News & Views: What proportion of OA is Diamond OA? Can AI help us find out?

Dan Pollock and Ann Michael • February 20, 2025

Overview


A recent post on the Open Café listserv posed a question about the true extent of fee-free open access publishing, but it noted the incomplete coverage of the data cited. We have more comprehensive data, but just as we started our analysis, DeepSeek’s release sent markets into turmoil. The stage was set for a timely experiment. We first answer the question using our data. Then we see how the AI did.



Background



What proportion of open access is not paid for by APCs? In discussing this, a recent Open Café listserv post cited studies by Walt Crawford – a librarian, well-known in the academic library and OA communities for his analysis of open access. He has paid particular attention to “diamond” OA journals, which charge neither readers nor authors. His studies are based on data from the Directory of Open Access journals (DOAJ). Excellent though both sources may be – and, full disclosure, we contribute to the DOAJ – the DOAJ’s remit covers only fully OA (“gold”) journals.


As listserv founder Rick Anderson noted, “By counting only articles published in DOAJ-listed journals, Crawford’s studies radically _undercount_ the number of APC-funded OA articles published – because DOAJ does not list hybrid journals, which always charge an APC for OA and which produce a lot of genuinely OA articles (though exactly how many, no one knows).”



Using our data


Actually, we do know … or at least have some fair estimates of hybrid OA. Our data allows us to determine the share of open access output in APC-free journals, as follows.


Sources: Publishers' websites, OpenAlex, DOAJ, Delta Think analysis. © 2025 Delta Think Inc. All rights reserved.


The charts analyze shares of open access articles published in both fully OA and hybrid journals.


  • They show the split between output in APC-free journals (“No APC” in the orange) and in journals where APCs were charged (“APC” in the grey).
  • Journals count as APC-free if they were free of APCs in the year of publication.
  • The data cover the five years 2018-2023 (consistent with the Open Café post).
  • Around 13% of all OA output in our sample is APC-free (left-hand pie).
  • As we often discuss, individual publisher’s experiences may vary. The right-hand pie shows the subset of output for Physics. Note how the APC-free share rises to almost 47% in this case.
  • If we choose a slightly earlier timeframe – which will become relevant below – we see very little change. Over 2018-2022, there was very little change in the numbers.
  • Our data allows further analysis, for example by over 200 subjects, or comparing society titles with others. Please get in touch to find out more.


Our estimate of a 13% APC-free share of the total market compares with around 33% noted in the Open Café post, citing Walt Crawford’s data.


What about AI?


If we wanted to use public sources, we would consider this to be a classic web research problem. This should suit generative AI (GenAI). ChatGPT is arguably the most well-known example of this technology. With disruptor DeepSeek grabbing headlines recently, it seemed opportune to compare both with the benchmarks from our own data. We can also test DeepSeek’s claims to deliver an equivalent level of performance to ChatGPT.

The actual question


Users of GenAI tools rapidly discover that the phraseology and nuances of questions asked – aka the prompt – are vital. The more detail and direction you give them, the better. We used:


what proportion of open access is diamond open access? list numbers of articles


and share for each of the last 5 years


Some prompts can run to paragraphs. But we wanted to test the AI’s ability to process a natural sounding and simple question, to interpret the term Diamond, and to see if it would spot the inclusion of fully OA and hybrid. We needed to direct it to the specifics of our required results. The 5-year period was mentioned in the Open Café post. We chose a “one shot” prompt (not asking the model to refine its first answer), as we wanted to see how our virtual research assistant would perform without further direction.



The answer



The table below shows the results.


*ChatGPT 4o did not give the answer directly. Instead, it gave Diamond OA & APC-funded OA as separate shares of all output, so we needed to calculate Diamond share of OA.

†ChatGPT 4-turbo with “Reasoning” switched on.


The table above summarizes the AIs’ answers.


  • The technology continues to evolve, so there are different versions of it – different models – available. We see this for ChatGPT, where different versions of the same AI give different responses.
  • A particular model may give different replies if you ask the same question more than once. ChatGPT seemed stable, but DeepSeek exhibited this behavior, as indicated in the “Attempt” row.
  • ChatGPT 4-turbo has a “Reason” feature. When enabled, this shows its method (reasoning) and then appears to return more verbose results. The annual figures we cited result from the “Reason” mode. Without it, it returned similar numbers to ChatGPT 4o.
  • Looking at share of OA output, there’s a big variation between ChatGPT and DeepSeek. There’s a smaller variation between the two DeepSeek attempts. DeepSeek did very well and got close to our estimates.
  • The DeepSeek attempts showed big variations in the article numbers. DeepSeek attempt 2 returned similar article numbers to ChatGPT 4-turbo, but the respective shares were roughly 2.5x different.
  • The models were only able to show figures up to 2022.


A research assistant’s ability to explain their results is important, and so we assessed the models on their communication ability. In practice both research assistants and AI can be asked to refine their work. But for the purposes of this analysis, we used a one-shot prompt.


  • The word count of replies varied. Chat GPT-4o had single-paragraph, single-number brevity. The others mixed bullets, prose, and tables.
  • Only one model added a hyperlink to its source(s), noted as follows.
  • ChatGPT-4o cited encyclo.ouvrirlascience.fr. It was the one model to include a link.
  • ChatGPT 4-turbo volunteered sections describing definitions, included headline numbers from some sources, tabulated yearly estimates, discussions comparing results from different sources, caveats on data limitations, and a summary. It cited Wikipedia as its source. It noted that its 5 years of estimates were based on data from 2017-2019, but it’s unclear whether it calculated them or was citing some language it found.
  • DeepSeek volunteered brief discussions about definitions, trends, annual estimates, and key sources. The sources seemed to be directions for further reading, including the OA Diamond Journals Study (2021) by cOAlition S and OPERAS, which it considered to provide the most comprehensive view, the DOAJ, and Web of Science/Scopus (noting they underrepresent Diamond OA journals, as many are smaller or regionally focused).
  • The verbose models attached general caveats to the data, but none picked up on the subtleties of defining Diamond. If a journal has only one APC-free year, say as an introductory discount, does it really count as a true Diamond or sponsored journal?


Conclusion


In working with and assessing GenAI models, it is important to set expectations correctly. They are more like an inexperienced intern than a seasoned researcher. Prompts need to be refined, several attempts made, and human judgement exercised over the results.


DeepSeek’s results were surprisingly good. It undershot our estimate by a few percentage points. ChatGPT overestimated considerably – perhaps reflecting its use of the problematic sources noted in the original listserv post. The additional material selected by the models was useful in explaining the concepts and drivers behind that data. The differences between our samples and those the models cite explain some of the variance in the data. But this doesn’t explain some of the big differences – e.g. same #s articles, but different shares or vice versa, with 2.5x variance.


Perhaps the GenAI models, in their current forms, are not ideally suited this sort of numerically focused research. However, they appear to be progressing in the right direction. (Note: We did not attempt this query on either of the Deep Research models, currently offered by OpenAI and Google Gemini.)


For fun, we tail-ended our assessment by asking each of our AIs “what's the answer to life, the universe and everything?” ChatGPT went to the Hitchhiker’s Guide to the Galaxy answer of 42. DeepSeek was more expansive, noting how the question encompassed broader philosophic endeavors. However, a light-weight offline version of DeepSeek that we experimented with gave perhaps the best reply. It concluded that, while 42 is a clever punchline,


“…it also reflects on the human tendency to search for significance where there might not be one.”


Fair point.

---


This article is © 2025 Delta Think, Inc. It is published under a Creative Commons Attribution-NonCommercial 4.0 International License. Please do get in touch if you want to use it in other contexts – we’re usually pretty accommodating.


---


TOP HEADLINES


What Gets Missed in the Discourse on Transformative Agreements – February 12, 2025

"Transformative agreements have a crucial role to play in the transition to open access, and while they may not fix structural problems, they are a reflection of the dynamic publishing landscape and the diverse needs of the academic community."


Serbia Adopts Open Science Platform 2.0 – February 4, 2025

"The Ministry of Science, Technological Development and Innovation of Serbia has adopted a new national open science (OS) policy – the Open Science Platform 2.0 – that applies to all Serbian publicly-funded research projects and programmes. EIFL welcomes the adoption of the policy, which significantly expands the scope of OS efforts in Serbia, and updates the country’s first national OS policy, which was adopted in 2018."


cOAlition S to sunset the Journal Comparison Service in April 2025 – February 3, 2025

"cOAlition S announces it will discontinue the Journal Comparison Service (JCS), effective from 30th April 2025...The JCS was launched in September 2022 to support the Plan S principle that Open Access fees should be transparent and commensurate with the publishers’ services. It was developed following detailed consultations with librarians, publishers, legal experts, and software developers with the aim to shed light on open access publishing fees and services."


Announcing a New Report on Open Educational Resources – January 22, 2025

"In the fall of 2023 we announced the launch of a new research project, funded by the William and Flora Hewlett Foundation, designed to assess the impact and implementation of open educational resource (OER) initiatives at public institutions of higher education. Today, we are publishing the resulting report, based on an initial literature review and interviews with OER leaders in four US states."


European Diamond Capacity Hub launched to strengthen Diamond Open Access publishing in Europe – January 21, 2025

"The European Diamond Capacity Hub (EDCH), a pioneering initiative to advance Diamond Open Access publishing across Europe, was officially launched on 15th January 2025, in Madrid. OPERAS, the research infrastructure for open scholarly communication in the social sciences and humanities, will serve as the fiscal host of the European Diamond Capacity Hub."

 


OA JOURNAL LAUNCHES


All EMS Press journals open access in 2025 following another successful Subscribe To Open round - February 5, 2025

"EMS Press publishes all journals open access in 2025 for the second consecutive year, following a successful Subscribe To Open (S2O) round...all 22 journals in our Subscribe To Open (S2O) programme will be published as open access for the 2025 subscription period."

Impelsys and Delta Think Join Forces to Expand Strategy and Technology Capabilities

By Lori Carlin • December 4, 2025
Impelsys and Delta Think Join Forces to Expand Strategy and Technology Capabilities for Publishing, Scholarly Communications, Education, and Healthcare Communities

News & Views: Market Sizing Supplement 2025 – Your Mileage May Vary

By Dan Pollock and Heather Staines • December 2, 2025
Overview Each year, our scholarly market sizing update and analysis goes way beyond open access headlines. One consistent finding is that market share of open and subscription access is highly dependent on subject area. This month we look at how to best use our Delta Think Data and Analytics Tool (DAT) to understand and analyze these variations. With coverage of approximately 220 detailed subject areas, the data shows that headlines can sometimes mask important detail. Background Since we began our scholarly journal market analyses in 2017, one of our core objectives has been to enable deep analysis of our headline findings. Our annual market share updates represent a summing of data – more than 200 detailed subject areas, 200 or so countries, also split by society vs. non-society journal ownership. This level of detail is clearly too much for our monthly short-form analyses, so we present the market-wide headlines in our annual updates. However, by picking one subject area as an example, we can see how much nuance lies beneath the surface, and why these variations matter. Subscribers to DAT can use our interactive tools to quickly and easily see each level of detail and filter for just those relevant to their organization. Market Share Variation by Subject Area Our latest market headlines suggested that open access (OA) accounted for just under 50% of article output in 2024. However, this headline proportion varies considerably by subject area.

Charleston Conference 2025 Reflections

By Lori Carlin & Meg White • November 19, 2025
Navigating Uncertainty, Innovation, and the Winds of Change As the Charleston Conference 2025 wrapped up, one thing was clear: scholarly communication continues to evolve against a backdrop of uncertainty: economic, technological, and policy-driven. Yet amid the turbulence, conversations throughout the week pointed toward resilience, adaptability, and even optimism. As Tony Hobbs observed during the Shifting Tides policy session, “the good news for scholarly communication is that due to technology advances, it is now possible to sail into the wind.” The Elephant in the Room: Doing More with Less Heather Staines Every conversation I had in Charleston seemed to circle back to one thing: budgetary uncertainty. Whether the concern was policy changes like potential caps on overhead or shifting grant funding or the ripple effects of declining enrollment, both domestic and international, everyone was asking how to do more with fewer resources. This theme ran through the plenary Leading in a Time of Crisis, Reclaiming the Library Narrative, and even the lightning sessions, a shared recognition that we’re all trying to redefine what “enough” looks like. What stood out was how data-driven decision-making has become essential. Libraries, publishers, and service providers are not just analyzing what to add, but what to let go of, all in an effort to find a new balance. And then there’s AI. We have moved beyond “sessions about AI” to “AI everywhere.” I will admit that I once thought AI was a solution in search of a problem, but now it’s woven through nearly every conversation. Librarians are leading the way on AI literacy, while publishers and service providers are using AI to innovate to meet changing research needs. The uncertainty is real but so is the shared determination to adapt, learn, and move forward together. The Long Arm of the Law and Its Reach into Scholarly Communication Meg White One of the things I love about Charleston is that there is always a moment that challenges me to reframe how I think about the work we do. This year’s Long Arm of the Law session did exactly that. It was a vivid reminder that the legal and policy currents swirling around us are not abstractions; they shape our ecosystem in ways we can’t afford to ignore. Paul Rosenzweig set the stage with a fascinating and lively walk through the history of executive orders. Hearing that Washington issued just eight while later presidents relied on them more frequently primarily to advance political agendas made the evolution very real. What stood out was the fine line between legitimate executive authority and overreach, and how easily those boundaries can blur. Nancy Weiss then brought the conversation directly into our lane with her analysis of an Executive Order directing the Institute of Museum and Library Services (IMLS) to reduce its activities to the bare legal minimum. Her experience as former General Counsel gave us an inside view of what such a directive could mean for libraries, museums, and cultural programs, all places where so much of our community’s work takes root. Sessions like this are why Charleston continues to be invaluable to me. They stretch my understanding, give me new context, and remind me that staying informed is part of how we navigate change together. Data-Driven Insights: The 2025 Author and Researcher Survey Lori Carlin My week was cut unusually short (for me) by other meetings I had to fly off to, but I still managed to squeeze in 2.5 days of interesting sessions, discussions, and ‘business casual’ gatherings. The first two events I attended this year were definite highlights, both of which were the brainchild of and brilliantly orchestrated by my colleague, Heather Staines – the Vendor Meetup on Monday evening and the Leadership Breakfast on Tuesday morning. Both were jam packed and filled with lively conversation. If you’re not familiar, the Vendor Meetup is an open, casual gathering (sponsored this year by Get FTR) designed to give vendor representatives, especially early career attendees, who attend only for Vendor Day a chance to socialize and network, something they often miss when they’re in and out in a single day, but all are welcome to attend! The Leadership Breakfast, a smaller invitation-only event designed to give a more intimate networking experience within the larger Charleston Conference, is always a thoughtful session centered on a pressing issue of the day, and this year was no exception. The discussion focused on sustainability across the entire scholarly communication ecosystem—from funders to libraries to publishers. Frankly, no one can unhear the words of one of the panelists (a library director) when he commented that his budget has dropped from ~$7M to ~$5.4M in the last 24 months … with more to come. Finally, I’m a little biased, but I dare say I and my panelists were very pleased with the session I moderated focused on the impact of US research funding changes, which highlighted info from Delta Think’s Spring 2025 Author and Researcher Survey, along with how publishers who participated used the data to inform their strategies. We also had a librarian on the panel who informed the audience about the impact of these changes on universities overall and libraries in particular. As you may know, the survey data showed rising concern about institutional support, with many researchers rethinking how they publish and participate in conferences. Respondents also described how tightening budgets are straining peer review and research dissemination, while responses varied sharply between U.S.-based and international authors, reflecting distinct policy and institutional pressures, it also showed that the impact is being felt globally. In the tradition of Charleston, what made the session so powerful was the discussion. Colleagues from societies, publishers, and libraries focused on how they are using these insights to understand the challenges and to act on them. From adjusting publishing strategies to helping researchers to growing relationships in other markets, to shaping advocacy and outreach activities, organizations are using these insights to inform resource and budget direction in innovative ways. For me, that was the real takeaway: turning evidence into collaboration, and progress. Even in uncertain times. We’re running the survey again now with plans to compare results to the Spring version. If you’re interested, there is still time to sign up! End of An Era (Two, in Fact!) This year’s conference marked a pivotal moment: the first without the in-person presence of founder, Katina Strauch (though we were grateful for her virtual participation), and the well-earned retirement of longtime Conference Director Anthony Watkinson, who rang his iconic bell one last time. We would not be here without them and their visionary colleagues who built this community from the ground up. Thank you, Katina and Anthony. Charting What Comes Next If there was one metaphor that captured Charleston 2025, it was motion; not adrift, but deliberate progress in the face of resistance. From policy updates to AI integration to the enduring strength of the scholarly community, the week’s sessions affirmed that innovation often takes root during uncertainty. As Tony Hobbs reminded us, even headwinds can propel us forward — if we learn how to adjust our sails.

News & Views Webinar: Market Sizing Update 2025

By Heather Staines • November 6, 2025
We are proud to share a video recording of our October News & Views companion online discussion forum! Join us for our annual update on the volume and revenue associated with Open Access publishing. If you missed the session, or if you attended and would like to watch/listen again, or share forward with friends, please feel free!

News & Views: Market Sizing Update 2025 – Has OA recovered its mojo?

By Dan Pollock and Heather Staines • October 21, 2025
Overview After a rocky couple of years, the open access (OA) market may be finding its footing again. Each year, Delta Think's Market Sizing analyzes the value of the OA scholarly journals market—that is, the revenue generated by providers or the costs incurred by buyers of content. Our analysis estimates that the OA segment expanded to just under $2.4bn in 2024. Although growth has improved compared with last year’s deceleration, it continues to lag behind the broader historical trend for OA. The proportion of articles published as OA has declined slightly, likely driven by continued reduction in the output from the large OA publishers. This trend has benefited established publishers, who saw growth in OA activity and revenue as they continued to consolidate their positions. Looking ahead, OA could soon begin outpacing the broader journals market once again—but likely through different growth drivers than in the past. Read on to see what those shifts might look like. Headline findings Our models suggest the following headlines for the 2024 open access market:

Join the Fall 2025 Author/Researcher Survey

By Lori Carlin & Meg White • October 13, 2025
Collaborate with Delta Think to uncover how funding and policy uncertainty continue to reshape the research ecosystem — and gain tailored insights for your community.

Smart AI, Smarter Scholarship

By Lori Carlin & Meg White • September 25, 2025
Introduction: One question, two paths  A recent essay in The Conversation posed the question, “Is ChatGPT making us stupid?” The author examined emerging research suggesting that over-reliance on AI tools for writing can dull critical thinking, originality, and even memory retention. But as the author points out, AI has the potential to augment human intelligence when used well , acting as a catalyst for deeper thinking rather than a shortcut around it. We agree and seek to guide our clients in determining how to use AI to strengthen research and scholarship. From concern to opportunity When AI is approached as a collaborator, it sparks creativity, deepens inquiry, accelerates problem-solving, and amplifies creativity. It can strengthen teams, enhance services, and improve efficiencies across the publishing enterprise. Turning Ideas into Action Here’s how Delta Think can help you transform smart AI potential into purposeful, strategic action: Strategy and Market Research Focus: Identify where AI can deliver the most value for your organization, grounded in community needs and behaviors. Delta Think Approach: Gather and analyze evidence through quantitative and qualitative methods to uncover how your community – your researchers, authors, reviewers, and readers – are using AI now or, better yet, where and how they could be using it in the future. Marrying their unmet needs with your strategic goals creates your roadmap to future success. 2. Build vs. Buy Decisions for AI-Powered Products Focus: Develop proprietary AI solutions, partner with trusted vendors, or combine the best of both approaches to suit your needs. Delta Think Approach: Assess your current state and future needs, design decision frameworks that weigh cost, capability, risk, speed-to-market, and long-term scalability, and build the approach that will work best to support your business goals and community needs. 3. AI Policy and Governance Focus: Ensure responsible, transparent, and ethical AI use that safeguards scholarly integrity. Delta Think Approach: Facilitate the development of your AI governance with the creation of important guardrails and policies, working to mitigate bias and hallucination risks, safeguarding research integrity while enabling innovation. 4. UX/UI Testing for AI Products and Features Focus: Design AI experiences that enhance human engagement. Delta Think Approach: Test results, interfaces, prompts, and transparency signals to keep users informed, empowered, and confident in your products and tools. 5. Licensing and Partnership Strategy Focus: Leverage commercial arrangements to unlock AI potential while aligning with your mission and values. Delta Think Approach: Guide you through licensing agreements, proprietary data partnerships, and collaborations that create sustainable competitive advantage and strategic revenue streams. Turning Ideas into Impact By reframing the conversation from Can AI substitute scholarship? to How does AI amplify scholarship? , publishers can lead the next wave of innovation. Delta Think’s collaborative approach ensures that your organization’s adoption of AI enhances creativity, critical thinking, and trust. We can help you map out your bespoke AI-strategy roadmap, develop new products and services, test prototypes, and design governance guidelines. Reach out today or schedule some time at the Frankfurt Book Fair (10/14-16) to discuss how Delta Think’s expertise and proven methodologies can help your organization unlock key insights and drive innovation.

News & Views: How much will cuts to NSF funding affect the future of scholarly publishing?

By Dan Pollock and Heather Staines • September 9, 2025
How might planned cuts to funding of the US National Science Foundation affect scholarly output? In our last News & Views we analyzed how the headline cuts might apply to relevant activities. This month we examine how journals may be impacted and model some scenarios quantifying the impact on global scholarly output. Background The US National Science Foundation is an independent US federal agency that supports science and engineering across the US and its territories. In its 2024 financial year (FY) 1 , it spent around $9.4 billion, funding approximately 25% of all federally supported research conducted by US colleges and universities. In July we looked at how reported funding cuts and NSF budget cuts proposed by the US Government might affect the NSF’s output of research papers. We found that in the near term the effects would be limited, as the cuts focus on NSF activites that produce low volumes of papers. However, cuts proposed over the coming year may have a more profound effect as they are deep and affect research activities. We have also previously analyzed proposed cuts to funding of the National Institutes of Health (NIH). We noted how cuts to the world’s largest producer of biomedical research could have a profound effect on publication outputs. So how do cuts to the NSF stack up? The effects on journals As ever, the headlines and averages are unevenly distributed, so we looked at how individual journals might be affected. 

News & Views: Will cuts to National Science Foundation funding affect scholarly publishing activity?

By Dan Pollock & Heather Staines • July 29, 2025
The US Government has planned cuts to funding of the US National Science Foundation (NSF) in 2025 and 2026. Before we can undertake a full analysis of how these cuts might affect publishers, we must unpack some data. This month we put the cuts in context, looking at how the cuts impact research and the scale of NSF output. And we find they may not affect research in the ways the headlines suggest. We will follow up with a future analysis modelling specific scenarios of impacts on publisher submissions. Background The US National Science Foundation is an independent US federal agency that supports science and engineering across the US and its territories. In its 2024 financial year (FY) 1 , it spent around $9.4 billion, funding approximately 25% of all federally supported research conducted by US colleges and universities. In May 2025, the New York Times (NYT) published an article analyzing proposed cuts to NSF funding by the current US Government. The NYT’s analysis suggested a 51% cut in funding from 1 January through 21 May 2025, with a further 56% reduction proposed for next year 2 . We have previously analyzed effects of proposed cuts to funding of the National Institutes of Health (NIH). The proposed cuts to the NSF are deeper, so might they have an even greater negative effect on publication volumes? Understanding what the cuts apply to The 51% cut in 2025 covers 140 days, equivalent to a 20% annualized cut. So could we see the same level of reduction in papers this year? And could this be followed be a 56% drop next year, as the 2026 cuts cover a full year? As with our analysis of the NIH, we need to understand how the changes in funding translate into research activities, and thence into corresponding volumes and timing of publication output. We therefore analyzed the NSF’s own budgetary figures to put the cuts into context. 

Delta Think’s Survey Results: What 13,000 Researchers Revealed About the Fallout from US Funding Cuts

By Lori Carlin and Meg White • July 24, 2025
This spring, Delta Think collaborated with 27 professional societies and associations to launch a Global Author/Researcher Survey to understand the ripple effect of US government research funding cuts. Our goal was to explore how researchers are navigating a rapidly evolving landscape, especially as US federal funding and policy decisions cast long shadows over the global research community.  More than 13,000 researchers across every major discipline and 135 countries shared their voices through our survey. While the detailed findings are deep and wide-ranging, one thing is clear: the ground is shifting. Uncertainty Is Driving Change in Research Behavior Delta Think deeply analyzed the data by six major disciplines: Health Sciences, Life Sciences, Physical Sciences, Engineering & Technology, Social Sciences, Arts & Humanities. Nuances vary by each main field, but some factors were universal. US-based researchers are signaling deep concern – and they’re bracing for change. Many anticipate reductions across publishing output, participation in peer review, and conference attendance. For example, 62% of US authors across all disciplines expect to publish fewer articles in the next 1–2 years, citing policy and funding challenges . “My research progress is now in ‘conservative mode’ in case funding is pulled from us with no notice. We cannot plan further out and have lost our trust in the federal government.” Primary Investigator (PI) at a US Academic Medical Center But the concerning news isn’t limited to the US. International researchers indicated their intention to pull away from US-based journals, threatening to reshape the global flow of research. In fact, a full 50% of international authors across all disciplines indicated that it is now important to them to submit their manuscripts to non-US journals. “We're doing everything we can to reduce our connections to the US, including looking for journals to publish in that are not based in the US.” Mid-Career PI, Biological Sciences, Canada Top Concerns: What Keeps Researchers Up at Night? One of the clearest patterns that emerged is the contrast in what researchers view as their most urgent challenges: For US researchers , the top concern is straightforward: elimination of research funding . This fear extends beyond specific grants—it reflects a deep anxiety about career stability, institutional viability, and the future of scientific advancement. For international researchers , the primary worry is academic freedom and collaboration , with many expressing concerns about losing access to US research infrastructure, data, and professional networks if international cooperation is reduced. While these represent the top concerns, the survey results reveal many others by discipline, career stage, and other factors, including specific community details for each of the 27 participating societies and organizations upon which to develop their future strategies. Looking Ahead: Tracking Trends with Fall 2025 Survey This spring’s survey was just the beginning. Delta Think will conduct a follow-up survey in October/November 2025 to track how attitudes and behaviors continue to shift. This next phase will allow us and the participating organizations to move from snapshot to trend — providing deeper insight into the lasting impact of funding and policy uncertainty. Joining in for Survey 2 is NOT limited to Survey 1 participating organizations. All are welcome to participate in this next round and have access to the deep data behind these high-level insights and much more. Turning Ideas into Action The Delta Think team designed this initiative not just to gather data, but also to support our partners across the scholarly ecosystem. By combining rigorous research design with deep industry context, we’re helping publishers, societies, and institutions make informed, strategic decisions in uncertain times. If you're interested in learning more about the findings, discussing how they apply to your organization, or joining the Fall 2025 survey, we’d love to connect. Please email Lori Carlin to start the conversation.