What is the average length of stay in ICE detention facilities in 2025?
Executive summary
Across 2025, public datasets and watchdog reporting converge on a middle range for the average length of stay in ICE custody — roughly between the mid‑40s and low‑50s of days — but that headline number masks substantial variation by facility, case type and month and rests on imperfect and sometimes competing data sources [1] [2] [3] [4] [5].
1. National averages point to roughly 44–50 days, depending on the source and month
Federal and sector analyses published during 2025 show a cluster of average‑stay estimates: the American Immigration Council reports an average of 47 days for FY2024 and notes that by July 2025 that figure had risen to about 50 days [1], while an advocacy analysis cited an average of 44 days as of September 2025 [2], and other compilers place the typical range between about 46 and 52 days across 2025 [3] [4].
2. Monthly and facility‑level fluctuation undercut a single definitive number
Journalistic and research reporting emphasize that “the average” obscures large month‑to‑month shifts in bookings and length of stay — ICE publishes monthly book‑ins and average length of stay but the datasets are incomplete and reporters warn the picture is imperfect [5]; Vera and other trackers document that ICE often reports only a fraction of the facilities it uses, which can distort national averages [6].
3. Local realities diverge dramatically from the national mean
County jails, tent processing sites and different contract prisons show wildly different averages: some New York county jails report stays as short as two days or about 33 days in others [7], an El Paso tent processing site registered an average stay of about five days because most people were transferred or deported rapidly [8], and certain contract facilities and solitary‑confinement populations have much longer averages — in solitary cohorts the average was recorded at 57 days with medians and maximums far higher [9] [10].
4. Why averages move: policy, capacity and reporting choices
Shifts in enforcement priorities, expansions of capacity, transfers between facilities, and the pace of immigration court processing all alter average stay calculations; advocacy and government sources note that increases in bookings or large transfers can raise the systemwide mean even if many individual stays remain short [5] [11]. Moreover, budgetary and legislative moves to expand detention capacity can change both daily population and average duration in short order [2] [11].
5. Methodological caveats: what “average” actually measures and what it omits
Different organizations use different time windows, include or exclude certain facility types, and publish snapshots versus rolling averages; some ICE facility listings lack average daily population values or report only limited metrics, which researchers warn can bias calculations toward facilities with better reporting [12] [6]. Consequently, the reported national average should be read as a working estimate rather than a precise, unambiguous fact [5].
6. Competing narratives and institutional incentives
Advocacy groups, watchdogs and government entities each have incentives that shape emphasis: immigrant‑rights groups highlight long stays and harmful conditions to argue against detention expansion [9] [11], while some government statements emphasize compliance with standards and oversight programs to justify operations and funding [10]. Those differing agendas help explain why media summaries and policy papers sometimes headline different averages from the same year [1] [2].
7. Bottom line for 2025
A defensible synthesis of the available 2025 reporting is that the average length of stay in ICE detention clustered roughly between 44 and 50 days, with many credible trackers centering the number in the high‑40s; however, this national figure masks dramatic variation across facilities and populations, and it is vulnerable to reporting gaps, differing methodologies and rapid policy‑driven changes in detention patterns [1] [2] [3] [4] [5].