Academic scientists and research institutes are increasingly being evaluated using digital metrics, from bibliometrics to patent counts. These metrics are often framed, by science policy analysts, economists of science as well as funding agencies, as objective and universal proxies for scientific worth, potential, and productivity. In biomedical science, where there is stiff competition for grants from the National Institutes of Health (NIH), metrics are sold as a less arbitrary way to allocate funds, yet the funding context in which metrics are applied is not critically examined. Success by the metrics is in fact inextricably linked to the distribution of NIH funds, and from the 1980s to the 2000s, NIH funding has been marked by high inequality (elite investigators and institutes get the lion’s share of resources) and decreased mobility (those who start at the bottom are less likely to rise to the upper ranks). Elite investigators and institutes currently produce the bulk of prestigious publications, citations, and patents that commonly used metrics valorise. Metrics-based evaluation therefore reproduces, and potentially amplifies, existing inequalities in academic science and rich-get-richer effects.