How have ICE detention counts for non‑criminal detainees changed from 2024 to 2026 and what datasets track that trend?
Executive summary
ICE detention of people without criminal convictions rose sharply from late 2024 into 2026, driving most of the agency’s net increase in detained populations, and multiple independent and official datasets — including ICE’s own dashboards, the Deportation Data Project, TRAC, and academic analyses from UCLA and others — document that shift [1] [2] [3] [4] [5]. Analysts disagree about precise shares and definitions, but the consistent signal across sources is a major growth in detainees who lack recorded criminal convictions during this period [6] [7].
1. Growth concentrated in detainees with no criminal convictions
Public reporting and independent trackers show that most of the increase in detained people since late 2024 was among those with no criminal conviction: one analyst found that 72 percent of the FY2026 detention growth was non‑criminal detainees, with only 8 percent attributable to people with criminal convictions [5], while broader summaries from TRAC and Cato corroborate that a large share of the detained population in late 2025 and early 2026 had no conviction on their record [8] [6].
2. Timeline: late 2024 baseline, big increases through 2025 into early 2026
ICE’s dashboards published through December 31, 2024 provide the public baseline and, thereafter, the agency continued to publish snapshots that researchers have used to compare book‑in totals and custody counts; those snapshots show a noticeable jump from September–November 2025 into the early months of FY2026 [1] [9]. Reporting using ICE and leaked data indicates ICE’s detained population crossed roughly 65,000 in November 2025 — substantially higher than the September 2025 baseline — and continued to rise into January 2026, with non‑criminal detainees powering much of that growth [9] [5].
3. Who is counted as “non‑criminal” and why definitions matter
Different trackers use different operational definitions: ICE’s public dashboards classify detainees by conviction status and by pending charges (the agency’s documentation frames some categories as “criminal” even when charges remain pending), while independent projects and researchers often separate “convicted,” “pending charge,” and “no criminal charge/conviction” to expose enforcement emphasis [1] [10] [6]. FactCheck and other analysts note that shifts in how ICE labels “criminal” versus “pending” can materially change headline percentages, so comparisons across months require careful attention to field definitions and to whether the datasets include CBP arrests or only ICE administrative arrests [7] [6].
4. Primary datasets and trackers to follow this trend
The clearest primary sources are ICE’s own Enforcement and Removal Operations (ERO) dashboards and detention statistics, which for the first time publicly posted time‑series dashboards through December 31, 2024 and continued updates thereafter [1]. Independent, researcher‑curated datasets include the Deportation Data Project, which publishes detailed ICE FOIA data on arrests, detainers, detentions and removals and documents table schemas for longitudinal merges [2] [10], and TRAC Immigration, which maintains up‑to‑date detention population tables and quick‑fact summaries used widely by journalists and scholars [3] [11]. Analysts and think tanks — UCLA Luskin, Cato, and reporters such as Austin Kocher — have repurposed those sources to quantify the rise among non‑criminal detainees and to visualize trends [4] [6] [5].
5. Converging evidence, methodological disputes, and blind spots
Multiple independent analyses converge on a sharp shift toward detaining people without convictions, but they differ on magnitudes and causal framing: UCLA’s analysis flags a sharp rise in non‑criminal Latino detainees and longer detention spells [4], while Cato and TRAC emphasize the changing composition of arrests and the small share of violent convictions in the detained pool [6] [8]. FactCheck highlights that some headline percentages are sensitive to baseline choices and to ICE’s labeling conventions [7]. Open data gaps remain on granular offense types for those with pending charges, on the contribution of CBP transfers versus ICE administrative arrests, and on facility‑level practices — gaps the Deportation Data Project and FOIA releases partially address but do not fully close [2] [10].
6. What to watch next: data releases and policy signals
Monitoring future ICE ERO dashboard updates, continued FOIA‑based data dumps curated by the Deportation Data Project, and TRAC’s running tables will be the best way to track whether the early‑2026 spike in non‑criminal detainees stabilizes or continues to rise [1] [2] [3]. Parallel reporting on detention capacity expansion — including interest in warehouses reported by The Washington Post — and detention‑related mortality and medical incidents remain relevant context for the human and policy stakes of these numerical trends [12] [13].