United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT ____________ No. 25-5087 September Term, 2025 1:25-cv-00240-ACR Filed On: December 9, 2025 Nicolas Talbott, et al.,

Appellees

v.

United States of America, et al.,

Appellants

BEFORE: Pillard*, Katsas, and Rao, Circuit Judges

ORDER

Upon consideration of the emergency motion for stay pending appeal, the opposition thereto, and the reply; the Rule 28(j) letters, and the responses thereto; and the letter request to lift the administrative stay, and the response thereto, it is

ORDERED that the administrative stay entered on March 27, 2025, be dissolved. It is

FURTHER ORDERED that the motion for stay pending appeal be granted. A concurring statement of Circuit Judge Katsas, joined by Circuit Judge Rao, and a dissenting statement of Circuit Judge Pillard are attached.

Per Curiam

FOR THE COURT: Clifton B. Cislak, Clerk

BY: /s/ Selena R. Gancasz Deputy Clerk

* Circuit Judge Pillard would deny the motion for stay pending appeal. KATSAS, Circuit Judge, joined by RAO, Circuit Judge: The United States military enforces strict medical standards to ensure that only physically and mentally fit individuals join its ranks. For decades, these requirements barred service by individuals with gender dysphoria, a medical condition associated with clinically significant distress. This bar was partially relaxed in 2016, revived in 2018, partially relaxed again in 2021, and revived again in 2025. District courts preliminarily enjoined the 2018 revival as a likely violation of equal-protection principles, but this Court vacated one of those injunctions, and the Supreme Court stayed two others. This case presents equal-protection challenges to the 2025 revival.

The 2025 policy generally bars individuals with gender dysphoria from serving in the Armed Forces. The Secretary of Defense concluded that this policy would advance important military interests of combat readiness, unit cohesion, and cost control. In doing so, he consulted materials compiled to assess the 2016 and 2018 policy changes, as well as more recent studies regarding the impacts of gender dysphoria on those with the condition and on their military service. The district court nonetheless preliminarily enjoined the 2025 policy based on its own contrary assessment of the evidence.

In our view, the court afforded insufficient deference to the Secretary’s considered judgment. Accordingly, we stay the preliminary injunction pending the government’s appeal.

I

A

To join and remain in the United States military, service members must meet strict medical standards. Lengthy Department of Defense documents set out the requirements. 2 Add. 1301; see, e.g., Dep’t of Def. Instruction 6130.03, Vol. 1, Medical Standards for Military Service: Appointment, Enlistment, or Induction (May 2024), https://perma.cc/J67Q- 982F (“DoD Accession Standards”); Dep’t of Def. Instruction 6130.03, Vol. 2, Medical Standards for Military Service: Retention (June 2022), https://perma.cc/WFQ3-F229 (“DoD Retention Standards”).

Hundreds of medical conditions are “disqualifying” for accession to military service. DoD Accession Standards at 13– 54. They run the gamut from poor vision, poor hearing, asthma, and high blood pressure to various abdominal, heart, lung, neurologic, urinary, vascular, and other deficiencies. See id. at 13–50. Many mental-health conditions are also disqualifying. These include bipolar disorders, eating disorders, substance-related disorders, obsessive-compulsive disorder, and depression or anxiety disorders under certain conditions. See id. at 50–52. The military traditionally has aligned these disqualifying mental-health conditions with those listed in the Diagnostic and Statistical Manual of Mental Disorders (DSM), a diagnostic and treatment manual published by the American Psychiatric Association. See Doe 2 v. Shanahan, 917 F.3d 694, 709 (D.C. Cir. 2019) (Williams, J., concurring in the result).

These requirements advance many military objectives. Among other things, they ensure that service members can complete required training, serve in harsh or remote environments, and perform their duties as safely as possible. DoD Accession Standards at 5. The standards also reduce the

1 “Add.” citations refer to the government’s addendum, while “App.” citations refer to the plaintiffs’ appendix. 3 risk that service members will “require excessive time lost from duty” for medical reasons. Id. at 4.

B

This case involves application of these principles to gender dysphoria, a medical condition experienced by certain transgender individuals. The term “transgender” describes an individual whose self-perceived gender is incongruent with his or her biological sex. Am. Psychiatric Ass’n, Diagnostic and Statistical Manual of Mental Disorders 511 (5th ed. Text Revision 2022) (“DSM-5”). A subset of transgender individuals experience “gender dysphoria,” a psychological condition stemming from such an incongruence. See id. at 512–13. Gender dysphoria “is associated with clinically significant distress or impairment in social, occupational, or other important areas of functioning.” Id. at 513. Individuals with gender dysphoria often seek specialized treatment, including hormone therapy, surgery, or other substantial interventions. See id. For example, women with gender dysphoria may take testosterone to induce hair growth, a deepened voice, and increased muscle mass. They also may undergo surgery to remove their breasts or change the appearance of their genitalia. Men with gender dysphoria may seek comparable hormonal or surgical interventions.

Before 2016, the medical standards discussed above broadly barred transgender individuals from military service. Add. 125; see Doe 2, 917 F.3d at 696–97 (Wilkins, J., concurring). Over the last decade, however, the governing rules have changed repeatedly. One way or another, recent standards all distinguish between transgender individuals and individuals with gender dysphoria. 4 1

In June 2016, Secretary of Defense Ashton Carter issued a policy permitting military service by transgender individuals not suffering from gender dysphoria. App. 17 (“Carter Policy”). For individuals seeking to join the military, a “history of gender dysphoria [was] disqualifying” unless a licensed medical provider certified that the individual had been “stable without clinically significant distress or impairment in social, occupational, or other important areas of functioning for 18 months.” Id. at 20. Likewise, a “history of medical treatment associated with gender transition [was] disqualifying” unless a medical provider certified that the treatment had concluded and that the individual had been clinically stable for 18 months. Id. For individuals already serving in the military, the policy permitted “[g]ender transition” despite noting that it “presents unique challenges” for “military mission and readiness needs.” Id. at 21.

In significant part, the Carter Policy rested on a report by the RAND Corporation, which recommended that transgender individuals be permitted to serve in the Armed Forces. App. 116 (“RAND Report”). Despite reaching this conclusion, the RAND Report included significant caveats. It flagged as a “critical limitation” the “lack of rigorous epidemiological studies of the size or health care needs of either the U.S. transgender population or the transgender population serving in the military.” Id. at 138. Likewise, it acknowledged a lack of data “from which to estimate” how many transgender individuals have gender dysphoria. Id. at 141. Regardless, the report found that hormone treatments make an individual undeployable for a year. See id. at 202. But it discounted this concern based on an assertion that “only a small proportion” of transgender service members would seek the treatment in any given year. Id. at 204. 5 After instituting the new policy, Secretary Carter issued a 71-page “implementation handbook” addressing “some of the issues” posed for military commanders. U.S. Dep’t of Defense, Transgender Service in the U.S. Military: An Implementation Handbook at 1, 8 (Sept. 30, 2016), https://perma.cc/B9NH- 6YVT (“Carter Handbook”). Among other things, the handbook asked commanders to evaluate whether any “treatment plan” would cause “impacts to the mission (including deployments, operations, training and exercises).” Id. at 53. It discussed deployment to any nation that “view[s] transgender people as culturally unacceptable,” requiring commanders to consider reassignment and to “[p]roceed with caution for the safety of the Service member.” Id. at 69. It anticipated “discomfort” from service members not wanting to share showers, bathrooms, and other living facilities with individuals of the opposite sex. Id. at 60, 64. And it flagged fairness concerns if a transitioning service member could not meet the physical standards established for his or her “preferred gender.” Id. at 48–49.

2

In August 2017, President Trump ordered reconsideration of the Carter Policy. As the President saw it, Secretary Carter had “failed to identify a sufficient basis to conclude that terminating” the prior “longstanding policy and practice would not hinder military effectiveness and lethality, disrupt unit cohesion, or tax military resources.” Military Service by Transgender Individuals, 82 Fed. Reg. 41,319, 41,319 (Aug. 30, 2017). The President reinstated the previous accession policy pending further study, and he substantially restricted government funding for “sex-reassignment surgical procedures.” Id. For individuals already serving in the military, he delayed any changes until Secretary of Defense James Mattis could implement a new policy. Id. at 41,319–20. 6 In February 2018, Secretary Mattis issued a policy titled “Military Service by Transgender Individuals.” Add. 119 (“Mattis Policy”). This policy barred from military service individuals who would require or had undergone “gender transition.” Id. at 120. Likewise, it barred individuals with a history of “gender dysphoria” from joining the military unless they had been “stable for 36 consecutive months in their biological sex prior to accession.” Id. Service members diagnosed with gender dysphoria after joining could continue to serve if they “d[id] not require a change of gender and remain[ed] deployable.” Id. Service members diagnosed with gender dysphoria while the Carter Policy remained in effect could continue serving “in their preferred gender” and receiving “medically necessary treatment.” Id. Transgender individuals without a history of gender dysphoria could join and serve “in their biological sex.” Id. at 121.

The Mattis Policy rested on the “professional military judgment” of “senior uniformed and civilian Defense Department and U.S. Coast Guard leaders.” Add. 119. It also rested on recommendations set forth in a Defense Department report reflecting “input from transgender Service members, commanders of transgender Service members, military medical professionals, and civilian medical professionals with experience in the care and treatment of individuals with gender dysphoria.” Id. at 119–20; see id. at 122–66 (“Mattis Report”). Based on this information, Secretary Mattis concluded that the RAND Report “contained significant shortcomings.” Id. at 120. Specifically, it relied on “limited and heavily caveated data to support its conclusions, glossed over the impacts of healthcare costs, readiness, and unit cohesion, and erroneously relied on the selective experiences of foreign militaries with different operational requirements than our own.” Id. 7 These actions spurred litigation. Before Secretary Mattis issued his policy, four district courts enjoined portions of President Trump’s memorandum. Doe 1 v. Trump, 275 F. Supp. 3d 167, 177 (D.D.C. 2017); Stone v. Trump, 280 F. Supp. 3d 747, 769 (D. Md. 2017); Karnoski v. Trump, No. 17-cv- 1297, 2017 WL 6311305, at *10 (W.D. Wash. Dec. 11, 2017); Stockman v. Trump, No. 17-cv-1799, 2017 WL 9732572, at *16 (C.D. Cal. Dec. 22, 2017). After Secretary Mattis issued his policy, we granted the government’s motion to dissolve one of those injunctions. Doe 2 v. Shanahan, 755 F. App’x 19, 22 (D.C. Cir. 2019). We held that the Mattis Policy “plausibly relied upon the considered professional judgment of appropriate military officials and appeared to permit some transgender individuals to serve in the military.” Id. at 25 (cleaned up). Soon after our decision, the Supreme Court stayed two of the other injunctions against the presidential memorandum and interim guidance. Trump v. Karnoski, 586 U.S. 1124 (2019); Trump v. Stockman, 586 U.S. 1124 (2019). The fourth district court then followed suit. See Stone v. Trump, No. 17-cv-2459, 2019 WL 5697228, at *3 (D. Md. Mar. 7, 2019). With these stays in place, challenges to the presidential memorandum and the Mattis Policy fizzled out. The Mattis Policy thus remained in force from January 2019 until President Biden assumed office in January 2021.

3

Upon taking office, President Biden issued an executive order shifting the policy once again. Enabling All Qualified Americans to Serve Their Country in Uniform, 86 Fed. Reg. 7,471 (Jan. 28, 2021). President Biden directed Defense Secretary Lloyd Austin to “immediately prohibit involuntary separations, discharges, and denials of reenlistment or continuation of service on the basis of gender identity.” Id. at 7,472. And he directed military officials to “establish[] a 8 process by which transgender service members may transition gender while serving.” Id.

In April 2021, the Department of Defense issued new guidance. App. 49; see also Dep’t of Def. Instruction 6130.03, Vol. 1, Medical Standards for Military Service: Appointment, Enlistment, or Induction (Apr. 2021), https://perma.cc/3W6K- 2XHW (“2021 DoD Accession Standards”) (collectively “Austin Policy”). The Austin Policy largely restored the Carter Policy, disqualifying gender-dysphoric individuals from acceding unless they had been “stable without clinically significant distress or impairment in social, occupational, or other important areas of functioning for 18 months.” 2021 DoD Accession Standards at 48. But the Austin Policy permitted “in-service transition,” App. 49, while recognizing that it poses “unique challenges” to “military mission and readiness,” id. at 55.

4

Upon returning to office, President Trump revoked the Biden executive order. Prioritizing Military Excellence and Readiness, 90 Fed. Reg. 8,757, 8,757 (Feb. 3, 2025). President Trump cited longstanding DoD guidance requiring service members to be “free of medical conditions” that may “require excessive time lost from duty.” Id. (cleaned up). He further emphasized that the Armed Forces “must adhere to high mental and physical health standards” to ensure that they can “deploy, fight, and win, including in austere conditions and without the benefit of routine medical treatment or special provisions.” Id. And he concluded that these high standards are “inconsistent with the medical, surgical, and mental health constraints on individuals with gender dysphoria.” Id. The President directed Defense Secretary Pete Hegseth to update the military’s medical standards accordingly. See id. at 8,757–58. 9 Secretary Hegseth issued various memoranda disqualifying from military service individuals with a diagnosis of, history of, or symptoms “consistent with” gender dysphoria. Add. 99–109, 113–16, 198–202 (collectively “Hegseth Policy”). For purposes of the policy, symptoms “consistent with” gender dysphoria means symptoms that “would be sufficient to constitute a diagnosis.” Id. at 200 n.2. In limited circumstances, the policy allows gender-dysphoric individuals to apply for a waiver to join or remain in the military. Id. at 198–99.

Several studies and reviews undergird the Hegseth Policy. These include the Mattis Policy and Report, a 2021 DoD study analyzing appropriate accession standards for transgender individuals, a 2025 literature review on the efficacy of hormonal or surgical treatments for gender dysphoria, and a review of cost data for treating gender-dysphoric service members through 2024. Add. 115–16.

C

Plaintiffs are transgender individuals with gender dysphoria diagnoses. Most of them are active-duty service members. They contend that the Hegseth Policy violates equal- protection principles incorporated into the Fifth Amendment.

The district court preliminarily enjoined the Hegseth Policy. The court concluded that intermediate scrutiny was appropriate because the policy assertedly discriminates based on sex and transgender status. Talbott v. United States, 775 F. Supp. 3d 283, 316–22 (D.D.C. 2025). The court further held that the policy was unlikely to survive intermediate scrutiny. See id. at 322–26. Alternatively, the court held that the policy was motivated by animus toward transgender individuals and thus would fail even rational-basis scrutiny. See id. at 326–29. 10 And it found that the equitable factors favored the plaintiffs. See id. at 332–34.

The government appealed the preliminary injunction and moved for a stay. While that motion was pending, the Supreme Court stayed a preliminary injunction against the Hegseth Policy entered by the Western District of Washington. United States v. Shilling, 145 S. Ct. 2695 (mem.) (May 6, 2025). Soon after, the Supreme Court held that a law prohibiting the use of hormones to treat gender dysphoria in minors does not discriminate based on sex or transgender status and survives rational-basis review. United States v. Skrmetti, 605 U.S. 495, 510–25 (2025).

II

The grant of a stay pending appeal depends on whether (1) the appellant has made a strong showing that it will prevail on the merits; (2) the appellant will be irreparably harmed absent a stay; (3) issuance of the stay will substantially harm plaintiffs or others; and (4) the stay is in the public interest. Nken v. Holder, 556 U.S. 418, 425–26 (2009). Applying these factors, a stay is warranted.

A

The Hegseth Policy likely does not violate equal protection. We doubt that the policy triggers any form of heightened scrutiny. In Skrmetti, the Supreme Court held that a law prohibiting the use of hormones to treat gender dysphoria in minors “classifies on the basis of medical use” and thus does not discriminate based on either sex or transgender status. See 605 U.S. at 510–19. The same reasoning would seem to cover the Hegseth Policy, which classifies based on the medical condition of gender dysphoria. Skrmetti also casts significant doubt on the proposition that a classification based on 11 transgender status would trigger heightened scrutiny. See id. at 517 (“This Court has not previously held that transgender individuals are a suspect or quasi-suspect class.”); id. at 557 (Barrett, J., concurring) (“The Equal Protection Clause does not demand heightened judicial scrutiny of laws that classify based on transgender status.”); id. at 558 (Alito, J., concurring in part and concurring in the judgment) (classification based on transgender status “does not warrant heightened scrutiny”).

Even if the Hegseth Policy contained a classification triggering some form of heightened scrutiny, decades of precedent establish that the judiciary must tread carefully when asked to second-guess considered military judgments of the political branches. And we must do so even in cases involving sex-based or other quasi-suspect classifications. Moreover, the Supreme Court already has held that the government is likely to succeed on its contention that the Hegseth Policy does not violate equal protection. Shilling, 145 S. Ct. at 2695. And this Court, applying settled doctrines of deference to the military, has strongly suggested that the Mattis Policy did not violate equal protection. See Doe 2, 755 F. App’x at 25. Given these decisions, and for reasons summarized below, the Hegseth Policy is likely constitutional.

1

The Constitution vests the “complex, subtle, and professional decisions as to the composition, training, equipping, and control of a military force” in the political branches. Gilligan v. Morgan, 413 U.S. 1, 10 (1973). For example, Article II makes the President the Commander in Chief, U.S. Const. Art. II, § 2, cl. 1, while Article I gives Congress the power to declare war and make rules for the Armed Forces, id. Art. I, § 8, cls. 11, 14. So, the President and Congress have primary responsibility for the “delicate task of 12 balancing the rights of servicemen against the needs of the military.” Loving v. United States, 517 U.S. 748, 767 (1996) (cleaned up).

Given this allocation of powers, the judicial branch must be “reluctant to intrude upon the authority of the Executive in military and national security affairs.” Dep’t of Navy v. Egan, 484 U.S. 518, 530 (1988); see Orloff v. Willoughby, 345 U.S. 83, 93–94 (1953) (“judges are not given the task of running the Army”); Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 645 (1952) (Jackson, J., concurring) (courts should give the President, as Commander-in-Chief, “the widest latitude … to command the instruments of national force”). Indeed, it is “difficult to conceive of an area of governmental activity in which the courts have less competence.” Rostker v. Goldberg, 453 U.S. 57, 65 (1981) (quoting Gilligan, 413 U.S. at 10).

Our review of constitutional challenges to military rules thus “is far more deferential than constitutional review of similar laws or regulations designed for civilian society.” Goldman v. Weinberger, 475 U.S. 503, 507 (1986). In the military context, courts consider whether the policy at issue resulted from the professional judgment of military authorities and reasonably furthers legitimate military interests. Id. Moreover, we must respect the military’s judgment on “the relative importance of a particular military interest.” Id. Courts “must be particularly careful not to substitute our judgment of what is desirable … or our own evaluation of evidence for a reasonable evaluation” by the political branches. Goldberg, 453 U.S. at 68.

Consider two leading precedents in this area. In Goldberg, the Supreme Court rejected an equal-protection challenge to a statute compelling men but not women to register for the draft. 453 U.S. at 59, 83. The Court stressed that Congress had made 13 a considered judgment that this classification minimized “burdens” and “administrative problems” involving “housing … and physical standards,” while promoting “the important goal of military flexibility.” Id. at 81–82. In the “context of military preparedness and the exigencies of a future mobilization,” the Court refused to “dismiss such problems as insignificant.” Id. at 81. In Goldman, the Court similarly rejected a free-exercise challenge to a military regulation restricting the wearing of yarmulkes. 475 U.S. at 509–10. The Court declined to apply the heightened scrutiny normally used to assess First Amendment claims. See id. at 506–07. Instead, it took a “far more deferential approach” than it would have if the restriction had been “designed for civilian society.” Id. at 507. Citing Goldberg, the Court confirmed that “judicial deference is at its apogee” when the political branches make “rules and regulations for [military] governance.” Id. at 508 (cleaned up). The Court thus credited the “professional judgment of military authorities” that the restriction advanced legitimate goals including “obedience, unity, commitment, and esprit de corps.” Id. at 507.

Of course, the Constitution binds the government in the realm of military affairs. See Goldberg, 453 U.S. at 67; Parker v. Levy, 417 U.S. 733, 758 (1974). As we have noted, “the cost of military service has never entailed the complete surrender of all basic rights.” Singh v. Berger, 56 F.4th 88, 92 (D.C. Cir. 2022) (cleaned up). Particularly when Congress has mandated heightened scrutiny, a military policy may be found unlawful where, for example, it is riddled with unexplained exemptions for some but not others. See id. at 98–107 (applying the Religious Freedom Restoration Act). But even then, courts must afford deference to the executive’s considered judgment that a military policy furthers legitimate interests. See id. at 98–99. 14 2

The Hegseth Policy is likely constitutional because it reflects a considered judgment of military leaders and furthers legitimate military interests.

a

The Hegseth Policy reflects the considered judgment of the Secretary of Defense and military leaders in the Defense Department. The policy was informed by “existing and prior DoD policy, and prior DoD studies and reviews of service by individuals with gender dysphoria, including a review of medical literature regarding the medical risks associated with [the] presence and treatment of gender dysphoria.” Add. 115. In significant part, the Hegseth Policy relied on the Mattis Policy, which “conclude[d] that there are substantial risks associated with allowing accession and retention of individuals with a history or diagnosis of gender dysphoria.” Id. (Hegseth Policy) (quoting id. at 120 (Mattis Policy)). In turn, the Mattis Policy “was informed by an extensive inquiry conducted by a panel of experts” who considered the views of many stakeholders, including transgender service members and their commanders. Id. at 115 (Hegseth Policy); see id. at 119 (Mattis Policy). And we previously concluded that, because the Mattis Policy relied on the considered judgment of military officials, the government had “substantial arguments” that it complied with equal-protection principles. Doe 2, 755 F. App’x at 25.

The Hegseth Policy also relied on more recent materials, including a 2021 study addressing appropriate accession standards for transgender individuals, a 2025 literature review on the efficacy of treatments for gender dysphoria, and a review of cost data for treating gender dysphoria. Add. 115– 16. The 2021 study suggested that up to 40 percent of individuals with a gender dysphoria diagnosis would be 15 “deemed non-deployable due to mental health reasons at some time during the 24 months following initial diagnosis.” Talbott, No. 25-cv-240, Dkt. No. 86-1, at 15 (D.D.C. Mar. 17, 2025) (“AMSARA Study”). The 2025 literature review indicated that transgender individuals are far more likely than others “to receive a psychiatric diagnosis” of mood disorders, anxiety disorders, or psychotic disorders, with a suicide attempt rate 13 times higher than that for non-transgender individuals. Add. 189. The literature review also found that the “strength of the evidence” supporting the efficacy of hormonal or surgical treatments for gender dysphoria is “low to moderate.” Id. at 189–93. And the cost data indicated that, between 2015 and 2024, DoD spent over $52 million providing treatment for gender dysphoria to active-duty service members. Id. at 116.

b

The government identifies several legitimate military interests that the Hegseth Policy advances.

Chief among them is military readiness. On its face, the policy applies only to individuals with gender dysphoria, a medical condition that involves “clinically significant distress or impairment in social, occupational, or other important areas of functioning.” DSM-5 at 513. Secretary Hegseth had before him evidence that treatments for gender dysphoria would limit deployability and may not be effective. AMSARA Study at 15; Add. 189–93. His predecessors unanimously agreed that these concerns are serious ones: The Carter and Austin Policies barred accession by individuals with a history of gender dysphoria unless they had been “stable without clinically significant distress or impairment” for 18 months, App. 20; 2021 DoD Accession Standards at 48, while the Mattis Policy drew the line at 36 months, Add. 120. The Mattis Report noted that individuals “receiving hormone therapy and surgery” 16 could not deploy to “austere environments where their healthcare needs cannot be met,” which showed that “the inclusion of individuals with gender dysphoria in the force will have a negative impact on readiness.” Id. at 156. The Carter implementation handbook also anticipated “impacts to the mission (including deployment, operations, training, exercise schedules, and critical skills availability)” from both treatment requirements and international cultural constraints. See, e.g., Carter Handbook at 14, 29, 48, 69. Even the RAND Report, despite recommending that individuals with gender dysphoria be allowed to serve, recognized that hormone treatments could render them unable to deploy for a year. App. 202.

The Hegseth Policy also rests on concerns about unit cohesion and good order. The military’s interest in “fostering cohesion and unity among its members” is “surely” a “compelling” one. Singh, 56 F.4th at 99. And even when applying strict scrutiny, we must “indulge the widest latitude in considering” interests based on an asserted need for unit cohesion. Id. (cleaned up).

One such interest involves the privacy of service members who, as the Mattis Report explained, must “live in extremely close proximity to one another when sleeping, undressing, showering, and using the bathroom.” Add. 159. To protect privacy, Congress has required the military to provide separate housing for male and female service recruits. See, e.g., 10 U.S.C. §§ 7419 (Army), 8431 (Navy), 9419 (Air Force). Even beyond statutory requirements, the military “has long maintained separate berthing, bathroom, and shower facilities for men and women.” Add. 159. Distinctions like these— keyed to biological differences between the sexes—are unquestionably constitutional. As Justice Ginsburg explained for the Supreme Court, it is not only permissible, but indeed “necessary to afford members of each sex privacy from the 17 other sex in living arrangements.” United States v. Virginia, 518 U.S. 515, 550 n.19 (1996). And as Judge Williams later elaborated, in explaining why the Mattis Policy likely did not violate equal protection, “there is no constitutional right for, say, biological males who identify as female to live, sleep, shower, and train with biological females.” Doe 2, 917 F.3d at 707–08 (Williams, J., concurring in the result). The Mattis Report and Carter implementation handbook both acknowledge that allowing gender-dysphoric individuals to serve consistent with their perceived gender identity, as opposed to their biological sex, would impinge on these privacy interests in intimate, communal living spaces. Add. 159 (Mattis Report); Carter Handbook at 60 (issues with “open bay shower configuration”); id. at 63–65 (sleeping areas).

The Hegseth Policy also implicates the military’s sex- based standards for physical fitness. As the Mattis Report explained, “to allow biological males who identify as female to be held to female standards” would raise concerns of unfairness and discrimination. Add. 158. Among other things, such individuals would gain competitive advantages over other males held to the male strength and fitness standards. See id. Likewise, females required to compete against such individuals would be disadvantaged. See id. “Even more importantly, in physically violent training and competition, such as boxing and combatives, pitting biological females against biological males who identify as female, and vice versa, could present a serious safety risk as well.” Id. We cannot lightly dismiss these concerns.

Finally, the Hegseth Policy cited cost issues, noting that the military had spent more than $52 million on treating service members with gender dysphoria between 2015 and 2024. Add. 116. The Mattis Report likewise cited cost concerns. It noted that after the Carter Policy was implemented, medical costs for 18 service members with gender dysphoria increased nearly three times as much as did costs for service members who did not have gender dysphoria. Id. at 163. Moreover, it noted that units sometimes had to use “operations and maintenance funds to pay for” the “specialized medical care” needed by gender- dysphoric service members. Id.

c

The district court held that the Hegseth Policy does not advance these interests. In our view, it gave no sound reason for overriding the Secretary’s considered judgment.

The court dismissed the Mattis Policy and Report as “outdated.” Talbott, 775 F. Supp. 3d at 325. The court concluded that in 2018, when Secretary Mattis issued his report, “medical literature concerning the study and treatment of gender dysphoria” was “in a state of flux.” Id. at 303 (cleaned up). But, the court continued, “medical studies now overwhelmingly conclude that gender dysphoria is highly treatable.” Id. at 304. In support, the court cited a declaration from one doctor who simply stated, in one sentence and without citations, that “gender dysphoria is ‘highly treatable.’” Id. (citing Decl. of George Brown at 4, Dkt. No. 72-77).

We are wary of such a sweeping conclusion. For one thing, the Mattis Report reflects a thorough assessment of evidence available as of 2018, not much time has passed since then, and we are skeptical that the constitutionality of a military policy ebbs and flows as new medical studies emerge every few years. For another, the Hegseth Policy did consider more recent studies, which provide further support for its restrictions. To reiterate, the policy rests in part on a 2021 DoD study suggesting that up to 40 percent of individuals being treated for gender dysphoria may become non-deployable at some point over a two-year period, AMSARA Study at 15, and a 2025 19 literature review finding that transgender individuals are more likely than others to suffer various mental-health issues, Add. 188, and that the “strength of the evidence” supporting hormonal or surgical treatment for gender dysphoria is “low to moderate,” id. at 189–93. Moreover, at least with respect to adolescents, evidence regarding treatment efficacy may be even less settled today than it was a few years ago. See Skrmetti, 605 U.S. at 524 (“Recent developments only underscore the need for legislative flexibility in this area.”); id. at 532–40 (Thomas, J., concurring); Br. of Alabama as Amicus Curiae at 28–32, Skrmetti, 605 U.S. 495 (Oct. 15, 2024). Given all this, the military was not constitutionally required to credit plaintiffs’ side of a fraught medical divide. Finally, the Hegseth Policy advances some interests that do not turn on medical questions at all. For example, the military continues to observe sex-based distinctions for sleeping areas and physical competitions. And the privacy and safety interests advanced by these standards do not turn on whether gender dysphoria is more treatable today than it was seven years ago.

The district court also conducted a searching review of the post-2018 studies relied on by DoD. Talbott, 775 F. Supp. 3d at 301–08, 322–25. We find many of its criticisms unpersuasive. For example, as to the 2021 study, the court concluded that a 40 percent non-deployability rate was “not informative” without knowing the comparable rate for non- transgender individuals. Id. at 304. But with accession standards that screen out everything from asthma to high blood pressure, it strains credulity to suppose that military-wide averages for non-deployability would be anywhere near two out of five. Next, the district court downplayed the 2025 literature review because it addressed all gender-dysphoric individuals, rather than only those serving in the military. Id. at 305. But that broader focus surely understates the challenges faced by gender-dysphoric individuals in the military, where 20 life is often more stressful and demanding than it is for civilians. The court further downplayed evidence of greater mental-health issues faced by transgender individuals as reflecting “external factors.” Id. Maybe so, but that says little about the cost to the military from taking on individuals who have a medical diagnosis associated with significant mental distress. To be sure, the district court fairly noted that the 2021 study and the 2025 literature review contain some evidence cutting in favor of permitting service by individuals with gender dysphoria. Id. at 303–06. But the RAND Report that informed the Carter Policy likewise contained cross-cutting conclusions. Again, the military is entitled to substantial deference in setting risk tolerances and evaluating conflicting evidence bearing on matters like force composition. See, e.g., Trump v. Hawaii, 585 U.S. 667, 707–08 (2018); Goldman, 475 U.S. at 507; Goldberg, 453 U.S. at 70–72; Doe 2, 917 F.3d at 728 (Williams, J., concurring in the result). Here, the evidence supporting the Hegseth Policy was more than sufficient to support the choices made.

3

The district court also held that the Hegseth Policy is rooted in animus against transgender individuals and therefore would fail even rational-basis scrutiny. Talbott, 775 F. Supp. 3d at 326–29. The court invoked a line of cases invalidating statutes that disadvantaged unpopular groups. See, e.g., Romer v. Evans, 517 U.S. 620 (1996); City of Cleburne v. Cleburne Living Ctr., 473 U.S. 432 (1985); USDA v. Moreno, 413 U.S. 528 (1973). In our view, the district court overread these precedents.

These decisions reflect ordinary rational-basis principles; they are not an independent basis for striking down statutes or rules. In Board of Trustees of the University of Alabama v. 21 Garrett, 531 U.S. 356 (2001), the Supreme Court rejected a contention that “decisionmaking reflecting negative attitudes … necessarily runs afoul of” equal protection. Id. at 367 (cleaned up) (discussing Cleburne, 473 U.S. at 448). Instead, the Court explained, cases like Cleburne condemn negative attitudes “unsubstantiated” by anything that would indicate a legitimate government interest. See id. (emphasis omitted). So, “read in context,” the animus precedents “simply state the unremarkable and widely acknowledged tenet of the Court’s equal protection jurisprudence that state action subject to rational-basis scrutiny does not violate the Fourteenth Amendment when it rationally furthers the purpose identified by the [government].” Id. (cleaned up).

The Supreme Court confirmed this point in Trump v. Hawaii. There, it explained that the “common thread” uniting its animus decisions has been the lack of “any purpose other than a ‘bare desire to harm a politically unpopular group.’” 585 U.S. at 705 (quoting Moreno, 413 U.S. at 534) (cleaned up). Put differently, the key question is whether it is “impossible to ‘discern a relationship to legitimate state interests’” such that the policy is “inexplicable by anything but animus.” Id. at 706 (quoting Romer, 517 U.S. at 632, 635). If any legitimate state interest exists, “we must accept that independent justification” and thus reject an equal-protection challenge resting on asserted animus. Id. Because the Hegseth Policy advances legitimate military interests, the animus cases do not help the plaintiffs here.

The district court looked beyond the Hegseth Policy itself to derive animus from various statements made by the President or other officials. The Supreme Court rejected such an approach in Trump v. Hawaii, which upheld a presidential proclamation imposing restrictions on travel to the United States from six predominantly Muslim countries. 585 U.S. at 22 676; see also id. at 736 (Sotomayor, J., dissenting). The plaintiffs in that case highlighted several of President Trump’s social media posts which, in the dissenting justices’ view, “express[ed] animus toward Islam.” Id. at 731 (Sotomayor, J., dissenting); see also id. at 728 (Breyer, J., dissenting). But the Court discounted the “significance of those statements in reviewing a Presidential directive, neutral on its face, addressing a matter within the core of executive responsibility.” Id. at 702 (majority opinion). The task, as the Court viewed it, was to “consider not only the statements of a particular President, but also the authority of the Presidency itself.” Id. With that principle in mind, the Court looked to the proclamation’s text, which was “expressly premised on legitimate purposes.” Id. at 706. Given that conclusion, the statements of government officials about the proclamation were not dispositive. So too here. Our review must focus on the text and sources informing the Hegseth Policy, not extrinsic statements made about it.

B

The remaining stay factors also favor the government. For one thing, the Supreme Court’s stay order in Shilling indicates that the government would suffer an irreparable injury if the Hegseth Policy were enjoined pending appellate review of preliminary injunctions. See Shilling, 145 S. Ct. 2695; Nken, 556 U.S. at 425–26. This is hardly surprising. Injunctions that “improperly intrude” on the Executive Branch often inflict irreparable injury, Trump v. CASA, Inc., 606 U.S. 831, 859 (2025) (cleaned up), especially where the injunction interferes with the government’s “compelling need to provide for the Nation’s security,” Trump v. Int’l Refugee Assistance Project, 582 U.S. 571, 583 (2017) (per curiam). We can imagine few if any government interests more compelling than a strong and ready military. And while the district court concluded that 23 allowing gender-dysphoric individuals to serve in the military would advance the cause of military strength and readiness, the Constitution assigns that assessment to the political branches, as cases like Goldberg and Goldman make clear.

On the other side of the balance, we acknowledge that the plaintiffs have significant interests at stake—particularly those who have already acceded. But “[t]he possibility that adequate compensatory or other corrective relief will be available at a later date … weighs heavily against a claim of irreparable harm.” Chaplaincy of Full Gospel Churches v. England, 454 F.3d 290, 297–98 (D.C. Cir. 2006) (cleaned up). In particular, loss of a valued government position “does not usually constitute irreparable injury,” especially if the terminated individual or officer may obtain backpay or reinstatement if he or she ultimately prevails on the merits. Sampson v. Murray, 415 U.S. 61, 90–91 (1974). Courts applying these principles have held that even less-than-honorable military discharges do not constitute irreparable injuries warranting interim relief while challenges to the discharge remain ongoing. See Hartikka v. United States, 754 F.2d 1516, 1518 (9th Cir. 1985); Guerra v. Scruggs, 942 F.2d 270, 273–74 (4th Cir. 1991). And here, any service members separated under the Hegseth Policy will be honorably discharged. Add. 101.

Finally, the public interest does not warrant a stay. With the government likely to prevail on the merits, and with the balance of harm among the parties favoring the government, we can discern no other consideration here that would tip the scale in the opposite direction. 24 III

For these reasons, we grant the stay pending appeal.2

2 The district court independently erred in universally enjoining the Hegseth Policy rather than merely providing as-applied relief to the plaintiffs. See CASA, 606 U.S. at 839–47. Given our conclusion that the government should be able to enforce the policy while its appeal is pending, we need not address further questions about the appropriate scope of any injunctive relief. PILLARD, Circuit Judge, dissenting: Transgender members of the military have answered the call to serve this nation on the same terms as any other servicemember. For nearly a decade, they have met the same standards, risen to the same challenges, and defended our country with the same honor as their peers. They have passed intensive physical and mental health screening, completed demanding training programs, and met the rigors of their assigned duties. The achievements of the individual plaintiffs in this case include “more than 80 commendations” during a collective total of “over 130 years of military service.” See Talbott v. United States, 775 F. Supp. 3d 283, 291 (D.D.C. 2025). In the district court, high-ranking officials from the Army, Navy, and Air Force offered unrebutted testimony uniformly supporting the finding that transgender servicemembers “have had either no detrimental effect or have had a positive” effect on military readiness. Id. at 310. Defendants offered no evidence to the contrary.

This suit arises because the government has now decided to force every transgender servicemember out of the military through a policy based on nothing more than negative attitudes about transgender identity. The majority and I agree that we owe the greatest deference to the political branches’ expert judgments on military policy. Maj. 11-12. The Constitution vests control of the military in the political branches. Id. We must respect “the executive’s considered judgment that a military policy furthers legitimate interests.” Id. at 13. The district court, in accord with that deferential standard of review, Talbott, 775 F. Supp. 3d at 311-12, offered the government every opportunity to provide evidence of considered judgment supporting the Hegseth Policy, id. at 309. The government declined to do so. Id.

Underscoring that its actions are heedless of military need, the government cannot even say how many servicemembers its policy will exclude—but evidence suggests the number is in the thousands. Id. at 307; see generally Add. 129 n.10 (citing 2 a 2016 survey estimating that 8,980 active-duty servicemembers, approximately 1% of the service, identified as transgender). Prior defense secretaries in Republican and Democratic administrations adopted varying policies regarding transgender people’s accession. But none categorically barred current servicemembers who are transgender from remaining in their roles so long as they continue to meet military standards.

There may well be valid reasons to reexamine and alter military service policies set by previous administrations. But on this record, one cannot tell. Defendants provide no evidence that they based their new policy on any assessment of costs, benefits, or any other factor legitimately bearing on military necessity. Indeed, there is “no evidence that [President Trump or Secretary Hegseth] consulted with uniformed military leaders” before imposing their unprecedented ban on transgender servicemembers. Talbott, 775 F. Supp. 3d at 292. The outcome was determined by animus from the start.

In Executive Order 14183, President Trump declared as “the policy of the United States” that “adoption of a gender identity inconsistent with an individual’s sex conflicts with a soldier’s commitment to an honorable, truthful, and disciplined lifestyle, even in one’s personal life.” Prioritizing Military Excellence and Readiness, 90 Fed. Reg. 8757, 8757 (Feb. 3, 2025). He asserted that openly identifying as transgender is “not consistent with the humility and selflessness required of a service member.” Id. Having thereby tagged all transgender servicemembers as shameful, dishonest laggards too arrogant or selfish to serve, he ordered Secretary Hegseth to implement policies within sixty days to carry out the Executive Order. Id. at 8757-58. 3 Secretary Hegseth responded by adopting the policy at issue here. Under the Hegseth Policy, all servicemembers who “exhibit symptoms consistent with[] gender dysphoria” or “have a history of” receiving gender-affirming care are summarily “disqualified from military service.” Add. 106. The policy purports to allow “consider[ation] for a waiver on a case-by-case basis, provided there is a compelling Government interest in retaining the servicemember that directly supports warfighting capabilities,” id., but the waiver is illusory. It is unavailable to anyone who has ever “attempted to transition to any sex other than their [birth] sex” and requires a commitment to “adhere to . . . the standards associated with the Service member’s [birth] sex.” Id. The so-called waiver thus excludes any transgender person who has ever attempted to serve or succeeded in serving—or just living—consistently with their gender identity, or who intends to do so.

The Hegseth Policy casts itself as concerned with the “medical, surgical, and mental health constraints on individuals with gender dysphoria.” Id. at 101. But the Policy is not a medical standard. It is a categorical exclusion of transgender people from military service—with “transgender” broadly defined. It covers not only those with diagnosed gender dysphoria, but also those with “a history of” gender dysphoria, or “symptoms consistent with” gender dysphoria. Id. As the district court found, it is a ban on transgender troops writ large. Talbott, 775 F.3d at 301-02; see infra Part I.A. What is more, the Hegseth Policy explicitly bars the military branches from using the military’s own Disability Evaluation System to determine on a case-by-case basis whether a transgender person has a disabling medical condition burdensome to the military or inconsistent with continued service. Add. 106. The government cannot identify any other medical condition that automatically triggers administrative separation instead of an individualized medical evaluation regarding continued fitness 4 to serve. Oral Arg. 27:30-28:30. Why do health concerns stemming from present, past, or suspected gender dysphoria require a process different from what the military uses for every other medical condition affecting servicemembers? The government has no answer.

The record of the military’s prior experience with individualized medical evaluations under Secretaries Carter, Mattis, and Austin undercuts the Hegseth Policy: It reveals that incumbent transgender servicemembers have earned their places in the military and that they can serve without health- related disqualification, Talbott, 775 F. Supp. 3d at 308-09, and often without any period of non-deployability, Wagner Decl. ¶ 14 (App. 332). Secretaries Carter, Mattis, and Austin understood that the military has strong reasons to continue to benefit from incumbent servicemembers’ valuable training and experience, in which the government has made significant investments, rather than arbitrarily removing them from their posts. And they recognized that transgender servicemembers have significant reliance interests in their military positions. See, e.g., Add. 128 (Mattis Policy).

In implementing the Hegseth Policy to exclude even current servicemembers, defendants double down on demeaning their own. The Air Force has released a memorandum requiring transgender servicemembers to either show up to their separation hearings wearing the uniform and meeting the grooming standards of their birth sex—not the uniform they have worn while serving—or else forego the right to participate in their hearings. Memorandum from Brian Scarlett, Acting Assistant Sec’y of the Air Force for Manpower, Pers., and Servs., to All Major Commands, Direct Reporting Units, Field Operating Agencies, & Commanders 4 (Aug. 14, 2024) [hereinafter Air Force Memorandum]. For transgender servicemembers who want to participate in their 5 own separation hearings, that might mean buying a new uniform, shaving their mustaches or shearing their long hair to adhere to different sex-based standards, and dressing in clothing not meant for their medically transitioned bodies. It is hard to see any purpose of those requirements other than humiliating the departing transgender servicemembers and deterring them from exercising their hearing rights. The threat that any transgender servicemember who does not separate voluntarily could face recoupment of past bonuses is added spite. Add. 107.

Based on its review of the uncontested facts of record, the district court preliminarily enjoined the Hegseth Policy. Applying existing, binding law to the uncontested record before us, I must agree with that judgment. I cannot credit the government’s bare assertions over the district court’s findings on the factual record, none of which the government challenges for clear error. The government has not satisfied the demanding standard to stay the district court’s preliminary injunction.

I write with full respect for the signal we have already received from the Supreme Court regarding the Hegseth Policy. Acting on its emergency docket without statement or opinion, the Supreme Court has stayed a separate preliminary injunction against the Hegseth Policy issued by the U.S. District Court for the Western District of Washington. See United States v. Shilling, 145 S. Ct. 2695 (2025) (mem). I am mindful of the Supreme Court’s guidance that its “interim orders . . . inform how [we] should exercise [our] equitable discretion in like cases,” Trump v. Boyle, 145 S. Ct. 2653, 2654 (2025), as well as its instruction that the Court’s “unexplicated decisions” are “not to be read as a renunciation . . . of doctrines previously announced” in its binding decisions, Ramos v. Louisiana, 590 U.S. 83, 104-05 (2020) (internal quotation 6 marks and citations omitted). The Supreme Court’s order in Shilling may well reflect its conclusion that district judges lack authority to grant universal relief. Trump v. CASA, Inc., 606 U.S. 831, 856 (2025).1 And this case, like Shilling, arrives in a preliminary posture, presenting the kind of important, fast- moving issue on which the lower federal courts can contribute to “the airing of competing views” that may “aid[] [the Supreme] Court’s own decisionmaking process” at the merits stage. Dep’t of Homeland Sec. v. New York, 589 U.S. 1173, 1176 (2020) (Gorsuch, J., concurring in grant of stay); see also Arizona v. Evans, 514 U.S. 1, 23 n.1 (1995) (Ginsburg, J., dissenting) (“[D]iverse opinions from[] state and federal appellate courts may yield a better informed and more enduring final pronouncement by this Court.”). As do my colleagues in the majority, I thus find it worthwhile to briefly explain my views on whether the district court’s preliminary injunction should be stayed pending resolution of the appeal on its merits.

I.

Entering a stay pending appeal is an “extraordinary” remedy. Citizens for Resp. & Ethics in Wash. v. FEC (CREW), 904 F.3d 1014, 1017 (D.C. Cir. 2018) (per curiam). In most circumstances, parties must vindicate their positions through the ordinary process of litigation. To stay relief that a district court entered for an opposing party, the applicant must (1) make a “strong showing that [it] is likely to succeed on the merits;” (2) demonstrate that it will be “irreparably injured” before the appeal concludes; (3) show that issuing a stay will not “substantially injure the other parties interested in the

1 As this case proceeds, the district court may need to reassess the scope of its injunction in light of the Supreme Court’s intervening decision in CASA. See Doe v. Trump, 142 F.4th 109, 112 (1st Cir. 2025). 7 proceeding;” and (4) establish that “the public interest” favors a stay. Nken v. Holder, 556 U.S. 418, 434 (2009). We review the district court’s decision to grant the preliminary injunction “for abuse of discretion, its legal conclusions de novo, and [its] factual findings for clear error.” Media Matters for Am. v. Paxton, 138 F.4th 563, 573 (D.C. Cir. 2025).

Even with the benefit of the “great deference” we owe “to the professional judgment of military authorities,” Winter v. Natural Resources Defense Council, Inc., 555 U.S. 7, 24 (2008), defendants have failed to make the necessary “strong showing” that they are likely to succeed on the merits of their claim, Nken, 556 U.S. at 426. Because the Hegseth Policy is openly fueled by animus towards transgender people and defendants have not shown that it is based on military considerations, it fails even the most deferential form of equal protection review. A. “[I]f the constitutional conception of ‘equal protection of the laws’ means anything, it must at the very least mean that a bare . . . desire to harm a politically unpopular group cannot constitute a legitimate governmental interest.” U.S. Dep’t of Agric. v. Moreno, 413 U.S. 528, 534 (1973). Accordingly, even reviewed with “the highest deference” that we owe the political branches “in ordering military affairs,” Loving v. United States, 517 U.S. 748, 768 (1996); see Dep’t of Navy v. Egan, 484 U.S. 518, 530 (1988) (collecting cases), a law or other official action that is driven solely by animus towards a disfavored group cannot survive. Animus fails as a justification for differential treatment even of groups unprotected by any heightened constitutional scrutiny. The Supreme Court has applied rationality review to strike down as animus-based a statutory provision rooted in negative attitudes about “‘hippies’ and ‘hippie communes’” that disqualified 8 unmarried cohabitants from receiving food stamps, Moreno, 413 U.S. at 534 (citing 116 Cong. Rec. 44439 (1970) (statement of Sen. Holland)), the amendment of Colorado’s constitution to prohibit the state from acting to protect gay people from discrimination, Romer v. Evans, 517 U.S. 620, 632-33 (1996), and a zoning ordinance rooted in fears of the mentally disabled, City of Cleburne v. Cleburne Living Center, 473 U.S. 432, 448-50 (1985). Even as the Court in Cleburne declined to apply heightened scrutiny to official action targeting “the mentally retarded,” it invalidated the denial of a permit for a group home because the denial was “rooted in considerations that the Constitution will not tolerate”— namely, “irrational prejudice.” Id. at 446, 450.

Animus need not be explicit on the face of a policy, as it is here, but may be evidenced by “contemporary statements by members of the decisionmaking body.” Village of Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 268 (1977). In Cleburne, for example, the Court relied on the district court’s findings that the Cleburne City Council passed the discriminatory zoning ordinance primarily based on the “negative attitude of . . . property owners . . . ” and “the fears of elderly residents,” 473 U.S. at 448—attitudes and fears accepted by the “five city officials most involved in the decisionmaking process,” Br. for Resp’ts at 34, City of Cleburne v. Cleburne Living Center, 473 U.S. 432 (1985) (No. 84-468). “[C]ircumstantial” factors may provide evidence of decisionmakers’ animus, including “sudden[]” changes in policy (especially unexplained reversals), “[d]epartures from the normal procedural sequence,” failure to address “the factors usually considered important,” Arlington Heights, 429 U.S. at 266-68, and creating a policy whose “sheer breadth” is “discontinuous with the reasons offered” for it, Romer, 517 U.S. at 632. The Defense of Marriage Act, for example, failed rationality review in part because its “unusual deviation from 9 the usual tradition of recognizing and accepting state definitions of marriage” provided “strong evidence of a law having the purpose and effect of disapproval” of same-sex couples. United States v. Windsor, 570 U.S. 744, 770 (2013).

Executive Order 14183, which triggered the Hegseth Policy, is openly hostile to transgender people. The Order denigrates anyone with “a gender identity inconsistent with [that] individual’s sex,” (i.e., anyone who is transgender) as categorically unable to fulfill “a soldier’s commitment to an honorable, truthful, and disciplined lifestyle”—a commitment required “even in one’s personal life.” Prioritizing Military Excellence and Readiness, 90 Fed. Reg. at 8757. The Order asserts that transgender identity is itself a “falsehood . . . not consistent with the humility and selflessness required of a service member.” Id. In other words, Executive Order 14183 vilifies transgender people as dishonorable, dishonest, and undisciplined. It treats their transgender identity as a defect that categorically cancels out their every individual strength, ability, and virtue. The military’s new official policy brands all transgender people, without regard to individual merit, as unworthy to serve in our armed forces solely because they are transgender.

Neither the government nor the majority disputes that Executive Order 14183 expresses animus towards transgender people. See Mot. for Stay 24-25; Maj. 20-22. They focus only on the Hegseth Policy. But that changes nothing. The Hegseth Policy springs from the Executive Order and espouses the same denigration and vitriol. Contextual factors—the statements of triumphant policymakers, the hasty and unsupported reversal of apparently successful policies backed by careful study, and the categorical sweep of the new policy—all confirm that the Hegseth Policy, like the Executive Order, is driven by animus. 10 Start with the substance of the Hegseth Policy. Secretary Hegseth’s memoranda repeat the disparaging claim that people “expressing a false ‘gender identity’ divergent from an individual’s [birth] sex cannot satisfy the rigorous standards necessary for military service.” App. 325 (quoting Executive Order 14183); App. 639-40. Additional implementing documents list military virtues, including “honesty, humility, uniformity, and integrity,” and then flatly declare that anyone who has “symptoms consistent with[] gender dysphoria” categorically lacks such virtues. See, e.g., Add. 101; Add. 114. As the district court concluded, the Hegseth Policy’s “language is unabashedly demeaning [and] its policy stigmatizes transgender persons as inherently unfit.” Talbott, 775 F. Supp. 3d at 326.

Next, consider some of the many statements from official government sources describing the Hegseth Policy as a total ban on transgender service and celebrating the wholesale exclusion of transgender individuals. For instance:

• “Transgender service members are disqualified from service without an exemption.” @DODResponse, X.com (Feb. 27, 2025) (emphasis added), https://perma.cc/S4CE-EPFX.

• “Your agenda is illegals, trans & DEI — all of which are no longer allowed @ DoD.” @PeteHegseth, X.com (Apr. 20, 2025) (emphasis added), https://perma.cc/WX2M-R4K9.

• “No more dudes in dresses, we’re done with that shit.” Pete Hegseth, Sec’y Defense, U.S. Dep’t of Defense, Keynote Address at Special Operations Forces Week 2025 (May 6, 2025), https://perma.cc/T75U-93FR. 11

• “After a SCOTUS victory for @POTUS, TRANS is out at the DOD.” @PeteHegseth, X.com (May 8, 2025) (emphasis added), https://perma.cc/ZUV2- 6QNJ.

Add to defendants’ overt derision their unusually “rushed and haphazard manner” of issuing the Hegseth Policy. Talbott, 775 F. Supp. 3d at 302. Previous defense secretaries devoted months of study, with the help of experts, to understanding the facts and circumstances relevant to transgender service. Id. Secretary Hegseth promulgated his policy in just one month. Id. That was remarkable because, by the time Secretary Hegseth took office, the military was poised to shift from the prognostication on which past administrations had to rely to draw on the military’s own substantial and growing experience with transgender servicemembers.

Before our court, defendants rely almost exclusively on the 2018 Mattis Report and the Policy that Report recommended. Mot. for Stay 13, 23-24. But when Secretary Mattis acted, the military had negligible experience with openly transgender servicemembers. Of necessity he based his policy on “risks” and predictions about what “could” happen gleaned from broader studies not focused on the military. Add. 120. Even then, the Mattis Policy made no move to purge transgender people already serving; it disallowed only new accessions. Add. 120-21. Secretary Hegseth, in contrast, had a wealth of relevant evidence available to him that Secretary Mattis had lacked: the military services’ own experience during the intervening years with the thousands of transgender persons among their ranks. Secretary Hegseth deliberately failed to account for that highly probative experience—or even explain why he eschewed it. Ignoring realities that might undercut his chosen course, Secretary Hegseth rested his policy on 12 categorical disdain. That head-in-the-sand approach yielded a policy driven by animus, not military need.

The majority reverts to three pieces of evidence that defendants had featured in district court: a 2021 review conducted by the Department’s Psychological Health Center of Excellence and the Accession Medical Standards Analysis and Research Activity; a 2025 medical literature review conducted by the Office of the Assistant Secretary of Defense for Health Affairs; and a review of cost data on treatment of active-duty servicemembers currently experiencing gender dysphoria. Maj. 14-15. But the district court found, as to the first two sources, that each “contradicts, rather than supports, the conclusions” the Department drew from it. Talbott, 775 F. Supp. 3d at 305-06. And as to the cost data, the district court explained that the Department could not substantiate its assertion of the disproportionate cost of treating servicemembers with gender dysphoria without any baseline for comparison with costs of other types of covered medical care, or any analysis of the costs associated with discharging all transgender servicemembers. Id. at 306. The government does not challenge the district court’s findings of fact as clearly erroneous, and we are bound to respect them.

Finally, consider the draconian scope of the Hegseth Policy. It imposes a complete ban on military service by transgender individuals. See id. at 301-02. And, unlike every policy since 2016, it does not just affect the standards for prospective recruits but drives out current, active-duty soldiers. It does so in vivid contrast to the Mattis Policy, which recognized that the military generally applies “less stringent standards to retention than to accession in order to preserve the Department’s substantial investment in trained personnel” and to honor the reliance interests of incumbent transgender servicemembers. Add. 127-28; see id. 164 (“[R]etention 13 standards focus squarely on whether the Service member, despite his or her condition, can continue to do the job.”). Accordingly, servicemembers who had been diagnosed with gender dysphoria yet permitted under the Carter Policy to serve openly could remain in the service under the Mattis Policy, continue to receive benefits including medical treatment, and serve consistently with their gender identity. Add. 120. That “reliance exemption” was critical to our court’s understanding that the Mattis Policy did not operate as a blanket ban on transgender servicemembers. See Doe 2 v. Shanahan, 755 F. App’x 19, 24 (D.C. Cir. 2019); Doe 2 v. Shanahan, 917 F.3d 694, 700 (D.C. Cir. 2019) (Wilkins, J., concurring); id. at 722 (Williams, J., concurring). In contrast, the Hegseth Policy not only prevents transgender persons from acceding to the military but also summarily ousts all those already in it. Both groups are disqualified from service without exception. Thousands of talented and patriotic individuals who met rigorous accession standards and have been serving military need are being pointlessly expelled.

The Hegseth Policy also subjects transgender servicemembers to a gratuitously harsh administrative separation process. Normally, when a current servicemember is diagnosed with a medical condition that the military considers potentially disabling, the Department of Defense refers the servicemember to its Disability Evaluation System for an individualized assessment of the servicemember’s continued ability to serve and suitability for deployment. Soper Decl. ¶ 13 (App. 446). The Department uses the results of that individualized assessment to determine whether the servicemember can continue in the military. But the Hegseth Policy specifically bars referrals to the Disability Evaluation System based on gender dysphoria. Instead, it requires transgender servicemembers to undergo administrative separation, a process typically reserved for misconduct or 14 failure to meet standards. Id. at ¶¶ 14, 16 (App. 446). Administrative separation proceedings can cause immediate and severe career harm: Servicemembers in those proceedings are automatically designated non-deployable and cannot be promoted. Id. at ¶ 14.

There is no other medical condition that automatically triggers administrative separation from the military rather than individualized medical evaluation followed by determination of an appropriate response. Oral Arg. 27:30-28:30. Defendants offer no explanation why they devised a new, categorical approach just for gender dysphoria when existing medical screening procedures serve for all other medical conditions. As the district court found, this sui generis treatment of gender dysphoria underscores that the Hegseth Policy’s asserted grounding in “the military’s rigorous physical and mental fitness requirements” is pretextual. Talbott, 775 F. Supp. 3d at 328. Such rigorous fitness “standards are already in place, and each active-duty Plaintiff already meets them.” Id. If servicemembers become unfit or if their participation poses unusual risks to others, the military’s preexisting procedures enable it to assess and remove them.

The district court also found that the government plans to implement the Policy by assigning servicemembers to the task of “rummag[ing] through private medical records” to identify those “who have a current diagnosis or history of, or exhibit symptoms consistent with, gender dysphoria.” Talbott v. United States, 775 F. Supp. 3d 445, 449-50 (D.D.C. 2025) (emphasis added). No previous policy has devised such a costly and intrusive implementation scheme. No other medical condition involves servicemembers scrutinizing their peers’ medical records to root out vaguely defined disfavored symptoms—in the name of unit cohesion, no less. The Hegseth 15 Policy is not designed to further military objectives. It is designed to stigmatize and drive out an unwanted group. B.

In advocating for a stay, the government does not dispute the animus pervading the Executive Order but shrugs it off as legally beside the point. Animus matters to the constitutional analysis, say defendants, only where a policy “lack[s] any purpose other than a ‘bare . . . desire to harm a politically unpopular group.’” Mot. for Stay 25 (quoting Trump v. Hawaii, 585 U.S. 667, 705 (2018) (quoting Moreno, 413 U.S. at 534)). The majority agrees, concluding that “[i]f any legitimate [governmental] interest exists” alongside the animus, “‘we must accept that independent justification.’” Maj. 21 (quoting Trump, 585 U.S. at 706). We need not address whether the presence of animus is fatal to a policy with a genuinely independent justification because here, the government fails to offer any plausible, distinct rationale for its policy.

Trump v. Hawaii exemplifies the requisite inquiry. There, challengers asserted that a Presidential Proclamation barring the entry to the United States of nationals from certain countries was invalid because it was fueled by anti-Muslim animus. Trump, 585 U.S. at 699. In crediting the government with an independent and legitimate national security justification, the Court emphasized that the revised Proclamation before it “reflect[ed] the results of a worldwide review process undertaken by multiple Cabinet officials and their agencies,” limited the ban to “countries that were previously designated by Congress or prior administrations as posing national security risks,” and provided for meaningful exceptions and waivers in keeping with the limited national security rationale. Trump, 585 U.S. at 706-710. 16 But the Court has refused to so credit alternate explanations for policies driven by animus where, as here, those explanations lack indicia of truth and independence. In Moreno, for example, the government claimed the challenged policy was designed to prevent welfare fraud; the Court disbelieved that explanation because other statutory provisions dealt with “those very same abuses.” 413 U.S. at 535-37. In Cleburne, the government claimed its policy forbidding a group home for the developmentally disabled was justified by the possibility that local “students might harass” the disabled residents, that its siting in a floodplain might endanger them, that the city might face legal liability, and that the group home would create residential crowding. 473 U.S. at 449-50. There, too, the Court rejected the alternate explanations because the targeted group presented no “different or special hazard” from groups that remained unregulated. Id. at 449. And in Romer, the Court deemed several explanations pretextual because the challenged policy’s “sheer breadth” in depriving the targeted group of every potential legal protection from discrimination “across the board” was notably “discontinuous with the reasons offered for it.” 517 U.S. at 632-33.

The Supreme Court’s instruction to credit only actual justifications “independent” of animus aligns with broader principles of equal protection analysis. Trump, 585 U.S. at 706. Screening out animus is a chief concern of constitutional scrutiny. The Court reviews suspect classifications such as race and sex more skeptically because they “are so seldom relevant to the achievement of any legitimate state interest that laws grounded in such considerations are deemed to reflect prejudice and antipathy.” Cleburne, 473 U.S. at 440. In other words, the Court treats a suspect category as presumptively based in animus (and therefore unconstitutional) unless the law’s “suitabl[e] tailor[ing]” toward some legitimate and sufficiently weighty purpose shows otherwise. Id. Even under 17 the most deferential rational basis review, the government must do more than supply possible alternate explanations for a policy motivated by animus. It must show that such reasons better explain the policy than the “prejudice and antipathy” already identified. Cleburne, 473 U.S. at 440. To hold otherwise would too readily allow official actors to mask forbidden, irrational animus with pretext.

The government variously asserts that the Hegseth Policy is justified because transgender servicemembers harm military readiness, disrupt unit cohesion, or impose excess medical costs. Mot. for Stay 14-18. But none of those alternate rationales accounts for the focus, conclusory process, or harsh scope of the Hegseth Policy. Animus does. The government has failed to show that the Hegseth Policy was designed to promote its stated interests rather than the inverse: that those interests were retrofitted to justify the animus-driven policy required by Executive Order 14183.

Consider the two potentially strongest justifications for the Policy: improving military readiness and maintaining unit cohesion. The majority notes that previous defense secretaries had “concerns” about transgender persons’ health issues, stressing the predictions of the Mattis Report and the “anticipated” impacts of the Carter Policy. Maj. 15-16 (internal quotation marks omitted). Similarly, the majority draws on the Mattis Report’s discussion of potential unit cohesion issues stemming from gender-based privacy and physical fitness standards. Id. at 16-17. But the predictions and hypotheticals of past defense secretaries were developed nearly a decade ago, without the benefit of experience. As the district court observed, the Mattis Report was explicit that its “findings were limited by uncertainty” because “[t]ransgender persons had not served openly previously.” Talbott, 775 F. Supp. 3d at 303 (internal quotation marks omitted). 18

In adopting the Hegseth Policy, defendants failed to consider whether the “woefully stale” predictions of the Mattis Report have borne out. Id. at 324. The government is silent as to whether, in the intervening seven years, any of the challenges predicted by the Mattis Policy—or others it did not foresee—have arisen and, if so, how the military has addressed them. And the government fails to identify any way in which policies in place on the eve of the Hegseth Policy failed to meet legitimate military needs. Defendants provide no evidence that President Trump or Secretary Hegseth “consulted with uniformed military leaders” at all before imposing their unprecedented ban on transgender servicemembers. Id. at 292.

Contrast defendants’ suppositions with the district court’s findings of fact, based on undisputed record evidence that includes the observations of multiple high-ranking officials who implemented the Austin Policy. Former service secretaries and other leaders of the Department of Defense attested that they had never encountered a disciplinary case, readiness issue, or weakening of unit cohesion stemming from a servicemember’s transgender status. Id. at 309-10. What is more, the former Secretary of the Navy observed that “allowing transgender individuals to serve strengthens unit cohesion by fostering honesty and mutual trust” and that “excluding transgender individuals from military service is destabilizing to good order and discipline.” Id. at 308. The district court credited those officials’ uncontested avowals that “transgender service . . . has not negatively impacted readiness.” Id. at 309. As one official responsible for personnel and readiness explained, “policies that allow transgender service members to be evaluated based on skill and merit . . . do not jeopardize the military’s mission of protecting the United States, but strengthen it.” Id. at 309. 19 In accord with those attestations, the record substantiates that plaintiffs’ service has made our country safer. To name just a few, Captain Gordon Herrero kept American soldiers in the Republic of Korea supplied with necessities during multinational military exercises and was recently selected to educate future officers as an instructor at the United States Military Academy at West Point. Herrero Decl. ¶¶ 3-8, ECF No. 72-25. Major Erica Vandal earned a Bronze Star for her meritorious service as a Battery Commander in Afghanistan. Vandal Decl. ¶¶ 8-9, ECF No. 72-20. Ensign Dany Danridge repaired and maintained Navy aircraft and equipment to support the military campaign against ISIL. Danridge Decl. ¶¶ 11-12, ECF No. 72-28. Commander Michelle Bloomrose, as Commanding Officer of the Naval Justice School, trained hundreds of enlisted and civilian professionals to deliver military legal services. Bloomrose Decl. ¶ 7, ECF No. 72-54. Each has served our country for between nine and nineteen years, each continues to meet the military’s demanding standards, and each wishes to continue serving. These servicemembers’ records, unmet by any contrary showing, support the district court’s conclusion that the service of openly transgender people “has resulted in a stronger, not a weaker military.” Talbott, 775 F. Supp. 3d at 308.

The government raises no challenge to the district court’s findings of fact. Despite being afforded the opportunity to cross-examine those officials and submit its own evidence, the government repeatedly declined to do so (with one immaterial exception).2 The government instead simply ignores the record

2 The government filed one declaration by Timothy Dill, Assistant Secretary of Defense for Manpower and Reserve Affairs, in response to the declaration of Gilbert Cisneros, the former Under Secretary of Defense for Personnel and Readiness. Id. at 309. Mr. Dill challenged whether a person in Mr. Cisneros’s role would have had firsthand knowledge of individualized servicemember complaints. 20 and the district court’s findings about the military’s experience since the Mattis Policy, and the majority does the same. But in the absence of any showing of clear error we are bound to respect those findings. As the case comes before us, the facts conclusively demonstrate that the government’s stated justifications for the Hegseth Policy are unsupported and pretextual.

The gratuitously demeaning nature of the administrative separation process that the military has adopted for transgender servicemembers underscores the point. No special concern for deployability, unit cohesion, training safety, or medical costs supports the use against this group of a separation process designed for cases of misconduct or failure to meet standards. And none of the government’s justifications could support the additional, degrading step of requiring transgender servicemembers to either attend their separation hearings dressed up according to the military standards for their birth sex or else waive the right to participate. See Air Force Memorandum 4. C. Because this policy fails rational basis review, it is unnecessary to explore the deferential version of heightened scrutiny for “the military context” that the majority assumes to be the most demanding standard that might apply here. Maj. 12. The majority’s test, however, notably overreads limited lines of precedent to downplay the role of overt bias or animus. And in any event, it fails to justify upholding the Hegseth Policy.

Id. at 310. But the district court credited Mr. Cisneros’s testimony over Mr. Dill’s, reasoning that Mr. Cisneros’s sworn account of his firsthand experiences was more authoritative than someone else’s suppositions about his experiences, id., and the government does not challenge that determination as clearly erroneous. 21

The majority asserts that, “[i]n the military context,” courts need only “consider whether the policy . . . resulted from the professional judgment of military authorities and reasonably furthers legitimate military interests.” Maj. 12. Neither of the cases it cites supports such a conclusion. Rostker v. Goldberg rejected an invitation to craft a distinct version of intermediate scrutiny of sex-based classifications for military issues, 453 U.S. 57, 69-70 (1981), and in any event concerned Congress’s constitutional authority over military affairs, id. at 64-65. The Rostker Court nonetheless scrutinized Congress’s decision-making to determine whether the contested statute was “the ‘accidental by-product of a traditional way of thinking about females.’” Id. at 74 (quoting Califano v. Webster, 430 U.S. 313, 320 (1977)). Goldman v. Weinberger continued a long line of decisions denying accommodation of certain aspects of religious free exercise in the military. See 475 U.S. 503, 507, 509-10 (1986). To be sure, “the complex, subtle, and professional decisions as to the composition and training of military forces are matters of expert military judgment” on which we “tread with great care,” but no special rule of deference shelters military policies from claims of unconstitutional bias or animus. See Singh v. Berger, 56 F.4th 88, 98 (D.C. Cir. 2022) (internal quotation marks and alterations omitted).

Even under the majority’s standard, the Hegseth Policy falls short: It did not result from the professional judgment of military authorities. We owe deference to the “considered professional judgment” of “appropriate military officials.” Goldman, 475 U.S. at 509. But where defendants made no attempt to collect or analyze the most relevant information available to the military or bring military expertise to bear on the issue before the court, there is no “considered professional judgment” to which we might defer. In Rostker, the Court 22 upheld a policy that was the product of extensive “hearings, floor debate, and . . . committee” discussions which “adduced extensive testimony and evidence concerning the issue.” 453 U.S. at 72. And in Goldman, the Court evaluated “AFR 35-10, a 190-page document” developed by the Air Force which “describe[d] in minute detail” what its servicemembers could wear and the limited “individualized options” allowed. 475 U.S. at 508. Those efforts helped demonstrate that the policies in question were the result of “considered professional judgment,” id., and “reasonable evaluation,” Rostker, 453 U.S. at 68, warranting deference from the courts. There is no record of informed and expert decision making in support of the Hegseth Policy’s abrupt upending of prior policies that were based on careful study.

The Department of Defense may forego consideration, study, or analysis of military experience or needs as it develops and implements its personnel policies. But if it makes that choice, it cannot then demand in litigation that courts evaluating the constitutionality of those policies defer to the Department’s unreasoned—even arbitrary or malign— conclusions. And it certainly cannot do so when it has failed to challenge as clearly erroneous the district court’s findings of fact—findings based on attestations of military officials who implemented the Austin Policy—that allowing open service by qualified transgender individuals has had a positive effect on military readiness and unit cohesion. See Talbott, 775 F. Supp. 3d at 309. Ignoring the overt animus fueling the Hegseth Policy by assuming the Policy is supported by valid military interests that the government did not even bother to substantiate in court would stretch “military deference” beyond any rational meaning. As the Supreme Court noted in Rostker, “deference does not mean abdication.” 453 U.S. at 70.

* * * 23 Neither the government nor the majority contests the existence of animus pervading Executive Order 14183. Mot. for Stay 24-25; Maj. 20-22. The Hegseth Policy carried that animus forward in the draconian ban it imposed, which it then sought to justify based on stale predictions that had supported a significantly more modest policy years earlier. None of the government’s alternate justifications can disguise the real impetus behind the Hegseth Policy: to purge transgender servicemembers from the military because the President vilifies transgender people and wishes to drive them out. The government cannot make a strong showing of likelihood of success on the merits when the Hegseth Policy fails even rational basis review. II.

The government must meet a “high standard” to show that it faces irreparable injury. Chaplaincy of Full Gospel Churches v. England, 454 F.3d 290, 297 (D.C. Cir. 2006). Staying a district court’s grant of preliminary relief “based only on a possibility of irreparable harm is inconsistent with our characterization of injunctive relief as an extraordinary remedy that may only be awarded upon a clear showing that the plaintiff is entitled to such relief.” Winter, 555 U.S. at 22. Instead, the government must show that the impending injury is “actual and not theoretical,” and “of such imminence that there is a clear and present need for equitable relief to prevent irreparable harm.” Id. (quoting Wisc. Gas Co. v. FERC, 758 F.2d 669, 674 (D.C. Cir. 1985) (per curiam)). Failure to demonstrate irreparable harm is, by itself, “fatal” to a request for a stay. See KalshiEx LLC v. CFTC, 119 F.4th 58, 64 (D.C. Cir. 2024).

The majority seeks to justify its stay based on the government’s compelling interest in a “strong and ready military.” Maj. 22. But the record in this case is devoid of 24 evidence that the preliminary injunction allowing continued service by fully qualified, healthy, and ready transgender troops will irreparably harm that interest. The majority simply asserts that in matters of military readiness “the Constitution assigns that assessment to the political branches.” Maj. 23. Indeed, we give “great deference” to the military’s judgments. Winter, 555 U.S. at 24. But the military must provide something more than the bare assertions in its briefing to be “weighed against the [other] interests that are legitimately before this [c]ourt.” See Winter, 555 U.S. at 25.

The government does not here assert that, pending expedited decision of the appeal, the preliminary injunction is likely to undermine military readiness and lethality. Rather, it relies on a generic claim that the Department of Defense suffers irreparably when it cannot immediately carry out a policy it thinks is valid. The majority refrained from embracing that claim. No wonder: That reasoning applies to every injunction against the executive branch, so to treat it as irreparable harm would effectively disable courts from preliminarily enjoining any unconstitutional executive action. “That cannot be so.” Doe #1 v. Trump, 957 F.3d 1050, 1059 (9th Cir. 2020). Even in Maryland v. King, the Chief Justice’s in-chambers stay order on which the government relies, the applicants demonstrated “ongoing and concrete harm.” 567 U.S. 1301, 1303 (2012) (Roberts, C.J., in chambers). Assertions of a generalized interest in the immediate effectiveness of an executive policy that a court deems likely unconstitutional cannot constitute irreparable harm. To hold that it does would negate the irreparable harm factor whenever the government seeks a stay.

Defendants have identified no irreparable harms that manifested under the many years of the Austin Policy. And the district court found that the Mattis Report’s predictions were disproven by the real-world experiences of commanding 25 servicemembers under that policy. Talbott, 775 F. Supp. 3d at 323-25. The military has robust policies in place to remove servicemembers who pose a threat to military readiness and lethality, and transgender servicemembers are governed by those just like everyone else. They are subject to the military’s regular disciplinary procedures. And if they have medical conditions affecting their ability to serve, the military can make individualized assessments through the Disability Evaluation System and remove any transgender servicemember as to whom the assessment confirms a service impediment. The government’s request for a stay therefore fails for lack of irreparable harm. III. The majority contends that a stay pending appeal will not cause “irreparable injuries” to plaintiffs because if they ultimately prevail they may obtain backpay or reinstatement. Maj. 23. “But in this procedural posture, it is not the plaintiffs’ burden to prove irreparable injury; it is the government’s.” J.G.G. v. Trump, No. 25-5067, 2025 WL 914682, at *12 (D.C. Cir. Mar. 26, 2025) (Henderson, J., concurring). Instead, “[w]e must consider whether a stay will ‘substantially injure’ the plaintiffs.” Id. (quoting Nken, 556 U.S. at 434). If a stay is granted, plaintiffs who are currently in the process of joining the military will be prevented from doing so, despite having trained to qualify and given up housing, other jobs, and educational opportunities in reliance on enlistment contracts. And the incumbent active-duty plaintiffs who are being separated from the service are already suffering the loss of a “job, income, pension, health and life insurance, and all the other benefits attendant” with military service, which by itself constitutes substantial and irreparable harm. McVeigh v. Cohen, 983 F. Supp. 215, 221 (D.D.C. 1998); see Elzie v. Aspin, 841 F. Supp. 439, 443 (D.D.C. 1993); see also Cy Neff, US To Begin Immediate Removal of Up to 1,000 Trans Military 26 Members, GUARDIAN (May 8, 2025), https://perma.cc/Q83G- XNNB. Plaintiffs are currently losing the opportunity to accept Senate-confirmed promotions, losing the ability to accrue time towards retirement, losing access to medically necessary healthcare, losing the opportunity to provide benefits for their children, and, in many cases, losing the only career they have ever known—for which many of them undertook years of specialized training. The Hegseth Policy stigmatizes plaintiffs and “subject[s]” them “to the indignity of being unable to serve for reasons that, on this record, bear no relationship to their ability to perform.” Singh, 56 F.4th at 110 (internal quotation marks and alterations omitted). Under our precedent, that is “irreversible and irreparable harm.” Id. Undoubtably, the likelihood of “substantial[] injur[y]” to the plaintiffs weighs heavily in favor of denying the stay. Nken, 556 U.S. at 434. IV. Finally, denying a stay serves the public interest. In Doe 1, we held that allowing a ban on the accession of transgender people to the military to take effect was contrary to the public interest because it would “directly impair and injure the ongoing educational and professional plans of transgender individuals and would deprive the military of skilled and talented service members.” 2017 WL 6553389, at *3. That reasoning has even greater force in this case. Allowing the Hegseth Policy to take effect would lead not only to the end of new accessions, but to the blanket removal of thousands of admittedly qualified and talented servicemembers fully able to continue meeting the military’s demanding standards. Removing those individuals based only on their transgender status is especially detrimental at a time when the services are “facing a reduced pool of American[s] who meet military physical and health standards” and must “build[] and grow[] [their] own skilled specialists and leaders organically,” a process that “typically requires years or decades.” Wagner 27 Decl. ¶ 41 (App. 12-13). The public interest suffers from the arbitrary and demoralizing depletion of our armed forces.

* * * Every factor in this case weighs against granting a stay pending appeal. Defendants have failed to introduce evidence that the Hegseth Policy serves any purpose other than the indulgence of animus. They have not demonstrated that the military is likely to suffer any harm, let alone irreparable harm, if the preliminary injunction remains in place while this case is litigated. On the other side of the ledger, plaintiffs and the public are already suffering substantial and irreparable harm as the military has begun to carry out the Hegseth Policy. All of that is enough to deny defendants’ request for a stay. But denial is even more urgent here to prevent grave disruption to the status quo that preceded the Hegseth Policy: what the record shows to be nearly a decade of successful service in our armed forces by transgender servicemembers. The majority’s decision makes it all but inevitable that thousands of qualified servicemembers will lose careers they have built over decades, drawn up short by a policy that would repay their commitment and service to our nation with detriment and derision. The majority grants this stay in the face of all evidence to the contrary. We should not accord deference to the military when the Department itself carelessly relied on no more than blatant animus.

I respectfully dissent.