How does grant termination convert to data-maintenance termination?

A subsidiary question under Q-0019. §3.3 documents the funding-termination mechanism in the four-category §3 taxonomy.

The answer is C-0019: when grants end, maintenance ends with them. Data infrastructure paid for by a project budget has the lifespan of that project budget — three to five years, after which the server contract lapses, the storage migrates ad hoc or not at all, and the dataset enters the failure modes catalogued in §3.1 (personnel) and Tier 1 (platform).

The 2025 NIH and NSF terminations are the cleanest available demonstration: 2,291 NIH grants ($2.45B) and 1,752 NSF grants ($1.4B) terminated between February and August 2025; the NSF STEM Education Directorate alone lost 839 grants worth $888M; FY2026 proposed cuts of ~56% to NSF, ~24% to NOAA, and ~57% to ARPA-E. Specific data infrastructure disappeared alongside the grants — NOAA's Billion-Dollar Disasters database, Mauna Loa's 68-year CO₂ record proposed for complete defunding, ~3,400 datasets removed from Data.gov, 191 of 411 long-term mammal studies terminated.

Every long-term dataset is a cumulative capital asset. When the funding stops, the asset is abandoned, and re-collection is rarely possible because the original ecological, cohort, or instrumental conditions cannot be reassembled.