Between January 2020 and September 2025, while the world was living through the worst pandemic in a century and struggling to absorb its lessons, research funders issued 26,028 awards for studies on pandemic-prone diseases. A sweeping new audit of that spending reveals an uncomfortable arithmetic: only 1.6 percent of those grants went to the field whose entire purpose is to detect outbreaks before they explode. The analysis, published on 29 August 2026 in the journal Health Research Policy and Systems by researchers from the University of Oxford, CPC Analytics and the World Health Organization’s Hub for Pandemic and Epidemic Intelligence, is the most systematic account yet of how the world finances the science of seeing epidemics coming. Its central finding is stark: despite everything the pandemic taught, the money has not followed the risk.

Pandemic and epidemic intelligence is the largely invisible discipline standing between a handful of unusual infections and a global emergency. It spans pathogen surveillance in every form: genomic sequencing of viruses circulating in humans and animals, scanning of municipal wastewater for genetic fragments that appear weeks before case counts climb, and sentinel testing in clinics. It includes event-based surveillance — the algorithmic mining of news reports, health-record patterns and online chatter for statistical anomalies — together with the modelling infrastructure, increasingly powered by artificial intelligence, that weighs whether a given signal is noise, a local cluster, or the beginning of something worse. The whole apparatus operates like an intelligence cycle: detect, verify, integrate genomic and epidemiological context, assess risk, and deliver the assessment to decision-makers while containment is still cheap. Increasingly the field is framed through a One Health lens, because most emerging infections begin in animals before spilling over into human populations. The COVID-19 response showcased what this machinery can do, from variant tracking that guided vaccine updates to analytics that shaped containment choices. Yet the study’s authors open with a blunt diagnosis: despite innovations driven by that response, research to improve surveillance and public health intelligence remains fragmented and poorly translated into practice.

That diagnosis is exactly what prompted the audit. In 2024, the WHO Hub for Pandemic and Epidemic Intelligence, based in Berlin, conducted an exercise to identify research priorities aimed at coordinating robust evidence generation and the methodologies needed to respond effectively to global health emergencies. A priority list, however, is only as useful as the funding that follows it, and funding landscapes largely determine which evidence exists when the next emergency begins. A team spanning Oxford’s Pandemic Sciences Institute — Emilia Antonio, Zaharat Kadri-Alabi and Alice Norton — joined by Isabel Redies of CPC Analytics and Barbara Tornimbene and Oliver Morgan of the WHO Hub, therefore set out to measure the alignment between the stated priorities and the real award record. Their question was deceptively simple: when funders around the world write cheques for research on epidemic-prone diseases, how much of that money is actually buying the intelligence capability the world says it needs?

To answer it, the researchers drew on one of the most comprehensive funding databases in existence: the Pandemic Preparedness: Analytical Capacity and Funding Tracking initiative, known as Pandemic PACT. Supported by Wellcome, the United Kingdom’s National Institute for Health and Care Research using UK aid, Canada’s International Development Research Centre, and UK Research and Innovation, the grant tracker systematically monitors research funding for pandemic-prone diseases, capturing award descriptions, funder identities, committed amounts and research locations. The team extracted the full dataset on 15 September 2025 and analysed awards made between January 2020 and September 2025, classifying each against the pandemic and epidemic intelligence priority themes. Supplementary files published with the study lay out the research categories and tags used, the definitions of each priority theme, and the step-by-step process for identifying intelligence-related awards from the tracker — an unusually detailed level of methodological transparency. The coding framework distinguished the focus of each award, the funding entity and the amount committed, and the locations in which the research would be conducted, allowing the investigators to map not merely how much money flowed, but from whom, to whom, and toward which pieces of the intelligence puzzle.

The headline numbers are sobering. Of the awards logged in the tracker across that window, just 1.6 percent centred on pandemic and epidemic intelligence — a few hundred grants scattered across an ocean of work on vaccines, therapeutics, pathogen biology and clinical care. What funding did materialise came from 29 funders, most of them based in North America and Europe, which invested at least US$503 million in the field. The qualifier “at least” matters: the figure is a documented floor rather than a ceiling, though the authors’ point survives even a generous reading, because even a larger sum would leave intelligence research marginal against the scale of the need. Equally striking is how rarely funders pooled resources. Jointly financed research awards were few, meaning the field’s limited budget was assembled in uncoordinated slivers rather than consolidated into strategic programmes capable of sustaining laboratories, data systems and expert teams between emergencies.

The shortfall was not uniform across the field; it was worst precisely where coordination and institutions matter most. Investments were limited across all of the research priority areas, but two domains stood out for their neglect: evidence translation to policy, and data governance. The first describes the “last mile” of surveillance science — the methods, institutional arrangements and evaluation research that turn analytical outputs into decisions a health ministry actually takes, at the speed an outbreak demands. The second covers the legal, ethical and technical scaffolding that determines whether data can move at all across borders and institutions: agreements for sharing pathogen genomic sequences, privacy protections that allow patient-level data to be linked without breaking public trust, and interoperability standards that let incompatible systems communicate. Starve those two areas, and even well-funded detection technologies stall in silos — samples sequenced but never compared, models built but never consulted, warnings issued but never acted upon.

The audit also exposes a stark geographic imbalance. Most of the funding identified supported research being conducted in high-income countries, even though pathogens with epidemic potential circulate disproportionately in lower- and middle-income settings where humans, livestock and wildlife intersect most intensively. Surveillance capability built in wealthy, well-resourced systems tells decision-makers little about how early-warning tools perform in district hospitals with intermittent electricity, in informal urban settlements where clusters first grow unnoticed, or in remote regions where a zoonotic spillover may go unrecorded altogether. When research capacity and financing are concentrated away from the front lines, the global system loses twice: it loses the local data quality that early detection depends on, and it loses the trained workforce that must operate surveillance infrastructure continuously, not only during crises. The authors argue that this pattern makes multi-country research collaborations essential — not as a matter of goodwill, but because an early-warning network is only as strong as its weakest, most under-resourced node.

The study’s conclusion lands on the machinery of funding itself. Targeted research investments and multi-country research collaborations, the authors write, will be crucial for enhancing effective global surveillance systems and future pandemic preparedness — and neither will emerge spontaneously from a landscape of 29 largely independent funders acting on their own priorities. Their prescription is improved coordination among funders: shared priority-setting, aligned funding calls, and transparency about who is financing what, so that gaps in data governance or policy translation become visible in time to be closed rather than discovered mid-crisis. Much of the infrastructure for such coordination already exists. Pandemic PACT is embedded in a wider ecosystem that includes the Global Research Collaboration for Infectious Disease Preparedness, or GloPID-R, a network of research funders created to coordinate investment in outbreak science. The audit effectively upgrades that transparency tool into an accountability instrument, handing every funder a view of the entire board rather than a single piece of it.

The economics of that argument are difficult to dismiss. Preparedness research is inexpensive relative to the alternative: the COVID-19 pandemic caused millions of deaths and inflicted economic losses measured in the trillions of dollars, while the intelligence systems that might have compressed its trajectory cost orders of magnitude less to build and maintain. Epidemic-prone diseases have not gone quiet in the meantime; the very existence of a tracker devoted to pandemic-prone pathogens is a reminder that the threat list is long and constantly renewing itself, and the World Health Organization’s decision to anchor its Hub for Pandemic and Epidemic Intelligence in Berlin reflects the recognition that data, not intuition, must drive the next response. Between emergencies, however, political attention decays and budgets contract, and surveillance is precisely the kind of invisible infrastructure that is noticed only when it fails. A funding audit cannot manufacture political will. What it can do is quieter and arguably more durable: measure the gap precisely enough that the next conversation between funders, governments and the World Health Organization begins from shared numbers rather than competing impressions.

There is also a paradox embedded in the data worth sitting with. The other 98.4 percent of awards did not vanish into irrelevance: work on pathogen biology, diagnostics and vaccines ultimately feeds the same early-warning ecosystem. The authors are not arguing that the rest of pandemic science is wasted money; they are arguing that the connective tissue converting those advances into detectable, interpretable and actionable intelligence has been left nearly unfunded. Published open access with its full extraction framework available, the study arrives at a moment when the quiet interval between outbreaks is exactly when such systems should be constructed. Whether the US$503 million identified since 2020 proves to be the floor of a rising investment curve or its ceiling will be decided by the 29 funders the study identifies — and by the many institutions around the world that will read its numbers and recognise, uncomfortably, their own absence from them.