The missing link in FAIR data policy: data resources

Abstract

The FAIR principles have reshaped research policy, but their implementation still relies largely on individual researchers—many of whom lack the expertise or support needed to make data truly reusable. Realising FAIR’s promise requires sustained investment in the infrastructures that organise, standardise, and curate data: deposition databases and knowledgebases. These data resources are especially critical for AI, which depends on large, high-quality, and consistent data. Landmark advances like AlphaFold and the COVID-19 response illustrate how sustained curation and standardisation in expert resources such as UniProt and the Protein Data Bank have enabled rapid innovation. Yet data resources remain precariously funded, jeopardising long-term sustainability and the expert workforce they require. To support ambitious, data-driven science, funders must align policy and budgets by establishing dedicated mechanisms that allocate a small (e.g., 1%), but strategic and stable share, of research funding to core data infrastructures. This would maximise the value of public investment, strengthen open science and international collaboration, and unlock the full potential of FAIR.