Airbus A350-900

* Structural ATA chapters use FAA K-code change rate. Verb-based propensity is suppressed there because "REPAIRED" in the SDR text usually refers to the airframe being repaired around the part.

US carriers only (BTS T-100, domestic + international segments) — foreign-carrier flying is excluded, so global utilization runs higher. Fleet size is reconstructed from the FAA registry (built on or before each year, not yet deregistered) — an approximation. Freighter share counts departures with zero passengers and freight aboard — a proxy for freighter/combi operations, not a tail-by-tail conversion count. Missing years render as gaps.

FAA registry data. Domestic deregistration is a teardown proxy — it also captures re-registrations and some unflagged exports, so it is not a confirmed part-out count; exported aircraft left the US fleet intact and are not USM supply. ATA shares reflect where this directory's parts for the family concentrate (parts in parentheses) — a coverage signal, not the aircraft's bill of materials or a teardown-yield forecast.

Engines account for roughly half of all MRO spend, so engine programs shedding aircraft are where retirement supply carries the most value.

FAA registry data, US-registered aircraft only. Counts reflect the engine model as registered — generic “series” rows coexist with thrust-variant rows, so per-variant figures are partial. Retired = domestic deregistrations (a teardown proxy, not a confirmed part-out); exported aircraft left the US fleet intact. Active tails span every family the engine flies on, not just this one.

BTS Form 41 data (Schedule P-5.2 maintenance expense over T-2 block hours), Group III US carriers only — filers above $1B annual revenue; smaller US operators, Part 135, and all non-US carriers are not in this data. Dollars are accrual-basis from regulatory filings (reserves and depreciation included), so they benchmark fleet economics and do not track to individual repair events. Averages are block-hour- weighted across every reporting carrier; the range spans per-carrier rates after excluding marginal reporting slices, and small carrier counts are noisy.