ROI on research data infrastructure is positive in every documented study, ranging 5x to 800x

The return on well-maintained research data infrastructure is positive in every documented study, across disciplines and geographies. Representative measurements:

• EMBL-EBI — operates on ~£50M/yr; generates an estimated £1B-£1.3B/yr in user value; 20:1 to 26:1 return (E-0059, S-0104).
• UK Archaeology Data Service — produces £13M/yr in efficiency gains; 5:1 return (E-0065, S-0105).
• Australian NCRIS — $7 returned per $1 invested (E-0066, S-0106).
• XSEDE — generated $4.7B-$22.7B on $257.5M; 18:1 to 88:1 (E-0067, S-0107).
• Apon et al. — every $100K in research-computing salaries associated with $14.3M increase in HERD; every 100 TFLOPs with $1.3M increase (E-0068, S-0108).
• Protein Data Bank — operates on ~$6.1M/yr federal funding; generates ~$5.5B/yr in economic impact; 800:1 (E-0060, S-0109).

The PDB 800:1 is the documented outlier. The cluster from the other measurements (5:1 to 26:1) sits well within an order of magnitude across radically different domains and scales.

The economic case for research data infrastructure investment is not contestable on the question of whether it produces returns. It is contestable on the question of how to capture the returns — institutional vs. sectoral, and the gap between proven Tier 2 infrastructure (where these measurements come from) and the Tier 3 deployment the paper recommends (whose returns measurable against a regime that does not truncate at the grant cycle have not yet been priced because the regime does not yet exist).

Infrastructure investment is also inseparable from R1 status: the 2025 Carnegie threshold requires $50M annual research spending and 70 research doctorates (S-0110), and every infrastructure investment that supports research at scale qualifies the institution for the next tier of funding eligibility.