The Deal of the Century

In “Who Will Keep Research Data Infrastructure Open and Running?” (Issues, Spring 2026), Jennifer Gibson and Kaitlin Thaney articulate a silent crisis in modern science: the systematic underfunding of the open research infrastructures that make global scientific breakthroughs possible. With nearly 200 research data repositories already permanently closed over the past 25 years, we are actively losing access to vital knowledge. And the picture may become even more bleak, as the full impact of closures will become apparent only with future knowledge gaps when it is too late.

From an economic perspective, the case for maintaining shared, open research infrastructure is a no-brainer and arguably the deal of the century. Our work at the Center for Open Science supports this view. We maintain and support the Open Science Framework (OSF), a platform hosting a variety of research assets. By prompting researchers to affiliate their funders and institutions in their research project metadata, OSF provides a live blueprint for mapping research outputs back to their financial and institutional sources. Approximately 1% of public research assets hosted on OSF include funder metadata, linking them to major funders such as the European Commission, the National Science Foundation, and the National Institutes of Health. Implementing a modest infrastructure support fee of $1,000 per verified research project would generate around $1.5 million to $2 million in direct, sustainable annual revenue. As participation expands across funders and infrastructures, costs could be shared more broadly, creating a sustainable ecosystem-wide funding model.

While the specific $1,000 per project fee is just an illustrative example, it highlights a concrete, multistakeholder road map for implementing a direct-funding model on a global scale. Extrapolating this model for funders to pay annual fees based on the number of verified research project assets hosted on open research infrastructure, along with other diversified revenue strategies, we could direct needed support to platforms—OSF and others—to transform the ecosystem into a sustainable utility.

As another option to stabilize the research ecosystem, the authors suggest funders allocate 10–15% of research grants directly to data infrastructure. But whatever mechanisms evolve, the nation must commit to sustained operational funding before more vital repositories go dark and we permanently lose access to this knowledge. No matter the application—whether for project, community, or whole disciplines—open research infrastructure supports them all for a fraction of the cost of building and maintaining bespoke solutions. If we do not urgently develop a plan to support open research infrastructure, we will inevitably spend more and gain less.

This fragility is especially urgent given the ongoing artificial intelligence boom, which relies heavily on these open research asset troves to train large language models. Not only does this add costs and unpredictable impacts on the infrastructure hosting the research assets, but without sustainable funding to ensure quality and longevity, AI risks becoming merely a “storyteller” rather than a reliable engine for scientific progress. As the authors warn, “If the federal government fails to act, it risks losing out on its own investments in open infrastructure.” Allowing these foundational repositories to wither not only would squander decades of high-yield public investments but also create a dangerous vacuum.

Nici Pfeiffer