How many U.S. citizens were detained by ICE in recent years (annual breakdown)?

1. What the official numbers show — steady climb through FY2022–FY2025

ICE’s public releases and analyses compiled by TRAC show a clear year‑to‑year increase in people held in ICE custody: 25,134 at the end of FY2022, 32,743 by end of FY2023, “over 37,000” by end of FY2024, and 39,703 on January 12, 2025 ^[1]. ICE posts detention and book‑in statistics on its statistics and detention‑management pages that researchers and journalists use for biweekly snapshots ^{[3] [5]}.

2. Mid‑2025 surge and competing peak estimates

Several outlets and trackers reported much higher mid‑2025 counts. TRAC and advocacy groups cited biweekly ICE snapshots showing nearly 60,000 people in detention as of late September 2025 and similar mid‑year tallies ^{[2] [6] [7]}. Some organizations and commentary referenced even larger peaks—one site flagged a 66,000 figure as a record high in 2025—though that comes from a commentary/blog source rather than ICE’s official spreadsheet ^{[8] [2]}.

3. Why published totals differ — methodology and undercount risks

The Government Accountability Office has found that ICE’s publicly reported detentions understate the total because ICE excludes certain temporary bookings and transfers in its “initial book‑ins” methodology; GAO says that exclusion can amount to “tens of thousands” of individuals ^[9]. That methodological gap helps explain why independent reconstructions (TRAC, Vera, Deportation Data Project) sometimes report different annual or cumulative totals than ICE’s headline numbers ^{[9] [10] [11]}.

4. Daily average vs. cumulative book‑ins — different metrics, different stories

Reporting alternates between average daily population (how many people are in custody on a given day) and cumulative “book‑ins” or arrests during periods. ICE’s Detention Management spreadsheets provide average daily population snapshots that journalists use; other datasets track arrests/book‑ins, which can be much larger because many people cycle through custody ^{[5] [3] [6]}. The Marshall Project noted ICE and CBP had booked more than 310,000 people into custody since October 2024 in a data release referenced in their reporting ^[7].

5. Criminality breakdown and policy framing

Sources note much of the detained population lacks criminal convictions: TRAC reported roughly 71–72% of people in ICE detention in mid‑2025 had no criminal conviction, a point emphasized by advocates and researchers ^{[2] [12]}. The Trump administration’s enforcement priorities and increased capacity in 2025 were cited by multiple outlets as drivers of the rise in detention numbers ^{[4] [10] [13]}.

6. Data access, transparency and recent reporting interruptions

Journalists and researchers warn that gaps in public reporting—caused at times by agency shutdowns or delayed releases—make trendlines harder to pin down. The Marshall Project reported that during an information blackout some enforcement numbers and facility daily counts were temporarily unavailable, complicating continuous tracking ^[7]. GAO recommended ICE improve reporting and explain exclusions so external analysts can reconcile totals ^[9].

7. How to interpret “how many U.S. citizens were detained” specifically

Available sources in this collection discuss total detainee counts and citizenship breakdowns for ICE book‑ins generally, but they do not provide an explicit, sourced annual breakdown limited only to U.S. citizens detained by ICE. The Deportation Data Project, TRAC, and ICE tables include citizenship fields for arrests and book‑ins, but the current set of summaries here does not list year‑by‑year counts of U.S. citizens specifically ^{[11] [3] [1]}. Therefore, “annual breakdown of U.S. citizens detained” is not found in the current reporting provided.

8. What a careful analyst should do next

To produce an annual U.S.‑citizen‑only breakdown you should (a) download ICE ERO book‑in tables and monthly tables that include “citizenship” fields from ICE and OHSS, (b) use TRAC and Deportation Data Project exports to reconcile differences, and (c) apply GAO’s caution about excluded temporary bookings when aggregating totals ^{[3] [6] [9] [11]}. Those primary tables are referenced in the sources above as the authoritative raw materials for such disaggregation ^{[3] [6]}.

Limitations: reporting and sources here mix average daily population, point‑in‑time snapshots, and cumulative book‑ins; GAO found ICE’s public methodology omits some detentions; and none of the provided summaries gives a ready annual table limited to U.S. citizens only ^{[9] [3] [1]}.