The 2026 State of Open Infrastructure report is out. This is our third annual analysis of where and how funding flows (or doesn’t) to the infrastructure that underpins research and scholarly communication. So, what can 25 years of grant data, coupled with the information IOI gathers, validates, and regularly updates in Infra Finder tell us about the open infrastructures that keep research running?
Some of the disruptions over the last two years have been very loud, with budget cuts and policy shifts across the globe. Herein, we try to spotlight some of the ripple effects that often are quieter and harder to see. The biggest take-away may well be that the fragility the field feels today is not new,: instead, it is a built-in characteristic of the structures and models used to fund open infrastructure for decades.
Where the money goes.
We’ve analyzed more than US$663 million in open infrastructure grant funding across 1,065 award clusters, 47 funders, and 94 open infrastructures from 2021-2026. Here’s what that money actually paid for:
- $192 million (29%) funded open infrastructures directly
- $411 million funded research that depends on open infrastructures
- $23 million funded adoption of open infrastructures
We want this to be crystal clear to our readers: the vast majority of the money flow has gone to the work that infrastructure makes possible, not to the infrastructure itself. That isn’t a bad thing on the surface: we want infrastructure to be strong and to enable lots of exciting research discoveries, publications, and dissemination. But the gap here tells a more complicated tale: Creative Commons sees nearly $29 spent on research that uses its work for every $1 that reaches it directly. DOAJ sees nearly $24 spent for every dollar it receives, and RRID sees more than $22 per dollar.
Again, this could signify efficiency: a modest open infrastructure investment carrying an enormous research enterprise. It could also signal that there is a growing body of research resting on infrastructure that is not funded to grow with it. At IOI, we think both of these readings are true and that the second one should keep funders up at night.
We pay to build, not to run.
As this year’s report shows, only a thin slice of direct funding has supported open infrastructures. And an even thinner slice of that $192 million funded operations rather than innovations for these open infrastructures.
So, rewinding, only 29% ($192 million) funded open infrastructures directly. Of that, $105 million went to research and development (R&D) while only $64 million went to operations, or the work of keeping systems available year after year.
This doesn’t stem from anyone’s bad judgment; it’s structural. Grant proposals are built around projects, and we expect that projects will build things. This produces an ecosystem where launching new things is very lucrative, but keeping them alive thereafter is not. Infrastructure organizations end up running just to stay in place, raising money for new features that often complicate their codebases and services so that they can very quietly keep their lights on.
Global reliance, single-source support
Research that depends on open infrastructure is funded from all parts of the world. That is not the case for the infrastructure, though. Of the 34 infrastructures in this year’s analysis that had more than one direct grant, 22 drew their support from a single country, even though the research relying on them is funded from nearly 3 countries on average. Nine countries, for example, fund research that depends on arXiv; one country funds arXiv’s grants. For OSF, that ratio is seven to one.
Funders are, of course, similarly concentrated. The US National Science Foundation appears in the direct funding of 13 different infrastructures, Institute of Museum and Library Services in 10, and Mellon, Sloan, and the European Commission each appear in eight. When one of these funders shifts priorities, the risk isn’t isolated; it’s explosive. Right now, $42.1 million in active US federal grants to open infrastructure is exposed to immediate budget threats, indirect cost caps, and the near-elimation of some agency activity. Infrastructure that the world depends on is only as durable as the priorities of the one government that sustains it.
The infrastructure of integrity
The report's second section asks how open infrastructure makes a published research output trustworthy, especially when generative AI is making research fabrication increasingly easier. The recently emerging threats include preprint flooding, AI-powered paper mills, ghost authorship, and hallucinated citations polluting the scholarly record.
The good news? We already have much of the infrastructure needed to validate and establish the provenance and trustedness of a research object. The bad news? Well, that’s consistent with the rest of this report. These are largely nonprofit, community-governed systems that are chronically underfunded relative to how much the ecosystem leans on them. Free to use does not mean free to run.
Signals we are watching.
The final report section tracks forces that are usually discussed separately but that we think are deeply intertwined: the AI investment boom, the severe pullback on public science budgets, and the rise of digital sovereignty. We sort seven signals into three groups: Risk, Watch, and Opportunity.
In “Risk,” we talk about the dangers posed by philanthropic funders leaving the ecosystem simultaneously with the federal open science pull backs. We also look at how an open tool, Tailwind, has been seeing its usage grow while its revenues are dropping alarmingly. In “Watch,” we share our perspectives on the EU’s new EDIC Digital Commons, COS’s decision to streamline OSF, and PLOS’s proposal to move beyond the article as the unit of value. And in “Opportunity,” we discuss the $500 million HumanityAI funder coalition and the Dutch government’s funding for infrastructure maintainers’ training in sustainability.
What could change the trajectory from fragility to resilience?
None of these patterns were designed on purpose, and no single organization created them. Responsibility has been passed around the same small table for years, and each player has a reasonable argument for why it belongs to someone else. This is the reason no single organization or stakeholder group can fix the problem alone.
Still, some of this is imminently fixable. The funder pool may need a new view of the ways it is funding research and infrastructure in order to strategically consider what reciprocal relationship those two things need to have. Keeping systems running may cost far less than rebuilding, and concretely, based on our findings in this report, we recommend:
- Fund infrastructure directly, not just the research it supports
- Treat maintenance as a budget line, not a one-time cost
- Build persistent identifiers into policy
Above all, this report demonstrates that all stakeholders today need to look hard at their dependencies and how and whether they are funding them.
Coordination begins when funders, institutions, and open infrastructure providers all can look at the same evidence. That’s why we produce this report.
- Read the 2026 State of Open Infrastructure Report
- Explore the Data Dashboards
- Download the Dataset (opens Zenodo in a new tab)
We're grateful to Crossref, our Supporting Sponsor, which was the first to step up and put its money where its values are. We also thank our Community Supporters, RDA, CLIR, AGU, and Cassyni, and our core funders, including the Andrew W. Mellon Foundation, Arcadia Fund, and the Wellcome Trust. That this report is funded by the communities it describes is perhaps a testament to the value of cross-sector coordination.
Get involved
We’d love to have your feedback on the report through our feedback form. Your feedback is very important and will help shape the topics for future editions of the report.
We hope this report sparks your curiosity about the tools and systems the research and scholarship community relies on and helps you to think about how your decisions line up with your vision for the future of research and scholarship. If you need additional information and support around these decisions, please reach out to us. We are experienced in providing tailored research and recommendations, including discovery, due diligence, and stakeholder feedback collection, to inform the decisions you are facing.
