That’s why the team at Relevant Research built the ICE Detention Mortality Dashboard: to put what does exist in one searchable place, with case-level records, preserved source materials, and an interactive map.

When we first released the dashboard, the data only went back to 2018. But we have since incorporated prior data available in the public domain to provide a more comprehensive, historic view that captures the pervasive consequences of civil detention rather than any one administration. Between October 7, 2003 and August 1, 2026, at least 289 people died in ICE detention. There’s no way around it, though. The numbers have grown under Trump 2: 51 of those deaths came after January 20, 2025, or roughly one every eleven days across the first eighteen months of the current administration.

On Monday, August 31, Adam Sawyer and I are hosting an open conversation about this work because I want folks to understand the decisions that we made as we built the dashboard, the challenges of working with this data, and how we can use this data responsibly to tell human stories about the individual lives and the collective cost of a discretionary detention system. The conversation is on Monday, August 31 at 3:00 pm Eastern. Please join us, and bring your questions.

REGISTER NOW

Here is some of what we will get into during the conversation.

• Why these deaths are documented at all, and what happens to accountability when the only account of a death is the one the agency chooses to publish.

• How the dashboard was built, including where the records come from, what the counting rules are, and why an enforcement death and a detention death are kept apart.

• A walkthrough of the resource itself, so you can see what is in it and how to search it.

• What the data may support and where it can be read too confidently, including the deaths that may be missing from it entirely and the fields that are only partly filled.

• How other researchers, journalists, and advocates are putting these numbers alongside their own work, and where our count agrees and disagrees with theirs.

Adam Sawyer is the Director of Research at Relevant Research, where he co-developed the Interval ADP method for measuring detention populations. He previously did data research for the Transactional Records Access Clearinghouse and for the International Organization for Migration.

Do you have questions for Adam and me? Leave them in the comments below and I’ll do my best to bring them into the conversation.

Leave a comment

This newsletter, and live conversations like this one, are only possible because of reader support. If you are able, a paid subscription helps me keep this kind of documentation going and keep it public.

Subscribe now

Subscribe now

The Data on ICE’s Summer Enforcement Surge

A quick reminder that the fall of 2025 already saw a surge in arrests, with an average of 36,348 arrests per month from September through January. This previous surge was characterized by aggressive, targeted operations that drew criticism and on-the-ground observation and counter-protests. These tactics also led to the death of two U.S. citizens and corresponding outrage, even among many Trump supporters. DHS Secretary Kristi Noem was replaced by Markwayne Mullin in March. The number of ICE arrests and detention numbers increased again only modestly between February and May but remained around 30,000–32,000, below the fall 2025 surge.

That all changed in June and July. June set a recent record of 42,637 arrests and July set a new record again with 49,138 arrests in a single month.2 I had previously estimated that ICE arrested about 1,300 per day for the month of June and 1,474 for the first 11 days of the month of July based on the most recent ICE detention management spreadsheet. DDP’s data allows us to be more precise. ICE arrested in 1,421 in June, 1,585 in July, and 1,469 for the first five full days of data in August. The increase from May to June this summer was one of the largest month-over-month jumps in arrests, making the surge visible in the data, even though ICE’s current tactics have been less visible than the chaos in Minneapolis at the start of the year.

Consistent with the Trump administration’s well-documented pattern of misrepresenting immigrants as threats to national security and public safety, the DHS propaganda machine spent the summer pushing a false narrative that the surge in arrests focused on the worst of the worst. DHS press releases highlight a small number of serious criminals, hoping that selective misrepresentation will give the impression that all immigrants are violent offenders. The Trump administration’s claims go on.

Just days ago, DHS told KSL: “From day one, DHS has remained singularly focused on removing the worst of the worst criminal illegal aliens from the United States and is always evaluating the best methods to do so. These heinous criminals, once arrested, should be removed at lightning speed, not housed on American soil at the taxpayer’s expense.”

An August 10 DHS press release quotes Secretary Mullin as saying: “While Americans enjoyed their weekend, the patriots of ICE worked around the clock to remove murderers, drunk drivers, pedophiles, sexual predators, drug traffickers, and other public safety threats from our communities.”

DHS Secretary Mullin’s claims are emphatically untrue. This is not a political opinion, this is simply what the data say—there’s no way around it.

The ICE summer surge in arrests—to record numbers, let’s not forget—was driven almost entirely by a focus on people with no criminal history; i.e., neither a pending charge nor a conviction. June and July, the last two full months on record, saw the highest percentage of people arrested without any criminal history going back to the start of this administration. In June, it was about 47% of all arrests and in July it was over 50%. Half of all people arrested have no criminal history on record whatsoever.

A declining percentage of people had charges or convictions and, in fact, just 1 in 5 arrests as of the limited August data had a conviction—the lowest in the dataset so far. Remember that very few charges and convictions are serious violent offenses; see my careful analysis of Delaney Hall.

If you are newer to this Substack, you might benefit from going back and looking at my prediction of this trend at the start of the Trump administration. This is exactly what the research said would happen.

The story continues in total numbers. Although there was an increase of a little less than 20% for people with charges and convictions, people with no criminal history increased over 100% (more than doubled). This is not what it looks like for ICE to go after the “worst of the worst.” This is what it looks like for ICE to round people up who represent no public safety threat as part of a misinformation campaign to do one thing and spin it as something else.

There is more to be said about this data, including what is and is not in the data this go-around, and how we can use data to understand the broader trends in immigration enforcement. But this post should at least give you an initial look at the data. Don’t forget to subscribe and consider becoming a paid subscriber so this work can continue.

Subscribe now

Join the Deportation Data Project Webinar

Great news! —> The Deportation Data Project is holding a webinar on this release on Wednesday, August 26 at 1:00 PM Eastern. Anyone planning to work with these files should hear how the people who obtained them describe what changed. Register below.

Register for DDP's Webinar

Leave a comment

You can keep up with the data between editions on Bluesky, X, Instagram, Threads, LinkedIn, Mastodon, and Facebook.

Share

1,000,000

The number of Temporary Protected Status holders whose protection this administration has ended or moved to end, out of roughly 1.3 million who had it.

Pew counted 1,044,700 people in that position as of August 12, measured against the 1.3 million who held TPS in March 2025, the last month with public data. The tally comes from the beneficiary estimates the Department of Homeland Security publishes in its own Federal Register termination notices, covering thirteen designations. Venezuela accounts for 616,400 of them across two designations, and Haiti for 334,900. Four countries still hold a designation that is not being wound down, El Salvador, Lebanon, Sudan and Ukraine, covering about 273,200 people. Two others, Ethiopia and Somalia, were protected only by court orders, and district judges dissolved both in the week after Pew published. Losing the status strips the work permit and the protection from removal on the same day, and it orders nobody deported by itself, so what it produces first is a large population living and working without authorization.

• Im, C., & Hennen, M. (2026, August 12). Second Trump administration has ended Temporary Protected Status for most immigrants who had it. Pew Research Center.

60%

The top of the sales decline one Brooklyn restaurant owner reports since Temporary Protected Status for Haiti ended.

Wesly Simon, who owns Djon Djon in Brooklyn’s Little Caribbean neighborhood, told Documented that sales have fallen 40 to 60 percent and that the decline is accelerating, after he lost about half his staff when the Haitian designation ended in late July. Those percentages are his own estimate and not an audited figure. A second Haitian restaurant owner described the same problem to the same reporter. The Supreme Court cleared the way for the termination in June, holding that the Homeland Security secretary’s judgment about ending a designation is not subject to judicial review on non-constitutional grounds. A restaurant that loses half its kitchen cannot replace it from the same neighborhood, because the neighborhood is where the terminated status was concentrated.

• Thomassaint Joseph, R. (2026, August 10). New York Restaurants Struggle to Stay Open as Haiti TPS Ends. Documented.

23,000

The number of people the United States has sent to countries that are not their own since January 2025.

Human Rights First and Refugees International count more than 23,000 forced transfers to 26 third countries between January 20, 2025 and August 15, 2026, assembled from transparency requests, on-the-ground sources, and public reporting. Nearly 20,000 of those went to Mexico, with more than 2,700 to other countries in Latin America and the Caribbean, more than 330 to countries in Africa, and more than 200 elsewhere. The transfers run on more than 35 bilateral agreements signed since January 2025, most of which have never been published, and the two organizations have traced $44 million in U.S. funding attached to them. Liberia agreed this week to accept up to 1,200 people, and the first twenty arrived on August 20. The count leaves out removals that merely stop over in a third country and ad hoc transfers on commercial flights, so this is a floor.

• Human Rights First & Refugees International. (2026, August 18). Human Rights First and Refugees International Third Country Deportation Watch Details the Expanding Scope and Toll of Third Country Transfer Agreements. Human Rights First.

3,000

The top of the bed range the government is asking contractors to be ready to supply for detained migrant children.

The Office of Refugee Resettlement told vendors at a market research roundtable in mid-August that it wants emergency influx facilities, sited anywhere in the country, holding between 100 and 3,000 beds, with individual task orders likely covering about 1,000 each. No site has been chosen and no contractor named. The agency aims to award by December 31 and to have facilities operating within 90 days of the award. Reported design details include an eight-foot fence and rooms holding eight children who share one toilet, shower and sink. The plan surfaced a day after Judge Dolly Gee appointed a special master and an independent medical monitor over compliance with the Flores settlement, having found that the average detention of children exceeded twenty days for five consecutive months, with individual stays of 50, 91 and 323 days.

• Billings, A. (2026, August 21). Trump administration plans new facility to house 3,000 migrant kids. Many of them lack legal representation. Reason.

38

The number of days Ana Kikabidze’s two children had been detained when she gave up their case and accepted deportation.

Kikabidze and her children, aged 11 and 14, held green cards the family won in the 2013 visa lottery. She was detained on arrival at Chicago O’Hare in May after 229 days abroad, roughly six weeks past the limit on time away, and the three of them were moved to the Dilley facility in Texas on May 17. An immigration judge ordered them removed during a video call and told her that appealing would mean months more in custody. She had spoken with her lawyer once, by phone, the day before that hearing. She withdrew the appeal when the children had been held 38 days, and the family was deported on July 13, 58 days after they were taken in. ICE returned their unexpired green cards on the way out. More than 66,000 parents and children have been detained and deported since 2025, and in July more than half the children at Dilley had been held past the twenty days the Flores settlement allows even during a surge.

• Villagran, L. (2026, August 18). Parents are giving up green card fights to get their kids out of immigrant jail. The Marshall Project.

132,167

The number of asylum applications USCIS expects to send to immigration court each year without ever interviewing the applicant.

A rule published on July 28 allows asylum officers to refer an affirmative application straight to a judge with no interview in four situations, when the applicant is barred from applying for asylum, barred from a grant, does not merit a grant as a matter of discretion, or is not eligible on the merits. It also removes the regulatory language describing an interview as a right. USCIS estimates 132,167 such referrals a year going forward and says as many as 444,724 pending applications may be affected. On August 19, four organizations sued in federal district court in Washington to vacate the rule, arguing that the department issued it without notice and comment, and no ruling has issued. The interview is where a person explains a late filing or a gap in the record, and removing it moves that explanation into a court where most cases completed in June ended with the person not in the room.

• Orozco, A. (2026, August 5). New USCIS Asylum Interview Rule Could Speed Up Deportations of Asylum Seekers. American Immigration Council.

• Center for Gender & Refugee Studies. (2026, August 19). Immigrant Rights Groups Sue to Protect Due Process for Asylum Seekers. Center for Gender & Refugee Studies.

78,882

The number of removal orders immigration judges issued in June, the highest monthly total in TRAC’s records.

Judges issued 78,882 removal orders in June, up from 58,339 in May after three months of flat numbers, with voluntary departure orders adding another 7,689. Total case completions reached 100,773, the first month above 100,000. About 68 percent of June’s removal orders were entered in absentia, meaning the person was not present when the case ended, up from 53 percent in March. Removal orders accounted for roughly 78 percent of all completed cases that month. The in-absentia share rose fifteen points between March and June, so most of what this record month added was orders entered against people who were not in the room.

• TRAC. (2026, July 21). Removal Orders Jump While Immigration Court Closures Top 100,000 in June 2026. Transactional Records Access Clearinghouse.

$2.5B

What ICE obligated in contract awards in July, its largest month of the year.

Roughly $1.5 billion of that went to buying two CoreCivic detention centers outright, the 2,560-bed California City facility in Kern County for $732.6 million and the 1,994-bed Otay Mesa center in San Diego for $739.2 million. CoreCivic keeps operating both under its existing management contracts, through August 2027 at California City and December 2029 at Otay Mesa. New GEO Group contracts for sites in Colorado and North Carolina add more than 2,500 beds between them. The agency also obligated more than $30 million on body cameras, following a $20 million appropriation in April. Owning a building the seller continues to run removes the annual question of whether to renew a lease and leaves only the operating contract to renew.

• Knepp, E., Wriston, M., & m. (2026, August 14). Follow the Money - July 2026: ICE Spends Year-High $2.5 Billion, Driven by Detention Expansion. Project Salt Box.

$250

What ICE proposes to pay each local officer per year so they can carry liability insurance for making immigration arrests.

A planning document published on August 14 outlines reimbursing participating officers up to $250 annually, roughly the expected cost of a policy covering up to $500,000 in personal liability. Officers would buy the policies themselves, and a contractor hired by ICE would handle outreach, training and reimbursement. The program attaches to 287(g), the partnership under which nearly 1,600 agencies across 32 states now carry out immigration enforcement, and where arrests have run about 3,000 a month in early 2026 against roughly 250 a month in 2024. David Bier of the Cato Institute said ICE is going above and beyond to guarantee law enforcement does not have even the slightest risk of liability for violating Americans’ rights while helping ICE arrest people.

• Foley, R. J. (2026, August 17). ICE pitches legal insurance to help shield local officers who make immigration arrests. Associated Press.

Support This Work

I’m honored that this Substack continues to drive media reporting, research, and public debate, but it does take effort and expense. If you can pitch in a few bucks a month or year, it would help enormously. Find an option that fits your budget at austinkocher.substack.com/subscribe.

Subscribe now

About a thousand Haitians in his district have no status now, he said, and ten to fifteen thousand live an hour away in Springfield. Many of them are in Ohio because congregations in his region spent fifteen or twenty years building relationships with people in Haiti through service trips and many then sponsored the people who fled. “That’s who brought most of these folks here,” he said, adding that this is not merely an historical footnote, but a practical problem for the administration. “In some ways they’re going to have to go through our churches,” he said, referring to ICE’s looming Haitian deportation push.

At Blue Ash every Sunday, hundreds of people report to ICE to be fitted with ankle monitors. While some protest in the street, clergy have gained permission to pray in the doorway of the ICE facility. Landsman described families tracking rumors by the hour about roving patrols and deportation flights, asking where else they could go, because they cannot go back to Haiti.

He also said he has not met any genuine opposition to allowing Haitians to stay in the country. “I’ve not met somebody who’s like, when do they leave? They’re ruining my lives. Nobody thinks that,” the congressman told me. He pointed to Gov. Mike DeWine, who told him that deporting these families means “sending them back to hell,” and to Springfield’s Republican mayor, who has described it to him also as a matter of dignity for their neighbors.

The resistance, as Landsman sees it, is “entirely driven by Stephen Miller.” Republicans have told him privately that the extension needed to happen and then declined to put their names on it, because signing means picking a fight with President Trump.

When I asked how to connect across political lines, he said: "The arguments that I think are most compelling on this issue with Republicans is the faith-based arguments that churches have spent so much time there. They've sponsored so many of these folks. They're so heavily involved that it is a dignity issue. Our faith is calling us to take care of these folks, not kick them out."

You can listen to the entire conversation above, which includes more information about how Congressman Landsman’s office is responding to concerns from the community and why clergy and people of faith can play an important role on this issue. I’ve include more information below about what you can do to take action.

My thanks to Congressman Landsman for meeting with me and to his staff for their assistance in coordinating. If you are an elected representative and you want to talk with me live about how immigration is impacting your constituents and what you’re doing to create change, reach out to discuss the possibility of joining me live.

Share

What can do this week?

H.R. 1689 passed the House on April 16 by 224 to 204, with ten Republicans voting yes, and it would designate Haiti for TPS through April 2029. It has been on the Senate calendar since April, and Sen. Ed Markey is the point of contact on the Senate version. Landsman’s advice was to email your senators first, because email is easier to organize and it registers in a member’s office the same way a call does, and then to call. Contact forms and phone numbers for both of your senators are at senate.gov, and the Capitol switchboard is (202) 224-3121.

Follow Congressman Landsman online for information about his office:

• Website: https://landsman.house.gov

• X: https://x.com/RepGregLandsman

• Bluesky: https://bsky.app/profile/repgreglandsman.bsky.social

• Instagram: https://www.instagram.com/repgreglandsman/

• Facebook: https://www.facebook.com/RepGregLandsman/

Greg’s Substack

My thoughts on what’s happening in Congress, across the country, and here at home in Southwest Ohio.

By Congressman Greg Landsman

Help Keep These Conversations Going

I’m honored that this Substack continues to drive media reporting, research, and public debate, but it does take effort and expense. If you can pitch in a few bucks a month or year, it would help enormously. Find an option that fits your budget at austinkocher.substack.com/subscribe.

Subscribe now

• Who: Live discussion with Representative Greg Landsman (D-OH-01)

• What: How the end of Haitian TPS hurts Ohio and what Congress can do about it. See HR 1689.

• When: Wednesday August 19, 2:00 pm Eastern

• Where: Substack Live (subscribe to get notified when we start)

Subscribe now

Do you have questions for Rep. Landsman? Leave them in the comments below and I’ll do my best to bring them into the conversation.

Leave a comment

Greg’s Substack

My thoughts on what’s happening in Congress, across the country, and here at home in Southwest Ohio.

By Congressman Greg Landsman

Background to My Discussion with Representative Landsman

In the run-up to the 2024 election, Springfield, Ohio—a town just a few miles from where I grew up—became synonymous with a claim by Donald Trump that was as absurd as it was dehumanizing: “they’re eating the dogs,” he falsely claimed of Haitian residents who had played a crucial role in revitalizing the Rust Belt city.

I publicly stood up at the time for Springfield as much as for Haitians. I was personally offended at the attempt to stir up artificial antagonisms in a state with plenty of real ones, in a way that masked the shared working-class hospitality that I witnessed among immigrants and citizens in my home state. I was reminded then and now of Ohio author David Giffels, whose book The Hard Way On Purpose, a collection of essays about the struggle for survival and dignity in the Midwest, captures the story of all of us who call Ohio home—no matter where we came from.

Now, a little more than 18 months into the current administration, Haitians are not merely suffering from grave misrepresentation—they are losing their legal status. Somewhere north of 300,000 Haitians who were granted Temporary Protected Status (TPS) have had, or will soon have, that status revoked, and with it, both their ability to work and their protection from deportation.

The nation’s eyes are back on Springfield as Haitians are forced to report to ICE field offices to get fitted with ankle shackles. They surrender their passports at the door and go home wearing devices that leave lesions and blisters on their ankles, that beep them awake in the middle of the night, and that tell everyone who sees them that these people are being tracked. This is being done to nurses’ aides and line workers and small business owners who have spent years building lives in a city they helped bring back from the edge.

Opposition to the termination of TPS does not fall strictly along party lines. Republican Governor Mike DeWine—not a radical leftist but an old-school Republican with a state to run—has opposed the termination of TPS. He said that sending people to Haiti would be tantamount to sending people “back to hell” and emphasized the contributions of Haitians to the regional economy:

“I think it is a mistake. I think that the people who come here to work, the Haitians who came to Springfield, were filling jobs that had not been filled before. They are part of the resurgence of Springfield.” — Mike DeWine, Ohio Statehouse News Bureau, July 29, 2026

Haitians can continue to be a part of the resurgence in Ohio, but doing so now requires an act of Congress. That’s where Congressman Landsman comes in. Landsman, whose district is home to the ICE field office in Blue Ash where Haitians are required to report, has been an outspoken opponent of the termination of TPS. In July, Landsman cited the cruelty and shortsightedness of putting hundreds of thousands of Haitians into the deportation process and called on his fellow Congresspeople to take action before it is too late.

Landsman, along with a bipartisan coalition that included two House Republicans from the great state of Ohio, passed HR 1689, which would require DHS to designate TPS for Haitians until April 2029. The bill passed the House on April 16 (224 to 204), with ten Republicans voting yes. Rep. Greg Landsman signed the discharge petition that forced the House vote. The legislation is a reminder that Congress, not only the Executive, has the power to act on immigration matters—and that it is up to the public to exercise its power to decide whether people in office are doing the job that Americans want them to do.

I am grateful that Representative Landsman will be joining me to talk about what he is doing to protect Haitian TPS holders in Ohio and what he and his staff are hearing from impacted constituents. Please join us (or listen later) as we discuss:

1. What are Haitian families and their employers in Ohio actually facing right now?

2. Why has the Senate not taken up HR 1689 since April, and what would it take to move it?

3. What is Rep. Landsman’s office hearing from their constituents?

4. What did it take to force the vote in the House, and what do those ten Republican votes tell us about what is still possible?

5. What can people do right now, and should people be contacting their senators?

Do you have questions for Rep. Landsman? Leave them in the comments below and I’ll do my best to bring them into the conversation.

Leave a comment

This newsletter, and live conversations like this one, are only possible because of reader support. If you believe in keeping this work free and open to the public, please subscribe and share it with others.

Subscribe now

Further Reading

• H.R. 1689, the bill to designate Haiti for Temporary Protected Status, 119th Congress, introduced by Rep. Laura Gillen (D-NY-04) on February 27, 2025. The bill text, the sponsors, and where it currently stands in the Senate.

• Roll Call 120, Office of the Clerk, U.S. House of Representatives, April 16, 2026, the full 224 to 204 vote with every member listed by name.

• Discharge Petition No. 15, Office of the Clerk, filed by Rep. Ayanna Pressley, the procedure that forced the bill onto the floor, and the members who signed it.

• Office of Rep. Greg Landsman, Landsman Statement on Trump Deporting Tens of Thousands of Haitians, July 31, 2026, his account of what the terminations do to Ohio families and his call for the Senate to act.

• Megan Henry, Haitians in Springfield are required to report to Ohio’s ICE offices and wear ankle monitors, Ohio Capital Journal, August 14, 2026, on the check-ins, the ankle monitors, and the passports collected at the door.

• Isabel Nissley and Tana Weingartner, Hundreds show support for Haitians attending ICE meetings in Cincinnati, WVXU, August 2, 2026, on the gathering outside the Blue Ash field office.

• With TPS now expired, DeWine and advocates for immigrants await fate of Haitians in Ohio, Ohio Statehouse News Bureau, July 29, 2026, the source of the governor’s remarks quoted above.

• Ohio’s GOP governor calls Trump’s plans to deport Haitians a mistake, CBS News, July 28, 2026, where DeWine said that sending people to Haiti would be sending them back to hell.

• Mullin v. Doe, Nos. 25-1083 and 25-1084 (U.S. June 25, 2026), the Supreme Court decision holding that federal courts may not hear most challenges to a TPS termination, which is why Congress is now the only route.

On July 8 the family drove nine hours from Fort Polk, Louisiana, to Montgomery, Alabama, for what they had been told was a routine interview to finalize Maisa’s green card. Their paperwork had been filed in Alabama, where Alexis was stationed before. Maisa was taken out of the room partway through. An officer brought Alexis into a conference room, with their five-year-old son standing next to him, and told him his wife was going to jail.

“They use these words: ‘Your spouse is going to jail,’” Alexis said. “I was in uniform. They didn’t care about it.”

Alexis has served eleven years as an aviation operations specialist and is now a senior instructor. He spent the next five weeks parenting a five-year-old alone while remaining on active duty. He said a peer and a superior are the reason he decided to speak publicly rather than handle it quietly.

Maisa spoke briefly in Spanish, with Danitza translating. She said she is happy and feels safe at home, and that she is still afraid to go out, because anyone she does not recognize might be ICE. She thanked people who supported her without knowing her or her case. The couple has an ICE check-in on Monday, August 17, and both of them said they are afraid of what that appointment will bring.

Danitza said the trauma does not end when the detention does. Every immigration appointment from here forward will be a reminder, and that is true of every family going through this. “Alexis is still my brother. Maisa is still my military family sister. I didn’t know you. I have never met you. But you were still my military family sister. This is who we are. This is what we sign up for,” she said.

More than 1.5 million U.S. citizens are married to someone who could face exactly what Maisa faced, Ashley said. Her organization’s inbox fills daily with those families, and she said resolving them one press conference at a time is not a response to a problem that size. An Associated Press investigation counted at least 51 spouses and parents of active-duty service members detained since January 2025, and everyone on the call agreed the real number is higher. She asked the administration to stop detaining non-criminal immigrants married to U.S. citizens while Congress works this out, and asked Congress to move the bills that have been sitting in committee for years.

Repatriate Our Patriots is working other active cases right now, most of them active-duty service members whose spouse or parent is detained. She said the scale of military family detention is unprecedented in her years doing this, and that the military has no standard operating procedure, guide, or memo telling commanders how to support a soldier in this situation. What has helped in the cases she tracks is individual commanders and first sergeants deciding on their own to connect their soldiers to organizations like hers and to congressional offices. That is improvisation by people who happen to care, in place of a system.

Maisa is home, and everyone who had a hand in that decision did the right thing. Her case is not resolved, and many other families are waiting for their version of this outcome. Both organizations asked for her case to be reopened and for the system that put her there to be fixed.

Press contact information

• Repatriate Our Patriots: rop@repatriateourpatriots.org

• American Families United: press@americanfamiliesunited.org

About the organizations

• Repatriate Our Patriots is a veteran-led nonprofit based in El Paso that supports veterans and military families facing deportation or detention. They are on Instagram at @repatriate_our_patriots.

• American Families United advocates for U.S. citizens whose spouses have no reasonable path to legal status, and published Collateral Damage earlier this year documenting the pattern nationally. Ashley is @AFUAshley on Instagram, X and Bluesky.

Press release available below

Background

To learn more, you can listen to “Why Is ICE Deporting the Families of Active-Duty Military?,” my conversation with Danitza James last week about what the AP investigation found.

Support the families

Legal fees in these cases are covered by neither the military nor insurance, and the families pay out of pocket. Repatriate Our Patriots is collecting for them here: https://www.paypal.com/ncp/payment/U8LKJ7YMYAYBG.

Subscribe now

Maisa’s release was announced today in a press release from Repatriate Our Patriots and American Families United. She was arrested in early July at a routine immigration appointment, in front of Alexis and their five-year-old son, Noah, who is a U.S. citizen. Alexis, an aviation specialist stationed at Fort Polk, Louisiana, has served in the Army for eleven years.

After Maisa was detained, Repatriate Our Patriots, an organization that supports active-duty service members, veterans, and the families affected by the immigration system, worked together with American Families United to raise awareness about the case and to advocate for her release so that she could continue going through the immigration process with her husband at home.

I spoke with Danitza James about this case last Friday, and she explained that what happened to Alexis and Maisa follows a pattern she has been tracking for years. Reporting by the Associated Press confirms what Danitza has seen across both Republican and Democratic administrations, which is that many military families get swept up in our complex and broken immigration system. Many service members who enlist with the promise of citizenship are never able to obtain it, and some veterans are deported after their service to this country.

American Families United, an organization that supports U.S. citizens married to immigrant spouses at risk of removal, has focused on the damaging consequences for families, often unintended, when immigration law does not provide a reasonable path to citizenship for their spouse.

The administration’s decision to release Maisa is the right one, and I commend everyone who played a role in that decision for doing what is right. There are many more cases that deserve the same outcome. I would encourage citizens and politicians across the political spectrum to do what is right for active-duty service members, for veterans, and for the spouses of U.S. citizens.

Tomorrow we will hear directly from Alexis and Maisa about what this family has been through, and from Danitza and Ashley about how their case fits into the larger pattern that both organizations are seeing across the country. You can register for the press conference below.

REGISTER NOW

Please share the image below widely.

Find a copy of the official press release below.

Subscribe now

Click here to see all of my posts about ICE’s detention deaths.

Like so much of my work, writing about these tragedies requires rigorous behind-the-scenes information management to ensure that I’m working with a validated data set of cases with preserved links, archived materials, and carefully structured metadata that facilitates independent, non-partisan analysis. I am a researcher, first and foremost. But my identity as a researcher is enabled by another core identity: I am a systems-builder. These systems, most of which are invisible, help me find patterns and write quickly for a broad audience.

But sometimes, in the course of building systems for my own work, I stumble onto something I realize would be valuable to make public, too. This is a principle best articulated in ’s book Show Your Work!, which I highly recommend.

The Detention Mortality Dashboard is an example of this. Launching today, the Detention Mortality Dashboard is a repository of official ICE custodial deaths from 2018 until now, with searchable case-level information, source materials, and an interactive map. What started out as a simple spreadsheet with a few fields is now a more robust public tool.

Much of the credit goes to my amazing colleague Bishal Timsina for doing the heavy lifting to make this project possible. Bishal is one of the developers on the team at Relevant Research, the same team behind Detention Reports, the Enforcement Dashboard, Relevant Football, and Paradarsita — as well as the migration of ImmProf Blog, the Women’s Refugee Commission’s Detention Pregnancy Tracker, and American Families United’s Collateral Damage Report. Huge thanks to everyone else on the team who provided support and feedback.

This work is intended to complement, rather than duplicate, the independent tracking conducted by people like Andrew Free at Detention Kills, Eyes On ICE, and others whose work I have relied upon and respect enormously.

The philosophy behind the dashboard

Although the concept behind the dashboard may seem simple, we take this work seriously and we put a lot of thought into how we executed this project. Let’s start with our shared humanity—and our shared mortality.

The death of anyone is a serious matter, and it is all the more serious when that person dies while in a controlled facility run by a government agency that chooses to take responsibility for their care and well-being. Collating these tragedies together under one dashboard has the potential to blunt the humanity behind each story and treat life and death as mere datapoints. Even if this is not the intent, it may be a regrettable side-effect.

At the same time, allowing the details surrounding these deaths to remain scattered across government webpages is its own kind of irresponsibility. The disorganized nature of this information makes it more difficult to see these stories, and the institutions where they occur, as part of a broader pattern that deserves scrutiny.

Whether we effectively balanced these interests will be up to the public.

Note that we refer to this as a “mortality” dashboard, not simply a “death” dashboard. This is intentional. Whereas death often connotes individuality, mortality reflects a public health framework that seeks to understand individual deaths as the product of socialized and institutionalized risk factors that can be mitigated only through systemic change.

About the ICE Detention Mortality Dashboard

I am eager to show you the dashboard, but bear with me. With the proliferation of AI, there is an explosion of overnight “dashboards” out there from unknown and often unnamed contributors and very few of them do the hard work of explaining the decisions that went into the work (sometimes because there’s no human with domain expertise actually making any decisions). So let’s go over the basics.

Who should use the Detention Mortality Dashboard?

We believe a wide audience will find this useful. Here are just a few examples:

• Reporters covering a death at a specific facility can use it to quickly see whether that death fits a pattern or stands alone.

• Researchers and advocates documenting conditions at a facility over time can use it alongside the underlying detention population data at Detention Reports.

• Families and legal counsel trying to understand what’s happened at a facility where a loved one or client is held now have a starting point that doesn’t require knowing which press release to search for.

• Elected officials and their staff can get the latest numbers and details from an independent, non-partisan researcher who can also answer further questions.

Where does the data come from?

The data we collected here are from publicly available ICE records. These come from two sources:

1. Detainee Death Notifications, the press releases ICE publishes on its News Releases and Statements page announcing that someone has died.

2. Detainee Death Reports, more detailed reports that ICE is required to eventually publish on its Detainee Death Reporting page.

These sources are imperfect, increasingly brief, and frequently delayed. If you want to see what these reports miss, listen to my conversation with St. John Barned-Smith and Ko Lyn Cheang of the San Francisco Chronicle regarding what they found when they compared ICE’s official narratives to family accounts and medical records. The Detention Mortality Dashboard is a starting point for that kind of scrutiny, not a substitute for it.

We do not currently attempt to collect additional information about ICE’s detention deaths. Finding and vetting such information would, by itself, require enormous effort for which we do not have the capacity, and would turn this into a full-blown research project. That said, we hope that this project provides a trusted resource and starting point for further investigation.

What data is available?

The main questions we sought to answer in this project were: who died, when did they die, and where did they die.

• Who. We include names, ages at time of death, and nationality. This information is included in nearly all ICE Detainee Death Notifications, and the database is searchable.

• When. We include date of actual death and date of ICE’s announcement. As I said above, these notifications can be delayed; this allows the public to view how late the announcements are.

• Where. This question is more challenging than it may appear. Although these deaths all involve people in ICE custody, “custody” can still mean in a hospital outside of a detention facility. (See my related conversation with NILC.) So we decided to capture both the detention facility where people were last held and the facility or location where they actually passed.

How can I help ensure that this work is sustainable?

I’m so glad you asked! Much of my public scholarship is powered through the generous support of paid Substack subscribers. Many people assume that this is extra “spending money” on top of a stable salary. This is not the case. My work has always been independently and creatively funded—and now I’m asking you to support this work with a few bucks a month or each year. I have years of evidence now to show that my research has been able to maintain rigorous quality while also generating an outsized impact on our national immigration conversation, including connecting with people across the political spectrum. Is ensuring that this work continues worth $7 a month or $70 a year to you? If your answer is “yes,” please click the link below to help me keep the lights on.

Subscribe now

What data is not available?

Notice that we did not seek to answer the question of why did someone die in ICE custody. ICE lists some of the circumstances surrounding each death in their official notice and adds more information in their subsequent reports. Based on the full range of available research, I am not convinced that either of these sources provides a sufficient accounting of the cause of the death.

My rationale is two-fold. First, from a purely medical perspective, I don’t believe they are thorough enough nor independent enough to provide a definitive medical conclusion about cause of death. Second, and to reiterate the systems-thinking point above, the cause of an individual death is not the same as the cause of patterns of death (i.e., mortality) among a subset of the population. If a person with diabetes died from medical complications inside a facility, we could say that diabetes is the cause of death; but if the institutional factors surrounding the same death were predictable and preventable, we could, from a social science perspective, say that the cause of death was detention itself.

The dashboard does not attempt to assess these factors (I am not qualified to do so) nor adjudicate various causal interpretations.

We also do not repeat ICE’s assertions about the legal histories, both immigration- and criminal-related, of people who die in detention. Two factors argue against including ICE’s claims about criminal and legal history. First, the agency has not produced records that would allow us to validate and situate their claims (same principle as the medical cause of death issue above). And second, I simply refuse to participate in the indignity of suggesting, even by implication, that because someone had a criminal history, that they somehow deserved their death. Immigration detention is a form of civil administrative custody. No one is doing prison time for a crime and even if they were, detention is not a death penalty.

How is the data visualized?

Effective data visualization is mostly a process of addition through subtraction, a principle I wish more dashboard designers out there would consider (or ask their AI to consider)—which is why the current version of the site is intentionally minimalist and functional.

The central feature of the Detention Mortality Dashboard is a map that shows the locations of deaths in the dataset. The dataset currently goes back to 2018, and we will attempt to backfill information over time. As noted above, since many deaths actually occurred outside of a detention facility, the map shows the detention facility currently, or most recently, associated with the deceased.

The dashboard opens on a map of the United States, pinned only at the facilities where someone has actually died in ICE custody. We do not include every detention facility ICE operates. Each pin is color-coded to show whether the person died inside the facility itself or died elsewhere, most often at a hospital after being transferred for emergency care. Seeing that split visually, facility by facility, makes a pattern visible that’s nearly impossible to piece together from reading press releases one at a time. Hover over any pin and the available details for that person and that death appear.

The default map shows all detention deaths for the time span of data. Use the filters to narrow to specific calendar years or fiscal years, states, and whether the death took place inside or outside a detention facility. The map filters work in real time, so if you are trying to prepare the map for a screenshot (say, for a presentation or social media), you should be able to set it up how you want, then grab the image.

Below the map, a second chart, “Deaths in ICE Custody Over Time,” turns that same data into a year-by-year bar chart, toggleable between fiscal year and calendar year, with a dropdown to drill into monthly totals for any given year. If you filter on a year, the graph will change to a month-by-month graph for that year.

Scroll further down and you reach the Case Registry: a searchable, filterable table listing every case individually, by name, nationality, age, last detention facility, date of death, and, notably, the date ICE actually reported it. Having that last pair of columns side by side is the point: you can see, case by case, exactly how much time passed between a person’s death and ICE’s public acknowledgment of it.

Each case is registered on its own page with data we have on record, as well as links to the underlying source material: all Detainee Death Notifications and, when available, the detailed report. To make it easy to learn more about each facility, we link directly to the corresponding report on detentionreports.com.

How do I make a suggestion, report a mistake, or ask a question?

Feel free to reach out to me directly if you see any inaccuracies or bugs, want to share a feature request, or if you are a reporter looking for support on a story.

Click to Email Austin

I’m sure we will make lots of tweaks now that the dashboard is public, so there may be minor differences between screenshots in this post and how the site looks when you view it live.

When does the data get updated?

We will maintain this dashboard for the foreseeable future, regardless of who is in the White House. When ICE publishes a new Detainee Death Notification, we add it to the dashboard as soon as the notification becomes public. More detailed Detainee Death Reports, along with any subsequent updates to existing records, are added as they become available, typically the same week. If you notice anything missing, don’t hesitate to let us know.

How will I know when the website is updated?

The best way to stay up to speed about this data is by subscribing to my Substack here, where I write about every new detained death, or follow me on pretty much any social media platform, where I will share updates. I’m on Bluesky, X, Instagram, Threads, LinkedIn, Mastodon, and Facebook. We do not currently have a way of subscribing to site-specific updates, but we are considering it for the future. Let me know if that would be valuable for you.

How do I see the data myself?

For anyone who’d rather work with the underlying data directly than browse case by case, there’s also a dedicated Download the Data page, sitting in the site’s navigation right between the dashboard itself and the documentation. It’s super simple: one click gets you the full dataset, along with the metadata describing what’s in it, the time span it covers, and the file’s format and size, so you know exactly what you’re getting before you download it.

Where can I see the Detention Mortality Dashboard?

The dashboard lives on its own domain at https://detentionmortality.com.

Visit the Detention Mortality Dashboard

Support this work

I’m honored that this Substack continues to drive media reporting, research, and public debate, but it does take effort and expense. If you can pitch in a few bucks a month or year, it would help enormously. Find an option that fits your budget at austinkocher.substack.com/subscribe.

Subscribe now

Leave a comment

I am including social-media-ready images and explanations for each of the data points so that you can help spread this information beyond Substack. You can keep up with the data between editions on Bluesky, X, Instagram, Threads, LinkedIn, Mastodon, and Facebook.

Share

3

Three people died in connection with the Delaney Hall detention center, including one ICE never announced.

ICE confirmed to Representative Rob Menendez during an August 3 oversight visit that a man had died after a medical emergency during intake at Delaney Hall in the week of July 20, a death the agency never announced and that his office had first heard about on July 24. ICE granted the man a discretionary release to hospital care before he died, which allowed the agency to treat him as no longer in its custody, and in July it dropped a Biden-era requirement to report deaths that occur after a release like that. His name has not been made public and his death appears in no public tally, including the one I maintain. Edwin López Cornejo died at the same facility on August 1 after six weeks in custody, and I wrote about that death earlier this week. A reporting rule that stops at the hospital door leaves the official death toll to the agency’s own discretion, and it took a member of Congress walking into the building to surface this one.

• Janoski, S. (2026, August 5). Third Delaney Hall death went unreported. The Jersey Vindicator.

$96M

The value of the medical, dental, and pharmacy claims the Department of Veterans Affairs processed for immigration detainees over nine months of 2025.

Records obtained through public records litigation document 233,772 claims worth more than $96 million, processed by the VA on behalf of ICE’s Health Service Corps and Customs and Border Protection between January 1 and October 7, 2025. Providers have apparently not been paid for care delivered on or after October 1, 2025, when ICE shifted the work to two companies under no-bid contracts worth more than $110 million. The detained population nearly doubled between January 2025 and January 2026, so the volume of care ICE is responsible for has grown while the mechanism for paying for it has stalled. Hospitals and physicians who go unpaid for ten months have every reason to stop accepting the work, and the people who lose access are locked inside facilities where they cannot seek care anywhere else.

• American Oversight. (2026, August 6). Records Raise Urgent Questions About Broken Medical Claims System for ICE/CBP Detainees. American Oversight.

46,000

The number of people ICE booked into detention in July, the highest monthly total of this administration.

ICE took more than 46,000 people into custody in July, surpassing the roughly 43,000 book-ins in June that had been the second-term high, at an average of about 1,500 people entering custody each day. The detained population stood at 68,000 over the first weekend of August, approaching the record of more than 70,000 set in January. Book-ins measure how many people enter the system over a month, while the population figure counts how many are held on a given day, and the two move together only when releases and removals keep pace. A month like July fills detention faster than it empties, which is what pushes the standing population back toward a record even as individual stays get shorter.

• Montoya-Galvez, C. (2026, August 3). ICE detentions set Trump-era high in July, as immigration crackdown widens. CBS News.

1,400

The number of beds ICE plans to create at a shuttered federal prison in North Carolina, which would make it one of the largest immigration jails on the East Coast.

Internal ICE documents show the agency intends to reopen Rivers Correctional Institution in North Carolina by the end of the year with 1,400 beds. GEO Group pitched the site months ago on the argument that its rural setting offers a secure operational environment with minimal community disruption. Reopening a prison moves faster than building a facility from scratch, and it inherits a physical plant designed for criminal punishment rather than civil detention, which is the legal category immigration detention occupies. Rural placement also puts distance between a detained person and an attorney, and that distance shapes how a case turns out long before a judge hears any of it.

• Lartey, J. (2026, August 1). ICE Is Trading Warehouses for Old Prisons. The Marshall Project.

249

The number of parking spaces ICE is seeking in Boston, part of a solicitation covering more than a dozen cities.

A General Services Administration notice shows ICE conducting market research for 249 surface or structured parking spaces near existing or planned ICE facilities in the Boston area, on a lease of at least three years. Governor Maura Healey and Mayor Michelle Wu both objected, and Wu said she had heard nothing about ICE expansion in the city beyond the parking notice itself. Massachusetts enacted the PROTECT Act on August 5, which limits ICE enforcement activity across the state. A three-year lease on 249 parking spaces describes a scale and a duration of local operations that ICE has announced nowhere else.

• Scalese, R. (2026, August 5). Local officials line up against ICE quest for nearly 250 Boston parking spots. WBUR.

$3B

The amount the Justice Department is offering states and cities for immigration enforcement, available only to those that partner with ICE.

The Bureau of Justice Assistance is making roughly $3 billion available through a program the department named BIDEN, for Bridging Immigration-related Deficits Experienced Nationwide. An agency qualifies only if it participates in a 287(g) partnership with ICE or commits to entering one, and supports Homeland Security Task Force operations. The money may go to hiring officers, prosecutors, and detention staff, to enforcement and detention operations, and to surveillance technology including drones, license plate readers, and crime analysis platforms. Conditioning federal money on local participation turns 287(g) from a voluntary program into a budget decision, which reaches sheriffs and chiefs who have no particular interest in immigration enforcement but do want equipment they cannot otherwise afford. A second application round closes on August 19.

• Bureau of Justice Assistance. (2026, June 15). DOJ FY 2026 Bridging Immigration-related Deficits Experienced Nationwide (”BIDEN”) Program. U.S. Department of Justice.

143

The number of New York state and local officials who have signed a letter asking congressional Democratic leaders to defund and dismantle ICE.

A letter addressed to Senate Minority Leader Chuck Schumer and House Minority Leader Hakeem Jeffries carried 143 signatures from New York state and local officials by the end of July, up from 107 when it was first released on July 20. State Senator Julia Salazar, Assembly Member Diana Moreno, and City Council Member Shahana Hanif led the effort, and the signers include Bronx Borough President Vanessa Gibson along with officials from Long Island, Westchester, and Ulster County. The letter argues that ICE’s abuses of power predate this administration and will outlast it, which aims the demand at the agency itself and not at whoever is running it. The suburban and upstate signatures are what change the political arithmetic, because abolition has read as a position of urban progressives and a wider geographic base is harder to wave off as a New York City phenomenon.

• Lewis, R. C. (2026, July 29). More New York Dems are calling for ICE to be abolished. City & State New York.

42

The number of artificial intelligence tools ICE was running internally while disclosing 23 of them in the federal inventory meant to list them.

Records produced through public records litigation by Just Futures Law, Pangea Legal Services, and Mijente show ICE operating roughly 42 AI tools internally as of May 2024, with only 23 of them appearing in the Department of Homeland Security’s public AI Use Case Inventory at the time. CBP was running about 75 AI systems in the same period, 62 of them active. The inventory exists so the public can learn which government systems shape decisions about individual people, and a disclosure rate below 55 percent describes what an agency prefers to publish rather than what it operates. These counts are two years old, and it took litigation to surface them at all, when the inventory was built to publish them.

• (2026, August 5). Immigration as a Test Case for Executive AI Governance. Lawfare.

Support This Work

I’m honored that this Substack continues to drive media reporting, research, and public debate, but it does take effort and expense. If you can pitch in a few bucks a month or year, it would help enormously. Find an option that fits your budget at austinkocher.substack.com/subscribe.

Subscribe now

That includes the wife of Army Sgt. Hedar Leonel Turcios Juarez, an active-duty military service member who was stationed at Fort Bliss in Texas. She was snatched by ICE in front of their six-year-old daughter in a parking lot, and now he is worried about how to focus on his military career while trying to hold his family together.

The AP articles by Jack Brook, Safiyah Ridge, and Aaron Kessler looked at 50 cases, but there are no doubt many more cases that remain invisible. Like the myth of “automatic citizenship” that comes with marrying a U.S. citizen, there is a myth that simply serving in the military means you get citizenship and won’t face deportation. Not true. There is a path to citizenship, but there are all kinds of barriers to finishing the citizenship process when your priority is completing your military duties, not filling out paperwork.

Our country has broken its promise to immigrant veterans many times, both in times of war and times of peace. Just look at how President Truman betrayed Filipino World War II veterans through the Rescission Act of 1946. Or how military veterans have been abandoned by our social safety nets after their service, left to cope with substances, convicted of crimes for which most Americans would get a second chance, and then deported back to countries they barely know.

Our country is breaking our social contract with veterans and active duty military today by going after them and their families—and that’s what I find so alarming about the AP articles. As a veteran myself, I find the 50+ cases in the AP story to be a sign that our immigration system is so broken and ICE is so out of control that not even people who put their lives on the line in uniform are protected. Many of these spouses and children are literally going through the process they were told to go through—and ICE is still arresting them.

To understand the deeper story behind the story, I talked with Danitza James who runs Repatriate Our Patriots who fights for fairness for military families impacted by the immigration system. I’ve been fortunate to get to know Danitza recently through our conversation as part of my series “Veterans for Good” leading up to Memorial Day and through her work alongside American Families United, an organization that represents about 1.5 million U.S. citizens with spouses who have no pathway to legal status.

Danitza is a powerful voice, she speaks from experience as an Army veteran who wasn’t able to naturalize through military service, and her work now with Repatriate Our Patriots is a source of inspiration and hope for me when I feel hopeless. We talked on Instagram Live today about the larger context of the AP articles, the cases she’s working with right now, and why this reporting is just the tip of the iceberg of what’s to come. Follow Repatriate Our Patriots on Instagram.

Here’s the original announcement below, follow on Instagram for more information about the latest immigration data, policy, and research.

Read the two AP articles below.

• Takeaways from AP’s review of military families detained under Trump’s immigration crackdown

• AP Exclusive: Trump’s immigration crackdown has detained more than 50 military spouses and parents

Learn more below about my four-part series on the intersection of veterans and immigration, as well as my more recent discussion with Sui Chung of Americans for Immigrant Justice about what they are doing to protect veterans in Florida and across the country.

And remember: it’s not just immigrants and immigration organizations speaking out about these issues. I had the chance to meet Don, the powerful voice behind in DC this week. He is using his voice and his authority as a career military leader to remind Americans that this—tearing families apart and everything else we’re seeing—is not the country we agreed to defend when we took our oaths of service. We need more people to speak out.

Subscribe now

He is the second man to die after being held at Delaney Hall. Jean Wilson Brutus, a 41-year-old man from Haiti, died on December 12, 2025, the same day ICE detained him, and ICE attributed that death to natural causes. Lopez-Cornejo lived in Plainfield for years and leaves a 12-year-old daughter who is a U.S. citizen.

ICE arrested Lopez-Cornejo in Plainfield on June 18 and held him at Delaney Hall for roughly six weeks. ICE’s own boilerplate promises every detainee a medical, dental, and mental health intake screening within 12 hours of arrival and a full health assessment within 14 days of entering custody. Lopez-Cornejo was in custody long enough to receive both several times over, but he died anyway.

GEO Group owns and operates Delaney Hall. The company ran it as an ICE facility from 2011 to 2017, and after that contract lapsed the building was put to other uses and eventually sat empty. ICE awarded GEO a 15-year contract for the 1,000-bed facility in February 2025, worth roughly $1 billion over its term and more than $60 million in the first full year of operations, and the building reopened that spring. It is now the largest ICE detention center on the East Coast. I published a full data profile of Delaney Hall in May, covering who is held there, how they got there, and where the facility fits within ICE’s northeast detention network.

Reporting by the Jersey Vindicator and Gothamist fills in much of what ICE left out from their official report. Steve Janoski reported that Lopez-Cornejo managed diabetes, high blood pressure, and seizures, all of which required medication, and that his family says he did not receive it.

His mother, Maria Cornejo, told Movimiento Cosecha that she believes negligence caused her son’s death and that she wants an answer for it. His brother, Danilo Lopez, said the facility did not provide the supplies he needed. Michael Sol Warren and Arya Sundaram reported that more than 70 people held at Delaney Hall have filed federal suits alleging medical neglect, and that more than 30 of them say they were denied prescribed medication for chronic conditions. Citing conditions such as these, about 300 people inside the facility went on hunger and labor strike in May.

His attorneys had already gone to federal court over his health. NJ 101.5 reported that they filed a habeas petition on July 24 challenging his detention on the basis of his chronic conditions, and that Judge Brian R. Martinotti ordered ICE not to transfer him. He died eight days later.

Governor Mikie Sherrill said she was deeply disturbed by the death and opposes private for-profit detention. Newark Mayor Ras Baraka said GEO Group and ICE have repeatedly disregarded transparency about the health and safety of the people they hold. Baraka was arrested in May 2025 while trying to inspect Delaney Hall, and ICE and GEO have since denied entry to routine inspectors as well as to Sherrill and Representative LaMonica McIver. Representative Bonnie Watson Coleman called for Delaney Hall to close and asked ICE for the timeline of the medical emergency, the medical records including medication access, and the incident reports and oversight reviews tied to the death.

When I spoke recently with Heidi Altman, Vice President of Policy at the National Immigration Law Center, she explained that the amount of money ICE has drives how many people it locks up. The $70 billion enforcement law Republicans passed this summer through budget reconciliation committed roughly $38 billion to ICE through 2029 while stripping the transparency rules that normally ride along with the dollars, including the requirement that ICE publish basic data on who it holds.

, author of the forthcoming book How to Close a Camp, and I talked recently about whether these places are better understood as detention centers or as camps. A second death at the same facility within eight months will only deepen the concern that the ongoing deaths inside immigration detention are no accident, but the product of a dehumanizing apparatus that knowingly puts immigrants at risk and keeps expanding anyway.

Subscribe now

Leave a comment

New starting today, I am including social-media-ready images and explanations for each of the data points so that you can help spread this information beyond Substack. Wherever you share, consider tagging me and linking back to this Substack. I’m on Bluesky, X, Instagram, Threads, LinkedIn, Mastodon, and Facebook.

1,474

The average number of people ICE arrested per day over the first eleven days of July, the highest daily pace of this administration.

ICE released enforcement data for the first time since April, and it shows the agency arresting an average of 1,474 people a day through July 11. That is the highest daily pace since the president returned to office. The numbers point to a shift toward higher-volume operations that rely less on the large, visible raids that defined the spring. A higher daily arrest rate means more people moving into detention and removal proceedings every week, regardless of how visible the operations are.

• Reznick, A. (2026, July 24). ICE released data for the first time since April. It shows ramped up arrests, detention. KJZZ.

444,000

The number of asylum cases a new federal rule could redirect from asylum officers straight to immigration judges.

U.S. Citizenship and Immigration Services estimates that up to 444,000 cases could be affected by a rule announced this week that eliminates the asylum officer interview and sends applicants directly to immigration court. That interview has long been the first and less adversarial step where many claims are resolved. Routing cases straight to a backlogged court system could slow decisions and place applicants in front of judges without that earlier screening. The change may reshape how hundreds of thousands of people experience the asylum process, depending on how the rule is implemented and challenged.

• Prager, S. (2026, July 27). New Trump Asylum Rule Would Send Hundreds of Thousands of Applicants Straight to Deportation Proceedings. Common Dreams.

24,000

The number of unaccompanied children who could lose federally funded legal representation when a contract expires today.

More than 24,000 unaccompanied children facing deportation stand to lose access to federally funded legal help as the contract funding that work expires on July 31. Children have no right to appointed counsel in immigration court, so this program has been the main way many of them get a lawyer at all. Without it, thousands of kids may face a judge alone, some of them too young to understand what is being asked of them. Whether the funding is restored will decide if those children keep representation in the weeks ahead.

• Vives, R. (2026, July 30). More than 24,000 immigrant children could lose legal representation as a federal contract expires. Los Angeles Times.

460,000

The number of leads the Office of Refugee Resettlement has shared with immigration enforcement since January 2025.

Internal data shows that the Office of Refugee Resettlement, the agency responsible for the care of unaccompanied children, has passed more than 460,000 leads to immigration enforcement since January 2025. ORR exists to protect the children in its custody, and channeling this volume of information into deportation operations turns a child welfare agency into an arm of enforcement. It also raises real questions about the safety of the families who come forward to claim these kids, since a parent or relative who steps in to sponsor a child now risks handing their own information to ICE. A system that makes claiming a child dangerous will leave more children stuck in federal custody.

• Malleck, J. (2026, July 29). US Child Migrant Agency Tipped Off ICE, Leading to Arrests. Documented.

90

The number of people the Justice Department has moved to strip of citizenship since the start of 2025.

The Justice Department has filed to revoke the citizenship of nearly 90 people since the beginning of 2025, with at least 89 cases filed in federal court as of July 22. Denaturalization was rare for decades and is now being used at a pace with little modern precedent. It reaches people who already completed the full process of becoming citizens, which puts the security of naturalized status itself in question for a growing number of Americans. Civil denaturalization also carries fewer protections than a criminal case, since there is no right to appointed counsel and a lower burden of proof in some proceedings. The trend signals a broader effort to treat citizenship as reversible.

• Cann, C. (2026, July 26). Citizenship revoked? Trump administration ramps up controversial plan. USA Today.

4

The number of people who died in connection with federal immigration enforcement in a single month.

Over roughly one month, four people died in incidents tied to federal immigration enforcement. Two were shot by ICE agents, in Maine and Texas. One was hit by a truck while fleeing an operation in Florida. One died in ICE custody after being denied medication. Together these deaths track the same expansion in enforcement the other numbers this week describe, and behind this figure are four people who are now dead.

• The Intercept Briefing. (2026, July 24). Trump’s National Killer Police Force. The Intercept.

$776M

The additional amount ICE may pay to extend a no-bid contract at one of its largest Texas detention centers.

ICE plans to extend Amentum Services’ no-bid contract to run Camp East Montana through September 2027, at a potential additional cost of $776 million to taxpayers. Camp East Montana is one of the agency’s largest detention sites, and a no-bid extension keeps a single company in place without competing offers that might lower the price. The figure shows how detention spending compounds once a facility is running, since the cost of one more year at a single site runs higher than the $452 million contract that first installed the operator in March. Detention expansion is now as much a fiscal story as an enforcement one.

• Foley, R. J. (2026, July 27). ICE extends no-bid contract for major detention center in Texas. ABC News.

132,000

The number of people who have purportedly left the country through the administration’s self-deportation program.

Of the more than 200,000 people who signed up for Project Homecoming through CBP One, the administration’s “self-deportation” program, about 132,000 have perhaps left the United States, with another 70,000 waiting to depart. The completion figure runs well below the 200,000-plus sign-up number the administration tends to promote as evidence the program is working and the administration has not released data to confirm these numbers.

• Golding, B. (2026, July 28). More than 200,000 migrants have taken Trump’s offer to pay them for self-deporting. The Independent.

5

The number of federal appeals courts that have now rejected the administration’s claim that it can detain immigrants without a bond hearing.

The Ninth Circuit became the fifth federal appeals court to reject the administration’s argument that it may hold immigrants in the interior of the country without a bond hearing. Five circuits lining up the same way deepens a split that the Supreme Court will likely have to resolve. For now, people in those circuits may have a stronger claim to a hearing before a judge, depending on the facts of their case. A Trump-appointed judge joined the most recent decision, which makes the ruling harder to wave off as merely partisan.

• Crepeau, M. (2026, July 30). Trump Immigration Detention Authority Rejected by Appeals Court. Bloomberg Law.

287(g)

The federal program authorizing local police to enforce immigration law that Maine has now effectively banned.

A new Maine law effectively bans 287(g) agreements, the federal contracts that let local police carry out immigration enforcement. Only one Maine agency, the Wells Police Department, had ever signed one, and it withdrew last October, so the practical effect is less about terminating existing agreements than about preventing future participation. The law puts Maine among the states building legal walls between local policing and federal immigration enforcement, at a moment when the administration is suing other states to force the opposite. State law is becoming one of the few remaining checks on how far federal enforcement can reach into local communities.

• Fonseca, C. (2026, July 29). New laws in Maine taking effect Wednesday limit police cooperation, information sharing with immigration agents. The Boston Globe.

Support This Work

I’m honored that this Substack continues to drive media reporting, research, and public debate, but it does take effort and expense. If you can pitch in a few bucks a month or year, it would help enormously. Find an option that fits your budget at austinkocher.substack.com/subscribe.

Subscribe now

Thank you , , , , , and many others for tuning in!

Subscribe now

In the first months of 2025, ICE reopened family detention at the South Texas Family Residential Center in Dilley and at the Karnes County facility, both in Texas. CoreCivic, which runs the Dilley facility, says its five-year contract there will generate roughly $180 million a year, and the Karnes facility is operated by GEO Group. The reopenings reversed a decision the Biden administration made in December 2021 to end long-term family detention and replace it with electronic monitoring and case management. The Dilley facility had closed entirely in June 2024. I explained what family detention is, how we got here, and why it is so contested in an earlier post, Family detention: what you need to know, and I documented the rapid growth at Dilley this past winter in Family Detention at Dilley Triples in Size.

After family detention re-opened, the numbers of people held there increased in 2025, priced in early 2026, but have since declined substantially at Dilley and somewhat at Karnes. The decline is surprising since ICE has been arresting more people than ever this month. Detention Reports, the ICE facility hub built by Relevant Research, tracks each facility’s population using Interval ADP, the average number of people held during each reporting interval. As of the July 9, 2026 data, Dilley averaged about 161 people and Karnes about 914. Those are facility totals that count everyone held at each site, and Karnes in particular holds mostly single adults, so its number is not a count of detained families. The facility-level trend at each site is shown in Detention Reports’ own charts below. (As I explain below, the older higher numbers of people detained is not directly comparable to the current detained population.)

Despite some reduction in the population of detained families, ICE continues to deport large numbers of families. Counting from January 20, 2025 through the July 11, 2026 release, ICE recorded more than 66,000 removals of people identified as family-unit members, and at the current pace that figure passes 68,000 by the end of July. Family-unit removals have run at roughly 10 to 11 percent of all removals under this administration. In the release covering early August 2025, the family share climbed above 20 percent, more than 4,000 family-unit removals in a single two-week reporting period. The share has held fairly steady, but because total removals have surged, the absolute number of family members deported has grown with it. If the current pace holds, family-unit removals will reach roughly 185,000 by the end of the term in January 2029.

As the data show below, in fact, family units made up a higher percentage of deportations during the Biden administration—but the composition of who is being deported and why, as well as the total numbers of deportations, is really quite different. I would urge caution against any simplistic comparisons here.

A note on the terms, because ICE uses several and they do not all mean the same thing. These definitions come from the footnotes tab of ICE’s detention statistics spreadsheet. A family unit, which ICE abbreviates FAMU, is measured for each individual whose case family status is intact or intact-reunified, meaning that person was identified as a family member by CBP or ICE. It counts individual people, not families. Each adult and each child in a unit is counted separately, so one family of four traveling together contributes four to the 66,000 total. ICE also states that being identified as a family-unit member does not mean the whole family was removed together or at the same time. The removal totals themselves include removals, returns, and expulsions, which is how ICE defines its removals data. The two family facilities, Dilley and Karnes, are what ICE calls Family Residential Centers, written FRC in the footnotes and FSC in the data tables, and ICE says it resumed their use on April 7, 2025, after discontinuing them on March 31, 2022.

The facility charts come with a caveat that changes how they should be read. The same buildings have held very different populations over time. Dilley opened as a family detention center in 2014, held families through the Obama years, was converted to short-term processing under Biden, and then closed. Karnes held single adults for much of that period. Their populations before 2025 were largely not family detention in the current sense, so the pre-2025 portion of each facility’s chart should not be read as a record of family detention even though it was at the same facility that was previously (and now again) designed for family units. ICE’s own footnotes date the return of family detention to April 7, 2025, which is the point to read from on each chart.

ICE publishes the removal totals and the family-unit subset in the same biweekly spreadsheets it has released for years, and Detention Reports publishes the facility-level population with a documented method and stable links. The family detention centers are open again under multi-year contracts, and the deportation of family members is running in the tens of thousands and on pace to reach six figures before the term is out.

Subscribe now

Professor Rebecca Galemba at the University of Denver created the DU Court Transparency Project with a vision of documenting immigration court hearings with academic rigor and with the goal of having a scholarly and public impact. I wanted to talk to her and her students about their work because it raises so many important questions, from how to access detained courts, how to reliably document court hearings over time and across a rotating cadre of volunteers, and what her students are seeing first-hand inside the detained court in Aurora. And just as important, I wanted to know what drew Rebecca and these students to the work and what they wished more people knew.

I’ve respected Rebecca’s work for a long time even though I’ve only recently had the opportunity to get to work with her more closely. She is a professor at the University of Denver’s Josef Korbel School of Global and Public Affairs and co-director of the Center for Immigration Policy and Research, and she is the principal investigator of the Court Transparency Project. Ella Iveslätt is the project manager and lead research assistant, a 2024 DU graduate who built the project’s database and its public dashboard. Jasmine Salgado Simental is a lead observer and an undergraduate at DU who grew up in Aurora, and Jennifer Gutierrez Marquez is a research assistant and a graduate student at the Korbel School. Jasmine and Jenni both came to the work through their own families’ experience of the immigration system.

Much of what we talked about is simply how hard it is to document a place like this at all. Observers cannot bring in phones or laptops, so everything is captured by hand on paper and reconstructed later, which is why Rebecca built the project around a small, well-trained team that stays with it and can be retrained as the rules shift. She described building and rebuilding their methods as the government kept changing its practices, and learning to record both what a judge decided and the exact language used to reach it. The building itself sets the tone before a single hearing begins. Ella walked us through the entrance at the Aurora facility, which is run by the private prison company GEO Group, where the skylights are covered with a painted image of the sky and the walls are lined with awards and value words like “accountability” and “character” despite the dissonance with what actually goes on there.

A lot of the conversation was about the human weight of the hallways and courtrooms. The students described families kept apart from the people they came to support, the confusion of respondents who do not understand what is being decided, and Jasmine’s observation that the only real difference between her and the person in custody can come down to a uniform. Latina students described being mistaken for respondents when observing in immigration courtrooms. Several of them were drawn to this work through their own families’ journeys through the immigration system, and that history is part of why they stay with something this demanding. The idea they returned to most is that this system depends on not being seen, and that understanding it begins with people who are willing to sit in the room and pay attention, then communicate that out to the public through writing and presentations like this one.

Jenni had a powerful encouragement at the end that I 100% agree with:

“Most courts across the country should be public, so if anywhere in your area is public, please go bear witness. Bearing witness, even if you might not understand all the complexity of legal jargon, will teach you a lot.” —Jenni Gutierrez

If you can visit an immigration court, do it. It’s one of the best learning experiences you’ll ever have. It is harder to get into court these days, with the agency creating new barriers to entry and flooding the courts with mega-masters hearings where respondents themselves can barely fit into court. But if you can go with someone who knows the ropes, don’t miss the opportunity.

Watch or listen to the discussion above. I learned so much and I’m sure you will, too. Huge thank you to Professor Rebecca Galemba. Her projects illustrate the value of impact-driven scholarship at a time when skepticism about higher education is at a high. And she is, no doubt, doing a lot of invisible labor mentoring students and collaborators into this work—for which I have tremendous appreciation and admiration. Don’t forget to share this conversation with others so we can spread good information.

Share

Timestamp of key themes

• 00:00 — Opening and introductions

• 4:50 — What inspired the DU Courtwatch and Court Transparency Project

• 12:01 — How the students found their way into the work

• 20:16 — Why detained court is the black box of the immigration system

• 21:07 — Walking into the Aurora facility, from the parking lot to the courtroom

• 33:24 — How identity shapes who gets questioned in the room

• 34:32 — Turning courtroom observation into data you can defend

• 38:06 — What dismissals and pretermission actually mean

• 45:40 — The findings on bond, jurisdiction, and who is detained

• 49:02 — Mandatory detention and “weaponized incompetence”

• 52:30 — Where the project goes next

• 57:57 — What the team wishes the public understood

About the Guests

Rebecca Galemba is a professor at the University of Denver and a colleague of mine who has studied immigration from many angles for years. Trained as a qualitative researcher, she leads the DU Courtwatch Project, which grew out of a smaller court watch effort she built with attorney Christina Brown of the Colorado Asylum Center around the Biden-era dedicated docket in 2022 and 2023. She also runs a wage theft research project with immigrant workers, and much of what she described is the invisible labor of mentoring students into this field and keeping a small team trained well enough to document a moving target.

Ella Iveslätt is the project manager of DU Courtwatch and a recent University of Denver graduate who started on the project as a student observer in 2023. She is the team’s quantitative lead, and she wrote her undergraduate thesis on the project’s dedicated-docket data under the Biden administration, with Galemba as her advisor. She first found her interest in immigration law in a high school class at the University of Chicago, and she is now applying to law school with a commitment to stay involved in immigration law in some capacity for her career.

Jasmine Salgado Simental joined the project in 2023, and her interest in immigration law began in a freshman course called Deportation Nation taught by Professor Lisa Martinez. She grew up in Aurora, where the detention facility sits, in a family and community shaped by immigration. She was drawn to the work by the lack of transparency in the courts and by a desire to make what happens inside them legible to the general public rather than only to people in academic spaces.

Jennifer Gutierrez Marquez, who goes by Jenni, came to this work through her own family. Her parents began their immigration process around 2010, and it took roughly a decade of lawyers and hearings before they became permanent residents now working toward citizenship, with Jenni often translating for them along the way. She worked in higher education before her master’s program, and she started with Galemba on the wage theft research project before joining the courtwatch team, where she found the legal impact she wanted to have for her immigrant community.

Five Things You Can Do Right Now

1. Explore the data yourself. The Court Transparency Project’s Phase One dashboard and charts cover the Aurora detained court, with an updated version through June on the way. Look at the numbers on bond, jurisdiction, and representation and see what the team documented one hearing at a time.

2. Go bear witness. Most immigration courts are open to the public. If there is one near you, you can sit in for a day as a curious member of the public. As Jenni said, bearing witness will teach you a lot even if you do not follow every piece of legal jargon, and nothing explains how the system works better than watching it.

3. Support or join the project. You can reach the DU team at ducourtwatch@gmail.com, and there are ways to follow and get involved on their website. Attorneys and advocates who want to be interviewed are especially welcome.

4. Read the “no jurisdiction” write-up. Ella and Rebecca’s piece for Immigration Law Unhinged lays out how a jurisdictional finding is used to keep people detained.

5. Learn more about the courts. I have a whole section of the Substack newsletter with detailed discussions of the immigration court system going back several years at austinkocher.substack.com/immigration-courts. Visit those posts to learn more about policy changes over time, what the data says, and to find additional resources and readings.

Subscribe now

How did I make this graph? This graph was created using the ICE detention management spreadsheets going back to the start of the Trump administration (and my database of these sheets for several years). I used available data to estimate the number of daily book-ins for the period starting on January 20, 2025 through the next data spreadsheet, used actual data through to the last update in FY 2025 (near the end of September), estimated the final book-ins through September 30, then picked up again with actual data starting on October 1, 2025, continued up until July 11, 2026, then estimated through the end of July using the most recent data. Just simple averages to fill in unknown values, nothing fancy.

As my frequent readers know, I take very few liberties with the data. I rarely speculate, project, predict, or prognosticate unless (a) I have substantial research to support it or (b) I’m undertaking some thought experiment. Today is a (b) day. The graph of cumulative book-ins over time seemed awfully linear, so that got me thinking: if this trend continues, how many people will be booked in by the end of the year? And by the end of the Trump administration?2

To do this, I used the average daily book-in rate for FY 2026 so far (October 1 to July 11), and extrapolated out through January 19, 2029, presumably Trump’s last day in office (although given his age, he could leave office sooner, and given his hubris, he may not leave even in 2029). Using this simple projection based on current FY 2026 ICE operations, we would expect ICE to book in over 800,000 people by the end of this calendar year and 1.75 million people by the end of his term. Even though these are just projections, it does put the scale of immigration enforcement in a longer-term perspective.

Although this is really a thought experiment, the linearity of the data so far suggests that these estimates are entirely plausible. The real proof will be how well actual data hews to projected data. Of course, we can’t reliably do this now because Republicans decided to close the door on ICE’s biweekly detention spreadsheets, so unless the agency continues to publish the data as they did last week—and kudos to them if they do!—the public will remain mostly in the dark about the reality of detention.

On a related note, most of these 600,000+ people won’t spend (or haven’t spent) another free day in the United States after entering detention, since most people are deported instead of released back into the country. This trend changed significantly and suddenly at the start of the Trump administration with a sudden increase from 60% of detentions resulting in removal to over 90% of detentions result in removal—although that total has now dipped below 80% (I explain why in my previous detention post, linked above).3

So here is where things stand. In the year and a half since January 20, 2025, ICE has booked close to 600,000 people into detention, and if the current pace holds the total will pass 800,000 by the end of this year and reach something like 1.75 million by the time this term ends in January 2029. Roughly 65,000 people are in detention on any given day, and a growing share of the people who leave detention are not being released back into their communities but deported, up from about 60 percent under Biden to around 80 percent now.

And these numbers could climb higher still.

The administration is stripping legal status from people who have held it for years, which turns lawful residents into people ICE can detain and deport. In June the Supreme Court cleared the government to end Temporary Protected Status for roughly 350,000 Haitians while litigation continues, and the same legal reasoning reaches the rest of the program, close to a million people from countries like El Salvador, Honduras, and Syria.

At the same time, the immigration courts are being used to manufacture removal orders in bulk through the mega master hearings I wrote about earlier today, where a single morning can put more than a hundred people, most of them without a lawyer, on a path to deportation. In June alone, 210 of these hearings across 24 courts swept up more than 27,000 people. Every one of those moves expands the population of people who can be booked into detention, which means the totals in this post are more likely a floor than a ceiling.

Today’s post continues this ongoing research, and adds unreported findings based on qualitative research about how mega-masters are being scheduled that deepens our understanding. I am grateful for the insights from experts who helped me to fill in crucial gaps in understanding. If you have more information you would like to share with me, you can contact me on any platform or reach out to me on Signal (@austinkocher.99).

Note: I have to turn this into more than one post because there’s too much information to share, so be sure to subscribe so you get the next one when it comes out shortly.

Using Court Data Through July 30

Before we talk about new findings, I want to set the stage with a quick comparison between the predictions for June versus the reality. The data in this post reflects the latest immigration court data updated through the end of June, the first full month of mega-masters. Last month we only had the scheduled mega-masters, which means we were working with more of an estimate.

When analyzing immigration court hearings, the only thing we know for sure is what was scheduled in the past. This creates a blind spot. Even at this moment, we don’t know what was actually scheduled on July 1 yet. We won’t know that until we get more data in August. We can be fairly certain that the hearing schedule data on June 30 is mostly accurate for July 1, because it’s just one day apart. But as we extrapolate, the likelihood that the schedule for July 2 changed is higher than on July 1; the uncertainty of the July 3 schedule is higher still, and so on incrementally into the future. See my previous explanation here.

To show this more concretely, here is the comparison of total in-person non-detained morning mega-masters predicted by the May data looking into June, and the June data once the hearing schedules were locked in. Altogether, I was undercounting morning non-detained in-person mega-masters by about 15%.

When we look at how this 15% shakes out over the month, we see an interesting trend. The first two months of June were actually over-projected while the later weeks (as expected) were far more under-projected. I won’t get into the details about why. I just wanted to illustrate the fluctuation in projected versus actual.

As I explained in the previous posts, I focused on morning in-person non-detained hearings to align more closely to the reporting and to sort of limit the complexity of what I was trying to explain when we were in the early days.

What Counts as a Mega-Masters?

While much of the reporting rushed ahead with assumptions about what a mega masters is, as a researcher, my first concern is with definitions. My previous posts sought to historically situate and analytically define mega-masters in a way that enables us to develop a shared critical conception of what’s going on, as well as standardize some of the language surrounding these hearings.

In those posts, I argued that when it comes to developing a threshold for how many cases constitutes a mega-masters (50? 70? 100?), we could either adopt a model of relative growth at the particular court or we could adopt the threshold that had been reported in the news so far of 100. I have been using the 100 cutoff, but was uncertain what—if any—official cutoff the EOIR had or how exactly particular mornings or afternoons were chosen as mega-masters.

We now have an answer to these and many more basic questions.

What Do We Call Mega-Masters?

If you are grammar obsessed, you may have noticed that I’ve been hyphenating these hearings as “mega-masters,” which is different than the unhyphenated “mega masters” I was using before. There’s a simple reason for this change: this is what the EOIR is calling it. So if you’re writing about a mega-master or mega-masters plural (lowercase on both), just be aware that this seems to be the EOIR’s official convention.

What is the Official Mega-Masters Threshold?

Based on conversations I’ve had and the internal documents I’ve reviewed, the answer to the official mega-masters number of hearings is unambiguous. The EOIR’s official internal goal is to have between 100 and 200 cases scheduled for each of these mega-masters. While it is possible (likely, even) that the total number of people physically in court will be under this, the agency is clear that a mega-masters is constituted through this 100-scheduled-cases threshold.

Analyses that use thresholds below 100 will overcount the number of mega-masters, and analytical thresholds above this will undercount the number of mega-masters.1 Downstream analysis from different thresholds will also reflect a different population and different findings than the official EOIR threshold. (See the next section for a caveat.)

Based on this definition, and my previous discussion of focusing on mornings and afternoons, the backwards looking hearing schedule for June shows a total of 210 in-person non-detained mega-masters across 97 judges in 24 courts, which impacted 27,496 people who were scheduled for those hearings.2 (Okay, technically 205 people were in more than one mega-masters, so it’s 27,496 mega-masters people slots or seats and 27,291 total people. I’m not a stickler about this for this post, but just so you know.) In total, 15.9% of all people scheduled for a master calendar hearing in June were scheduled for mega-masters.

The official goal for mega-masters is actually more specific than just the 100 cases, however. The EOIR isn’t just aiming for 100 total people per mega-masters, but to explain I need to digress briefly. In immigration court, cases that involve family members (mostly spouses and children) typically have one person that’s a “lead” and other people that count as “riders.” For example, a parent with two children equals one lead and two riders, or three total people who are linked together in a specific way in removal proceedings.

What I’ve learned is that, for mega-masters, the agency’s goal is for mega-masters hearings to have 100 adults and leads, not counting riders, with a goal of 150 total people (i.e., 50 additional riders). Not every adult is a “lead,” of course; many are simply single adult cases. And not all children are riders; many are unaccompanied children with no adult (more on this in later posts).

Is the information I have correct? The good news is, we don’t have to take this on faith. We can test this theory.

Let’s look more closely at the 27,496 people scheduled for mega-masters in June. Of those, 18,977 were primary respondents (either solo adults or leads) and 8,519 were riders. The primary group divides further: 4,695 of them were leads with family members attached to their case, while 14,282 were adults appearing on their own with no one else linked to them (i.e., solo adults). Altogether, a little under half the people called to a mega-master in June arrived as part of a family group, and those groups averaged about 2.8 people each. The riders are overwhelmingly children, with a median age of 13, two thirds under 18, and 43 percent under 12.

To the research point in question: on average, the typical mega-master scheduled about 131 people, of whom 90 were primary respondents and 41 were riders. This isn’t quite the 100-lead/solo threshold, but as I’ll discuss later, there are good reasons why it might be hard to hit that number. In my assessment, the data supports the findings from the qualitative research. The EOIR seems to be setting a target not just of any 100 people, but specifically setting a target of 100 adults/leads plus however many riders (mostly children) come along with those leads, which then pushes the hearing schedule well over 100.

In case you’re wondering if mega-masters are “targeting” families more than other hearings, the answer so far is: not really. Around 30% of all cases (non-detained, in-person) have riders and this is fairly consistent across mega-masters and non-mega masters cases. This is the national trend; some variation may occur across time and geography so when we look closer at individual courts, those numbers tend to change.

Posts like this require a lot of effort and expertise. Please consider supporting this work with a paid subscription. Your financial support isn’t extra for me—it’s an essential part of how I keep the lights on and keep this work going.

Subscribe now

When and How are Mega-Masters Scheduled?

Although there has been some reporting about mega-masters days in court, one research question I wanted to answer is: why on those days? Who decided when a mega-masters takes place? Based on conversations so far, it appears that working immigration judges in certain courts are required to do mega-masters, but that when they do them is up to each judge. Once a day of the week and time (morning or afternoon) is set, it’s up to the court staff to fill these mega-masters dockets with hearings. Based on the June mega-masters hearings, these are taking place any day of the week but mostly on Monday and Wednesday mornings.

The fact that judges may be choosing their mega-masters days has one important consequence for how we think analyze mega-masters. It is technically possible that a judge would have a mega-master slot that, for some reason, never gets to 100 people scheduled and therefore doesn’t meet the required threshold. This doesn’t seem likely based on the data, but there may be edge cases.

Which Cases are Prioritized for Mega-Master Scheduling?

The definition and the thresholds above tell us what a mega-masters is, but it doesn’t tell us which cases are assigned to mega-masters. Is it just any case, or are there specific cases that the EOIR is telling each court to schedule?3 Based on what I’ve learned, the EOIR is being quite directive, indeed.

First, cases prioritized for mega-masters are being drawn as much as possible from cases where people don’t have attorneys. This is less of a new finding as it is a confirmation of what court observers, attorneys, and reporters have already documented. As many of us have written about for years, immigration attorneys are absolutely necessary for immigrants to receive any semblance of due process in the immigration courts, so choosing cases without attorneys is a way for the EOIR to cherry-pick the most vulnerable and defenseless people, assign them to a mega-master docket, and churn out in absentia removal orders.4

The data bear this out. Looking at the June master calendar hearings alone, mega-masters cases were 19.5% represented while all other master calendar hearing cases were 30.5% represented according to the court data. I will note that these numbers are a little imprecise for the simple reason that the EOIR’s record keeping isn’t great on this point and the reality of what it means to be represented is not as straightforward as you might think. Nevertheless, this method is fairly standardized at this point so the comparison holds: mega-masters cases are 11 percentage points less represented than all other master calendar cases.

Second, cases prioritized for mega-masters are also being drawn from a variety of scheduling situations, including edge cases such as cases from an IJ cancelation or a closed court day that were never rescheduled.5 Some hearings are the first-ever hearing, some are rescheduled from past hearings, some are pulled forward from future schedules, and some are active but had no future hearing scheduled. The only thing that these case types all have in common is the agency seems to believe they can close them out quickly and with as little fuss as possible.

For what it’s worth, I will say that there very likely are groups of cases that are lingering in the court system that legitimately can be closed out easier than others (that’s true for any court system or legal question, see Dworkin’s theory of easy and hard cases). And barring malice or a willful attempt to demolish due process, I’m not principally opposed to that kind of approach. The immigration court backlog is massive and I certainly don’t begrudge any dedicated civil servant who aims to find creative ways to make the courts more efficient and allow judges to focus on the most crucial and complex cases.

The problem is, by my count, most of these specific case categories are actually quite small, numerically speaking, while the general thrust of mega-masters is to scoop up an enormous range of other types of cases which may not be so easy or straightforward except for the fact that the administration has made it easier to deport legitimate humanitarian migrants and long-time residents with strong ties to the U.S. This is by design. The administration can claim that they are simply trying to “clear out the backlog,” while at the same time de facto suspending due process (or whatever version of it used to exist in the immigration courts) the same way that they tried to de facto suspend habeas corpus.

We’ll talk more in the next post about why these cases are so predisposed to negative outcomes in ways that have little to do with the individual immigrants themselves, but for now I just want to emphasize that I am not convinced that mega-masters have anything to do with “fairly, expeditiously, and uniformly interpreting and administering the Nation's immigration laws.” Expeditiously, perhaps—but this is hardly fair or uniform.

What are Mega-Masters Doing to the Immigration Courts?

We have not exhausted our exploration of the latest data on mega-masters. We still need to explore precisely how much notice is given to immigrant respondents (spoiler alert: not enough), what the outcomes of these cases are (mostly in absentia, as others have noted), and how many mega-masters are currently scheduled out for the future.

But I’ll have to address those questions in the next post because I have something more important to share, something even more important than the simple data findings above. And it’s a question that bugged me since May when I first heard of these mega-masters. To put it simply, what are the institutional prerequisites and effects of shuffling such an unprecedented number of cases around? In short, how the hell are they doing it?

Here’s how they’re doing it.

Starting in late April, the EOIR was already signaling internally that it would adopt this new scheduling strategy, and by early May, the courts were told they had to begin creating and filling mega-masters. As one person explained to me, every mega-masters you see is 100% human labor. Every single case is individually identified, selected, and scheduled by a human being who has to identify the case types handed down from the EOIR and move it onto a mega-masters docket. Those 27,500 mega-masters cases in June? Hand-picked by people at each court working under urgent instructions to fill up the calendar. And this fact has all kinds of real world consequences that are easily overlooked.

For example, apparently the EOIR rarely, if ever, permits overtime—but given the heavy lift of rescheduling so many hearings at once, overtime has suddenly become available, if not all but required, in order to perform this Sisyphean effort. Somewhat ironically, while ICE conducts arrests with the aid of expensive, advanced facial recognition software and custom-built smartphone apps, court staff seem forced to work around the clock manually clicking away on the latest directives from the administration. This is part of what mass deportation looks like in 2026: “docket politics”, as I called it the other week.

Before you equate court staff with ICE officers, please know that, in my experience, they have very little in common. Over years of immigration court research, I’ve had a full range of experiences with friendly to less-than-friendly court staff, to be sure. But I have found many court administrators, staff, attorneys, interpreters, and guards to be altogether helpful, to believe in the basic integrity of what they do, and even to see themselves as the frontline of being a welcoming and professional face to immigrants who are navigating a complex and scary system. Unlike ICE officers, court staff were not recruited with white nationalist images and job advertisements and do not see themselves in so politicized of a frame. In short, I have profound respect for the decency of the front-line workers at the EOIR in a way that I find impossible for today’s cadre of ICE officers.

All of the urgent manual work has real consequences for quality and thoroughness. The risk for mistakes is high, mistakes that could have life-altering consequences for immigrants facing deportation. And it’s not just on the front-end with scheduling. After hearings are over, record keeping for what happened at the hearings also has to be entered manually, and given that there are north of 200 cases at some of these mega-masters, there’s no way to do that in real time. From what I have seen, the urgency surrounding scheduling mega-masters comes straight from headquarters with a single-minded obsession with increasing the sheer numbers of closed cases—and that also means a scramble to enter in case outcome data by the end of every single work week so that the administration can tout its numbers on internal scoreboards that ignore quality and hyperfocus on removals, removals, removals. The EOIR is currently sending around regular emails that gamify case closures by highlighting courts and judges with the largest number of closures, and adopting racing-style language between courts and judges like an announcer at a NASCAR event.

The systemic effects are even bigger than this. With all the energy going to mega-masters, who is processing change of address forms? Do judges have any time to actually prepare for asylum cases? Are respondents able to get a copy of their file to review for appeal before the 30-day deadline is up? Is anyone looking at attorneys’ motions? While it is justifiable to be focused on the data on mega-masters, this approach risks reproducing a dangerously myopic view on what this administration is doing to the court system as a whole. The truth is, the mega-masters are not only impacting the 27,500 cases at mega-masters hearings in June — the whole mega-masters tactic could be undermining the routine functions of the court for everyone and further contributing to the delegitimization of the immigration court system as we know it.

Coming Up Next…

In the next post, I’ll continue to explore the new mega-masters trends and compare the findings of qualitative research to what the data says. We are still only scratching the surface of this new practice, so stick around. Subscribe to make sure you get the upcoming posts. And like I said, if you have more information you would like to share with me about mega-masters or the courts in general, you can contact me on any platform or reach out to me on Signal (@austinkocher.99).

Subscribe now

Adam Sawyer, the Director of Research at Relevant Research, joined me to talk about why he and the team built Detention Reports, a look under the hood at how the analysis works, and a tour of the new features on the site. Detention Reports maps every major facility and lets you filter by state, city, owner, and operator, with a chart on each facility page that shows its population over time and the role it plays in the system, from county jails that hold few people for short stays to the mega detention centers that hold many for months.

The newest additions are our own research on who owns and operates each facility and how it is classified. California City is the clearest case, a facility CoreCivic sold to the Department of Homeland Security after California passed a law letting state officials inspect it, so it is now federally owned and privately operated. Interval ADP also catches facilities that open or expand mid-year faster than the official average can, and at California City the difference is concrete, roughly 1,650 people against the 1,426 ICE reports. Every facility page carries a stable URL built on the official facility code, so a citation made today still resolves months from now.

VISIT DETENTION REPORTS

Since 2018, Congress required ICE to publish detention data every two weeks, a bipartisan provision that the latest party-line appropriations removed, so the July release arrived with no clear legal mandate behind it and after a gap of more than ninety days. Even so, this is a data-rich moment for immigration research by historical standards, and the work that makes it usable, documented methods, stable links, and transparent decisions, is what separates a citable source from a dashboard that may vanish in six months.

Our work is nonpartisan by design. The data is the data, and the July 9 snapshot already understates the surge in arrests during the first days of the month, which is one more reason to read any single number as a point in time. We will update the site whenever ICE publishes next, whether that is in two weeks or another three months, and our full methodology is at relevant-research.com.

Watch the full conversation above. If you’re a journalist or researcher with questions, feedback, or corrections, email Adam at adam@relevant-research.com.

Timestamp of key themes

• 00:00 — Opening, and why tracking ICE detention is difficult

• 03:22 — The gap between ICE’s numbers and local reporting, and the two-week data mandate

• 06:05 — Why the fiscal-year average distorts the count, and how Interval ADP fixes it

• 08:18 — The three-month data gap and the law that ended ICE’s reporting mandate

• 11:04 — What the 65,765 snapshot captures and what it misses

• 13:16 — A tour of detentionreports.com

• 16:57 — Inside a facility report, Moshannon Valley and the roles facilities play

• 20:41 — Why owner and operator matter, and ICE buying California City

• 23:12 — How Interval ADP catches fast-growing facilities, from California City to Fort Bliss

• 26:01 — Building data people can cite, and fighting link rot

• 32:47 — Nonpartisan by design

• 34:48 — Deaths in detention, and immigrants in the emergency room

• 42:25 — What the July 9 data still misses, and how this differs from the Deportation Data Project

• 44:30 — The data moment we’re in, and how to help

Help Keep These Conversations Going

Thank you , , , , , and many others for tuning into my live video with ! Join me for my next live video in the app. And don’t forget to subscribe.

Subscribe now

Subscribe now

65,765

The number of people held in ICE detention nationwide as of July 11, from the first enforcement data ICE has released since April.

ICE released enforcement data for the first time since April, and my analysis of it shows the agency arresting more people than at any point in this administration, an average of 1,474 arrests a day over the first 11 days of July. Detention remains near record levels, at 65,765 people in ICE custody as of July 11, and the largest group is immigrants with no criminal history at all, a share that has kept growing since April even as the proportion with charges or convictions falls. The data also shows more than 10,600 people released from detention by court order since January, most of them likely cases where a judge found ICE was holding someone unlawfully, which is about a quarter of all releases that did not end in deportation. After a three-month blackout, the numbers describe a system arresting at record speed and detaining people who, by the government’s own data, are mostly not the criminals it says it is pursuing.

• Kocher, A. (2026, July 21). ICE Quietly Arresting More People Than Ever According to Latest Data. Austin Kocher.

214

The number of children swept up in ICE enforcement operations that a new ACLU report examined, 32 of them U.S. citizens.

The ACLU’s report Agents of Chaos and Cruelty reviewed more than 1,200 enforcement actions across eight states and found 214 children caught up in ICE operations, 32 of them U.S. citizens. The same review counts hundreds of protesters, journalists, and clergy targeted and roughly 400 incidents involving misconduct, the pattern that ties isolated raids into a policy. I spoke with two of the report’s authors, the ACLU’s Naureen Shah and Denise Bell, about what they found and why they describe it as reckless by design. Counting the citizen children matters because it shows an enforcement model operating with so little precision that it sweeps up the very people it has no authority to remove.

• American Civil Liberties Union. (2026, July). Agents of Chaos and Cruelty: How the Trump Administration’s National Deportation Policing Force Has Attacked American Communities. American Civil Liberties Union.

1st

The first case ever filed in the Alien Terrorist Removal Court, a national security deportation tribunal that went unused for the 30 years since Congress created it.

The Justice Department’s National Security Division filed an application on July 15 to the Alien Terrorist Removal Court, the first use of a tribunal that Congress created in 1996 and no administration had ever invoked. The court is built to deport noncitizens the government labels terrorists while withholding the classified evidence against them, and its chief judge, Joan Ericksen, held a closed hearing the next day before ordering prosecutors to explain the legal basis for their request. The respondent’s name is sealed, so the public knows a secret deportation court is now operating without knowing who it is being used against. Activating a dormant tool built to run on evidence a person cannot see establishes a precedent that will outlast whatever this first case decides.

• Pagliery, J. (2026, July 23). A Decades-Old Terrorist Deportation Court Is Being Used for the First Time. NOTUS.

14

The number of locations where ICE is seeking builders for new or expanded detention facilities, from the Virgin Islands to Guantánamo Bay, under a plan capped at $10 billion.

A federal request for construction proposals shows ICE looking to build or expand detention at as many as 14 sites, from the U.S. Virgin Islands to Guantánamo Bay, moving toward purpose-built facilities after its earlier attempt to convert commercial warehouses stalled in lawsuits. Each proposed building follows the same design, a single story of 26 cells sharing three showers, at an estimated $24 million to $40 million apiece, with total expansion spending capped at $10 billion and contracts expected by the end of September. Several of the sites, including Guantánamo Bay and St. Thomas in the Virgin Islands, have held few or no immigration detainees before. Building rather than renting signals a system meant to be permanent and harder to unwind than the county jails and converted warehouses ICE has leaned on until now.

• Wang, P. (2026, July 20). ICE Seeks New Or Bigger Detention Facilities at 14 Sites, Documents Show. TIME.

$22,500

The amount ICE will pay a partner agency for each unaccompanied child it locates and checks on under a new incentive program for local law enforcement.

Under a new incentive program for its 287(g) partners, ICE offers local agencies up to $22,500 for each unaccompanied child they track down and subject to a “safety verification,” combining a $2,500 case payment with up to $20,000 more per completed check. The program, documented through a service agreement with an Idaho sheriff’s office, treats finding and visiting children who were released from federal custody as a billable enforcement task with a price attached to each one. Framed as protecting children, the payments give local police a cash reason to locate kids whose households often include people the same agencies can then move to deport. A bounty structure attached to children turns their safety into the government’s entry point into their families.

• Thrasher, A. (2026, July 17). A $22,500 Per-Child Bounty System. Maxwell Commons.

350,000

The number of Haitians whose deportation protections and work permits the Supreme Court cleared the administration to end, with their work authorization lapsing this month.

About 350,000 Haitians hold Temporary Protected Status, and a 6–3 Supreme Court decision cleared the administration to terminate it, with the work permits that a last-minute reprieve had extended reaching their new expiration on July 24. Losing that status strips both protection from deportation and the legal ability to work from people who have built years of life in the United States, roughly a third of them in Florida. The same ruling found that TPS terminations are largely shielded from judicial review, which is why the same logic has since reached Yemenis and other nationalities. Ending a legal status for hundreds of thousands of people at once, and placing that decision beyond the reach of the courts, removes the check that once slowed revocations on this scale.

• Bustos, S. R. (2026, July 10). About 350,000 Haitian TPS Holders Get Last-Minute Reprieve: Work Permits Extended to July 24. WLRN.

950,000

The number of people DHS itself estimates could drop or forgo public benefits like Medicaid and food assistance under its new public charge rule.

DHS estimates that roughly 950,000 people could disenroll from or avoid programs such as Medicaid, SNAP, and housing assistance under a new rule letting officers weigh the use of those benefits in green card decisions. Most of the projected drop is a chilling effect. The rule itself denies few people benefits directly, and instead families forgo aid they already qualify for out of fear it will later be used against them or a relative. The people who go without are disproportionately children and mixed-status households, many of them citizens who lose coverage because someone else in the home is not. I talked with experts about how that same fear keeps immigrants out of emergency rooms, and the public charge rule is the mechanism that carries the fear across the rest of the safety net.

• Sganga, N. (2026, July 16). DHS Could Weigh Immigrants’ Use of Medicaid, Food and Housing Help in Green Card Decisions. CBS News.

6,600

The number of noncitizens New Jersey found on its voter rolls after a motor vehicle software error, fewer than 400 of whom ever cast a ballot.

New Jersey removed about 6,600 noncitizens from its voter rolls after a Motor Vehicle Commission software error registered them in 2023 and 2024, and Governor Mikie Sherrill said fewer than 400 of them had ever voted. The error is a real administrative failure, and it is also the opposite of the coordinated fraud that the phrase “noncitizens on the voter rolls” is usually used to imply. A glitch that produced fewer than 400 votes across two years, caught and corrected by the state itself, is evidence that the safeguards work rather than that elections are being stolen. The same 6,600 gets held up as proof of a stolen election precisely because the correction it triggered, and the fewer than 400 votes actually involved, go unmentioned.

• Mulvihill, G. (2026, July 21). 6,600 Noncitizens Registered to Vote by New Jersey Motor Vehicle System Error, Gov. Sherrill Says. PBS NewsHour.

5

The number of states the Trump administration has sued, along with the city of Philadelphia, over laws requiring ICE officers to show their faces and identify themselves.

The Trump administration has filed suit against five states, New York, Virginia, Connecticut, New Jersey, and California, and against Philadelphia, arguing that laws barring ICE officers from wearing masks and requiring visible identifiers would put agents in danger. To make their case, federal lawyers point to an online tool that runs officers’ photos through facial recognition, but the artist who built it says it draws on a decade-old list of about 700 self-identified ICE employees and cannot identify the deportation officers now in the field, most of whom joined after ICE roughly doubled its force from 10,000 to 22,000. The suits would strike down accountability measures on the strength of a threat the government has not shown is real. Whether officers carrying out home raids can hide their identities is being decided in the government’s favor before the factual claim behind it has been tested.

• Varner, M. (2026, July 22). States Want ICE Agents to Show Their Faces. The Trump Administration Is Blocking Them. WIRED.

~33%

The share of the roughly 2 million immigration court hearings held remotely by video in fiscal year 2025, according to the GAO.

The Government Accountability Office finds that almost a third of the roughly 2 million immigration hearings in fiscal year 2025 were conducted by video rather than in person. Stakeholders the GAO interviewed raise concerns that remote hearings can weaken a person’s ability to consult counsel, present evidence, and be understood through an interpreter, all of which bear directly on whether a hearing is fair. Video lets a strained court system move a historic caseload faster, but speed and access pull against each other when someone’s deportation turns on being heard. As the docket grows, how these hearings are held becomes as consequential as how many there are.

• U.S. Government Accountability Office. (2026, July 15). Immigration Courts: Stakeholder Perspectives on the Use of Remote Hearings (GAO-26–108110). U.S. Government Accountability Office.

2–1

The vote by which a federal appeals court affirmed the release of a Georgetown scholar whom ICE had detained over his speech.

A three-judge Fourth Circuit panel voted 2–1 on July 23 to uphold the release of Badar Khan Suri, a Georgetown postdoctoral scholar from India whom ICE arrested outside his Virginia home in March 2025 and held for six weeks in Texas. The majority rejected the administration’s argument that federal courts cannot review the constitutionality of an immigration detention until the immigration case itself ends, holding that habeas protections apply while a person is still being held. That reasoning reaches beyond Suri, because the government has used the same jurisdictional argument to move speech-based detention cases out of the reach of the courts that would hear them. A 2–1 margin also shows how contested even this check has become on the federal bench.

• Center for Constitutional Rights. (2026, July 23). In Big Win, Appeals Court Affirms Release of Georgetown Peace Scholar Detained by ICE for His Speech. Common Dreams.

Support This Work

I’m honored that this Substack continues to drive media reporting, research, and public debate, but it does take effort and expense. If you can pitch in a few bucks a month or year, it would help enormously. Find an option that fits your budget at austinkocher.substack.com/subscribe.

Subscribe now

Get Immigration Data Updates on Your Favorite Platform

Wherever you already spend your time, you can keep up with the data between weekly posts. Find me on Bluesky, X, Instagram, Threads, LinkedIn, Mastodon, and Facebook. Bluesky continues to be a really great place. If you’re not there yet, consider joining and following along there.

I just hit 5,000 followers on Instagram this week, which is an exciting milestone on a platform that just a year ago I really didn’t understand very well or care about that much. Now it’s the source of a lot of productive collaborations and conversations.

Share

First things first. DetentionReports.com, a project from and the Relevant Research team, is designed to make detention data more understandable and citable by reporters, researchers, policymakers, and interested members of the public. Using public data and a method developed by Relevant Research called Interval ADP, the team calculates the current average population at individual detention facilities in use across the country.1 The facility-level reports provide additional data in a searchable format with stable links to reports that can be easily referenced over time. If this is your first time hearing about Detention Reports, read our first announcement last year, see the updated design and features, and visit the about page on the site for a detailed log of changes, methodological decisions, and history of the project.

VISIT DETENTION REPORTS

Detention Reports would not be possible without the contributions of the Relevant Research data team, including Michael Ale, who has worked on the project since the beginning, as well as Nush Ojha and Rashmi Singh, with ongoing contribution from Riwaj Chalise and the development team. Sunita Kunwar was instrumental in the initial conceptualization of the project.

The updates were originally announced by Adam on Bluesky. Follow him there for more immigration data.

Long Time, No See!

You may have noticed that it’s been a while since we’ve updated the site with new data. There’s a simple explanation. This data previously came from a mandate in appropriation acts since 2018—but when Republicans move forward with a party-line funding bill earlier this year, they chose to exclude the simple provision they’ve agreed to for years that required ICE to provide at least some minimal transparency.

After the previous data release in early April, therefore, the agency’s immigrant detention data has gone completely dark. Why the agency chose to release data now is a mystery. But as my dad used to say, don’t look a gift horse in the mouth (although I’m still not sure what that means). We’re always happy to have new data, and we hope that our work illustrates the value to be gained of ICE continuing to release this data with or without a mandate. ICE is clearly still producing this data and they can proactively release it anytime they want. With that said, let’s look at key trends and new resources we’ve added to the site.

Detention Reports Findings

As Adam said to me when we looked at the data, it’s important to remember that the long break in data does blunt the precision of the observations. Interval ADP uses two consecutive data releases to calculate period averages—so longer periods between releases necessarily imply greater uncertainty. Real totals at each facility may be higher or lower than the data shows. Despite this caveat, these numbers are the best current data the public has on the detained population by facility.

Camp East Montana outside of El Paso (aka Fort Bliss) and Stewart Detention Center in Georgia are still the two largest facilities by capacity and current population (respectively). Stewart’s population remains high while Fort Bliss has dropped dramatically from its height at the beginning of the year (but still holds more than a thousand people).

Other facilities have exploded in population. Diamondback Correctional in Oklahoma shot up over 1,000 since the start of the year (and may be even higher right now), as has Folkston in Georgia, and the smaller but still significant Midwest Regional in Leavenworth, Kansas, which has seen huge growth since April. The Florida Soft Sided facility (Alligator Alcatraz) is not on the current list since it has closed.

California City Processing Center2 stands out not just for it’s growth in the past year but for one additional peculiar fact: it is one of the first facilities that ICE decided to purchase outright—the taxpayer now owns this facility—while maintaining the operating contract with Geo Group. Read my longtime colleague Wendy Fry’s article with Nigel Duara for CalMatters to learn more about what going on. (“What the sale of 2 ICE detention centers reveals about Trump’s new immigration strategy.”)

This new and increasingly complex detention system—especially where DHS is buying up warehouses and now buying up detention centers—presented us with an additional data question: how to represent these new entangled relationships so that the public can trace the ownership and operations of these facilities? ICE does not maintain a public repository of these facts and the detention spreadsheet itself does not include this data.

To solve this, Adam and his team went over all available public data to assign owner and operator to each facility, noted when the two are not the same, and created additional categories for the various permutations of federal, private, and local owner-operator relations.

When you visit Detention Reports now, you’ll notice new filtering options on the left that allow you focus on types of owner-operator relationships, as well as filter in on all facilities by specific contractors like CoreCivic. It’s a simple yet elegant way to get more information you need from a trusted source.

We also note that five new facilities have been added to the dataset since early April:

1. Cache County Jail, Logan, UT

2. Central Valley Annex, McFarland, CA

3. Charlotte County Jail, Punta Gorda, FL

4. Garvin County Detention Center, Pauls Valley, OK

5. Randall County Jail, Amarillo, TX

45 Detention Contract Documents Added

To enrich Detention Reports even further, the Relevant Research team is collecting and posting detention center contracts from a quiet publication source on ICE’s website as well as crowdsourced contributions and other sources available online. The latest update includes 45 more contracting documents across 18 facilities, for a total now of 164 contracting documents across 78 facilities. You can find those contracts at the bottom of any facility that has them. If you have any contracts you’d like to add, reach out to Adam.

Join the Conversation!

Adam and I will talk through the latest updates to Detention Reports on Monday, July 27 at 1:00 PM Eastern right here on Substack Live. The best way to join us is to make sure you subscribe and download the Substack app so you get a notification when we get started. Bring questions, comments, and feedback!

Subscribe now

Press Inquiries

If you are writing about immigrant detention and have questions about how Detention Report data fits into your story, reach out to Adam Sawyer at adam@relevant-research.com.

Relevant Research is Here to Help

Even in a period with less transparency than we would prefer, Relevant Research is committed to applying our specialized skills set and deep domain area expertise to making data available, understandable, and usable to a wide variety of audiences. If you are need on demand support for projects at the intersection of immigration research, data science, web/app development, and public communication, contact us at contact@relevant-research.com.