially very sophisticated autocomplete systems. They predict what words should come next based on patterns in their training data. When you ask them about theoretical physics, they’re not reasoning from first principles or accessing some hidden understanding of the universe, they’re pattern-matching against every physics paper, textbook, and Wikipedia article they’ve seen. It’s impressive mimicry, but it’s still mimicry. — Mark Hahnel