Appeal from the United States District Court for the District of Columbia (No. 1:25-cv-00240)

Abhishek Kambli , Attorney, U.S. Department of Justice argued the cause for Appellants. With him on the briefs were Brett A. Shumate , Yaakov M. Roth , Michael S. Raab , Ashley C. Honold , and Amanda L. Mundell.

Shannon Minter , argued the cause for Appellees. With him on the brief were Christopher F. Stroll, Jennifer Levi , Michael Haley , Joseph J. Wardenski , and Sara E. Kropf.

William J. Olson , Jeremiah L. Morgan , Michael Boos , Joseph W. Miller , J. Mark Brewer , Patrick M. McSweeney , and Rick Boyer were on the brief for amicus curiae Lt. Gen. Michael T. Flynn (USA-Ret.), America's Future, Citizens United, Public Advocate of the United States, Public AdvocatePage 2 Foundation, U.S. Constitutional Rights Legal Defense Fund, and Conservative Legal Defense and Education Fund in support of Defendants-Appellants.

Robert S. Chang , Shaleen Shanbhag , Rachel CroskeryRoberts , and Beatrice A. Tice were on the brief for amicus curiae Japanese American Citizens League, The Fred T. Korematsu Center for Law and Equality, and Additional Race Centers in support of Plaintiff-Appellees.

Charity R. Clark , Attorney General of Vermont, Office of the Attorney General for Vermont, Jonathan T. Rose , Solicitor General of Vermont, Ryan P. Kane , Deputy Solicitor General of Vermont, Justin G. Sherman and Samuel Stratton , Assistant Attorneys General of Vermont; and Nicholas W. Brown , Attorney General of Washington, Office of the Attorney General of Washington, Colleen Melody , Civil Rights Division Chief, Cassandra Baker , Assistant Attorney General of Washington were on the brief for amicus curiae States Of Vermont, Washington, Arizona, California, Colorado, Connecticut, Delaware, Hawai'i, Illinois, Maine, Maryland, Massachusetts, Michigan, Minnesota, Nevada, New Jersey, New York, North Carolina, Oregon, and Rhode Island in support of Plaintiff-Appellees.

Before: WILKINS and WALKER, Circuit Judges , and ROGERS, Senior Circuit Judge .

Opinion announcing the judgment of the Court filed by Circuit Judge WILKINS.

Opinion concurring in part and dissenting in part filed by Senior Circuit Judge ROGERS.

Dissenting opinion filed by Circuit Judge WALKER.

Page 3WILKINS, Circuit Judge :

In the beginning days of his second term in office, President Trump issued an Executive Order proclaiming that persons “expressing a false ‘gender identity’ divergent from an individual’s sex cannot satisfy the rigorous standards necessary for military service.' Exec. Order No. 14183, 90 Fed. Reg. 8757 § 1 (Jan. 27, 2025). The President, and later Secretary of Defense Pete Hegseth (now redesignated as Secretary of War), also declared that persons afflicted with gender dysphoria are unfit for military service because, among other things, the character of such persons (in the President’s and Secretary’s words) is 'inconsistent' with the 'high standards . . . [of] honesty, humility, . . . and integrity.' DEP'T OF DEF., ADDITIONAL GUIDANCE ON PRIORITIZING MIL. EXCELLENCE AND READINESS (2025) (hereinafter 'Hegseth Policy'); see also J.A. 50.

In this litigation, the government has not attempted to defend or provide any factual basis for these disparaging characterizations of American citizens. Indeed, the government has not contested that the Plaintiff-Appellees who are currently serving (and who have collectively earned more than 80 commendations) have served honorably and pose no threat to national security, even though they happen to be transgender and have suffered from gender dysphoria.

Instead, the government contends that this case is solely about whether, pursuant to the Hegseth Policy, the military can disqualify persons from military service because they have gender dysphoria, a mental health condition. But the record shows that the purpose of the Hegseth Policy is to target applicants and servicemembers who express what the Administration believes is a 'false gender identity,' and the Page 4Policy goes far beyond disqualifying persons currently or recently suffering from gender dysphoria. Some of those disqualifications are completely unexplained and have no reasonable justification. The sharp contrast to the Mattis Policy, adopted in the first Trump Administration, which allowed servicemembers who were transgender or who had suffered from gender dysphoria to remain in the military, appears to be driven by the bare desire to harm a politically unpopular group: persons who identify as transgender. As such, at this preliminary stage, I conclude that the Hegseth Policy is both arbitrary and based upon animus, and for those reasons the Policy violates Plaintiff-Appellees’ constitutional right to equal protection of the law.

These Plaintiff-Appellees consist both of active-duty service members and those desiring to enlist. Under the Hegseth Policy, those Plaintiff-Appellees currently serving in the military-who have been honorably serving for many years—now face expulsion. In a similar vein, in accordance with the Hegseth Policy, those Plaintiff-Appellees desiring to enlist and who have applied for admission to the military are now excluded from accession. I conclude first that the equities are much stronger in favor of the current servicemembers, as compared to those Plaintiff-Appellees who seek admission, and second that the District Court must carefully consider the implications of compelling the Administration to admit persons to the military based upon a preliminary ruling on the merits that could be revisited and reversed after a full trial on the merits. As a result, and for separate reasons, Judge Rogers and I vote to affirm the District Court's preliminary injunction enjoining the Hegseth Policy as it relates to those PlaintiffAppellees already in the military, while Judge Walker and I vote to vacate the preliminary injunction as it relates to the Plaintiff-Appellees seeking accession into the military.

Page 5After President Trump issued Executive Order 14183 ('EO 14183') on January 27, 2025, entitled 'Prioritizing Military Excellence and Readiness,” Plaintiff-Appellees filed their initial complaint and moved for a preliminary injunction.¹ Plaintiff-Appellees’ Motion for Preliminary Injunction, Talbott v. United States , No. 1:25-cv-00240 (D.D.C. Feb. 2, 2025), Dkt. No. 13. Once the government issued its implementing guidance-otherwise known as the Hegseth Policy 2 -on February 26, 2025, Plaintiff-Appellees filed their Third Amended Complaint, as well as a Renewed Application for Preliminary Injunction. Amended Complaint (Third), Talbott , No. 1:25-cv-00240, Dkt. No. 69; Plaintiffs’ Motion for Preliminary Injunction (Renewed), Talbott , No. 1:25-cv00240, Dkt. No. 72. The Plaintiff-Appellees’ complaint stated that the government violated the equal protection component of the Fifth Amendment because the Hegseth Policy is not based on any legitimate governmental purpose, instead “reflect[ing] animosity toward transgender people because of Page 6their transgender status.'³ Amended Complaint (Third), Talbott , No. 1:25-cv-00240, Dkt. No. 69 at 6, 68-72. The District Court heard oral argument on Plaintiff-Appellees' preliminary injunction motion over the course of three separate days, ultimately issuing the injunction on March 18, 2025. Talbott v. United States , 775 F. Supp. 3d 283 (D.D.C. 2025).

In its memorandum opinion, the District Court made a number of findings that are relevant to the proceedings before us. First and foremost among them-driving its ultimate conclusion enjoining the Hegseth Policy-was the District Court’s finding that the “Military Ban [referring to both EO 14183 and the Hegseth Policy together] is soaked in animus and dripping with pretext.” Talbott , 775 F. Supp. 3d at 326. Further, the District Court found that the Hegseth Policy, which disqualifies from service anyone with a history of gender dysphoria, constitutes a ban on all transgender troops even though it never utilizes the word 'transgender.' Id . at 301. The District Court ruled that the Policy’s purported exemption was 'one in name only,' because '[v]irtually no one' who was transgender could meet the criteria for exemption. Id . at 301– 02. Underscoring this finding was the District Court's understanding that the Hegseth Policy, which even disqualifies those who “exhibit symptoms consistent with [] gender dysphoria,' was an 'exclusion . . . so broad as to capture persons who have never had gender dysphoria[.]” Id. at 324 (emphasis in original). Additional findings from the District Court included that the Administration had 'rushed' the Page 7Hegseth Policy when it reversed an existing military policy without 'comprehensive review,' id . at 302, and that the studies relied upon by the government to justify the Hegseth Policy did not support the government’s actions in disqualifying transgender personnel from serving. Id . at 303– 308.

The District Court then turned to the preliminary injunction analysis, holding that Plaintiff-Appellees had a likelihood of success on their equal protection claim, that they would suffer irreparable harm absent injunction, and that the balance of equities and public interest favor PlaintiffAppellees. Talbott , 775 F. Supp. 3d at 311–33. In analyzing Plaintiff-Appellees’ equal protection claim, the District Court applied intermediate scrutiny, as it found that the Hegseth Policy discriminated on the basis of sex. Id. at 315–16. The District Court's conclusion ultimately relied on the Supreme Court’s decision in Bostock v. Clayton County , 590 U.S. 644 (2020), which held in the Title VII context that it was “impossible to discriminate against a person for being homosexual or transgender without discriminating against the individual based on sex.” Id . at 316 (citing Bostock , 590 U.S. at 660). Alternatively, the District Court also found that intermediate scrutiny should apply because the Hegseth Policy discriminated on the basis of transgender status, and transgender persons are a quasi-suspect class. Id . at 319–22. Under the more exacting standard, the District Court held that even with the heightened deference that the judiciary provides to the military, the military’s stated legitimate goals of military readiness, unit cohesion, good order, and discipline were not substantially related to the means utilized to achieve those goals in the Hegseth Policy. Id. at 322–26.

Even assuming that the Hegseth Policy discriminated on the basis of a medical condition, as the government contended Page 8throughout the course of the lower-court proceedings, and even assuming that the more deferential rational basis standard of review applied, the District Court found that the Hegseth Policy still violated equal protection because it was an “unadulterated expression of animus” and reflective of a “bare … desire to harm a politically unpopular group.” Id . at 327 (citing U.S. Dep't of Agric. v. Moreno , 413 U.S. 528, 534 (1973)). To arrive at this conclusion, the District Court relied upon the Hegseth Policy, as well as EO 14183, a White House Fact Sheet that accompanied that Executive Order, Secretary Hegseth’s contemporaneous tweet regarding the Policy, as well as other factors demonstrating that the government 'has targeted transgender persons writ large.” Talbott , 775 F. Supp. 3d at 330 (citation modified).

Three days after the District Court issued its memorandum opinion, the government filed a motion to dissolve the preliminary injunction, highlighting DoD's newly issued March 21, 2025 guidance. Defendants' Motion to Dissolve Preliminary Injunction, Talbott , No. 1:25-cv-00240, Dkt. No. 91 at 2 (citation modified). The government explained that the phrase “exhibit symptoms consistent with gender dysphoria” refers only to the “diagnostic criteria outlined in the Diagnostic and Statistical Manual of Mental Disorders and applies only to individuals who exhibit such symptoms as would be sufficient to constitute a diagnosis[.]' Id. (quoting March 21, 2025 guidance) (internal quotations omitted). The District Court held a hearing on the motion that same day, and denied the motion on March 26, 2025, finding that the government had 'regurgitat[ed]' its arguments from before, and that nothing relevant had changed that prompted the need to dissolve the preliminary injunction. Talbott v. United States , 775 F. Supp. 3d 445, 447 (D.D.C. 2025). The District Court also denied the government’s request for a stay pending appeal but temporarily stayed its injunction until March 28. Id.

Page 9Appellants then filed an emergency appeal in this Court for a stay of the preliminary injunction pending appeal. We granted an administrative stay on March 27, 2025, heard oral argument on the stay pending appeal the following month, and then subsequently granted that stay in December. J.A. 1311; Talbott v. United States , No. 25-5087, 2025 WL 3533344 (D.C. Cir. Dec. 9, 2025) (Motions Panel) (Pillard J., dissenting) (per curiam).

We now have before us the appeal of the preliminary injunction on the merits. For the reasons articulated below and in the separate opinions, we affirm the preliminary injunction as it pertains to the Hegseth Policy for current servicemembers, but we vacate the preliminary injunction with respect to the provisions of the Hegseth Policy applicable to those persons seeking to enlist in the military.

The government contends that this case is solely about whether, pursuant to the Hegseth Policy, the military can disqualify persons from military service because they have a mental health condition. As the government concedes in its opening brief, '[i]n general, the military has aligned . . . disqualifying [mental health] conditions with the ones listed in the Diagnostic and Statistical Manual of Mental Disorders (DSM), published by the American Psychiatric Association.” Appellants' Br. at 5. However, the record shows that the purpose of the Hegseth Policy is to target applicants and servicemembers who express what the Administration believes is a 'false gender identity,' see Exec. Order No. 14183, 90 Fed. Reg. 8757, and the Policy includes disqualifiers that go far beyond persons currently or recently suffering from gender Page 10dysphoria. As such, the Hegseth Policy implicates much more than just those mental health conditions appearing in the DSM.

First, we should define terms. As stated in the 2018 report by then-Defense Secretary James Mattis in the first Trump Administration, when using the term “sex” or “biological sex,” we refer to “a person’s biological status as a male or female based on chromosomes, gonads, hormones, and genitals (intersex is a rare exception).” J.A.⁷⁸ n.10 (quoting AGNES GEREBEN SCHAEFER ET AL., RAND NAT‘L DEF. RSCH. INST., Assessing the Implications of Allowing Transgender Personnel to Serve Openly , 75 (RAND Corporation 2016), https://www.rand.org/content/dam/rand/pubs/research reports /RR1500/RR1530/RAND_RR1530.pdf (“RAND Study”)). 'Transgender' means an 'individua[l] who ha[s] [a] sexual identity or gender expression that differs from their assigned sex at birth.” Id . (quoting RAND Study at 75). A subset of transgender persons often have a condition called gender dysphoria, which was defined in the fifth edition of the DSM as a “marked incongruence between one’s experienced/expressed gender and assigned gender' that 'is associated with clinically significant distress or impairment in social, occupational, or other important areas of functioning.” Id . (quoting AM. PSYCHIATRIC ASS’N, Diagnostic and Statistical Manual of Mental Disorders (DSM-5) , 452–53 (5th ed. 2013)).

Some transgender persons, especially those diagnosed with gender dysphoria, choose to “transition” to their preferred gender. Id . at 78-79. As explained in the Mattis Report, transition can occur in various ways. “Social transition” occurs when the person lives and works in their preferred gender without hormone treatment or surgery. Id . at 93. 'Medical Page 11transition” refers to using cross-sex hormone therapy and hair removal to align secondary sex characteristics, such as breast development, facial hair or voice depth, with the person's preferred gender. Id . 'Surgical transition,' also called sex reassignment surgery, involves surgery on the genitalia and reproductive organs to make them resemble those of the preferred gender as closely as possible. Id .

Prior to 2016, the military barred persons diagnosed with “transsexualism,” as defined in the third edition of the DSM, or persons with “gender identity disorder,” as defined in the fourth edition of the DSM. Id . at 78, 81-82. 'Transsexualism is a gender identity disorder, the sufferers of which believe that they are ‘cruelly imprisoned within a body incompatible with their real gender identity.'' Farmer v. Moritsugu , 163 F.3d 610, 611 (D.C. Cir. 1998) (quoting THE MERCK MANUAL OF MED. INFO. 418 (1997)); see also Farmer v. Hawk-Sawyer , 69 F. Supp. 2d 120, 121 n.1 (D.D.C. 1999). (“Transsexualism is a condition that exists when a physiologically normal person is extremely uncomfortable and discontent with his or her particular sex and prefers to be the other sex.'). 'Gender identity disorder is most simply described as an individual’s confusion or discomfort about his or her sexual status as a biological male or female.” Hawk-Sawyer , 69 F. Supp. 2d at 122. Prior to 2016, the military “[a]lso disqualified . . . persons who had undergone genital surgery or who had a history of major abnormalities or defects of the genitalia.” J.A. 78; see also id . at 81–82. These two categories of disqualifiers, based on medical diagnoses and medical treatment, had the effect of 'prevent[ing] transgender persons, especially those who had undergone a medical or surgical gender transition, from accessing into the military, unless a waiver was granted.” Id . at 78.

Page 12As discussed in the Mattis Report, the psychiatric community significantly modified the DSM when the manual’s fifth edition was published in 2013:

[I]t changed “gender identity disorder” to 'gender dysphoria' and designated it as a “condition”—a new diagnostic class applicable only to gender dysphoria—rather than a “disorder.” This change was intended to reflect the [American Psychiatric Association’s] conclusion that gender nonconformity alone — without accompanying distress or impairment of functioning -was not a mental disorder .

Id . at 83 (emphasis added). As one of our sister Circuits has explained, “[a]lthough the DSM once considered ‘transsexualism’ and ‘gender identity disorder’ mental conditions, the modern edition of the DSM explains that gender nonconformity is not in itself a mental disorder.' Cath. Charities of Jackson, Lenawee, & Hillsdale Cntys. v. Whitmer , 162 F.4th 686, 698 (6th Cir. 2025) (Bloomekatz, J. dissenting in part, and dissenting in judgment).

In 2015, then-Secretary of Defense Ashton Carter initiated a working group to identify issues related to open military service by transgender persons. As later described by Secretary Mattis, '[t]his reevaluation . . . was prompted in part by amendments to the DSM that appeared to change the diagnosis for gender identity disorder from a disorder to a treatable condition called gender dysphoria.” J.A. 83.

After a year of research, Secretary Carter’s working group concluded that transgender persons should be allowed to serve Page 13openly in the military, as the working group's findings demonstrated that doing so would have “minimal impact on unit cohesion,” a “negligible” impact on unit readiness, and any health care costs would be an “exceedingly small proportion” of military health care expenditures. Id . at 570-71, 649; see also id . at 84-85. The DoD then issued the Carter Policy, allowing transgender individuals already enlisted in the military to serve openly starting in summer 2016. Id . at 83, 85. The Carter Policy highlighted that for those individuals who were diagnosed with gender dysphoria while in the military, they could begin a treatment plan that could result in the member transitioning genders. Id . at 85-86. Such treatment culminated in the member requesting a change of gender marker on the Defense Enrollment Eligibility Reporting System ('DEERS'), with the servicemember beginning to serve in accordance with all standards associated with the member's transitioned gender. Id . at 86.

The Carter Policy also allowed the accession of transgender persons into the military starting in July 2017. Id . at 86, 88. To the extent that any person seeking to enlist had a history of gender dysphoria, such medical condition was disqualifying, unless the applicant had been stable (i.e . , 'without clinically significant distress or impairment”) for 18 months. Id . at 86. With respect to past medical treatment related to gender transition, the Carter Policy disqualified persons from joining who had a “history of medical treatment associated with gender transition,” unless (a) the applicant had completed all medical treatment associated with gender transition, (b) the applicant was stable in the preferred gender for 18 months, and (c) if the applicant was presently receiving cross-sex hormone therapy following the gender transition, the individual had been stable on such hormones for 18 months. Id . Similarly, a “history of sex reassignment or genital reconstruction surgery' was disqualifying unless 18 months had passed since the date of any Page 14such surgery, 'no functional limitations or complications' persisted, and no further surgery was required. Id . at 86–87. In other words, the Carter Policy disqualified persons with past genital surgery or medical treatments related to gender transition while they were still in the midst of medical or surgical transition and not yet 'stable' in their preferred gender. Once such stability was demonstrated, the individual was free to enlist and serve.

C. Transgender Policy from First Trump Administration to the Biden Administration

Before the Carter accession policy could go into effect, however, the Administration changed, and the new Secretary of Defense, James Mattis, deferred the July 1, 2017 date to January 2018 so that the military could review their accession plans. A few months after the election, in August 2017, President Trump issued a memo banning members of the military from serving in a sex different from their sex assigned at birth (hereinafter '2017 Memo'). This ban was enjoined by several courts—including the District Court for the District of Columbia-and this Court declined to stay that injunction pending appeal. Doe 1 v. Trump , No. 17-5267, 2017 WL 6553389, at *1 (D.C. Cir. Dec. 22, 2017).

For the next several months following the 2017 Memo, Defense Secretary Mattis convened a panel of military and medical experts, and he issued a new transgender policy in early 2018. The Mattis Report, which formed the basis for the Mattis Policy, began with the “fundamental premise” that “any eligible individual who can meet the high standards for military service without special accommodations should be permitted to serve. This is no less true for transgender persons than for any other eligible individual.' J.A. 73. The Mattis Report recognized that 'transgender status alone is not a [mental Page 15health] condition[,]” while “[g]ender dysphoria, by contrast, is a mental health condition that can require substantial medical treatment.' Id . at 91. Accordingly, the Mattis Policy expressed service qualifications strictly in terms of gender dysphoria and associated medical treatment, rather than on whether an applicant or servicemember expressed a certain gender identity.

Further in accordance with the Report’s identified “fundamental premise,” the Mattis Policy allowed transgender persons who had not transitioned to another gender and who had no history of gender dysphoria to join the military, so long as they were willing and able to meet the standards of their biological sex, a requirement for accession consistent with the Carter Policy. Id . at 75. Transgender persons who required gender transition or who had undergone transition were disqualified from joining the military. Id . at 103.⁴ Transgender persons with a history of gender dysphoria were disqualified from joining the military, unless the person could “demonstrate 36 consecutive months of stability (i.e., absence of gender dysphoria) immediately preceding their application[.]” Id . at 76. The Mattis Policy used the 36-month stability period as the standard, rather than disqualifying persons with a history of gender dysphoria altogether, because 36 months of stability '[wa]s the same standard the Department . . . applie[d] to persons with a history of depressive disorder.' Id . at 113. The Mattis Policy also explained that '[g]iven the documented fluctuations in gender identity among children, a history of gender dysphoria should not alone disqualify an applicant seeking to access the Armed Forces.” Id .

Page 16Secretary Mattis handled a diagnosis of gender dysphoria differently for persons who were already serving in the military. As the Mattis Policy explained:

Retention standards are typically less stringent than accession standards due to training provided and on-the-job performance data. While accession standards endeavor to predict whether a given applicant will require treatment, hospitalization, or eventual separation from service for medical unfitness, and thus tend to be more cautious, retention standards focus squarely on whether the Service member, despite his or her condition, can continue to do the job. This reflects the Department’s desire to retain, as far as possible, the Service members in which it has made substantial investments and to avoid the cost of finding and training a replacement . To use an example outside of the mental health context, high blood pressure does not meet accession standards, even if it can be managed with medication, but it can meet retention standards so long as it can be managed with medication.

Id . (emphasis added).

Accordingly, while a current diagnosis of gender dysphoria was disqualifying for accession, a person already serving could be retained in the military without the need for a waiver, so long as they were willing and able to serve under the standards of their biological sex, did not require gender transition, and were not non-deployable for more than 12 months (or the applicable policy of their Service branch). Id .

Page 17To the extent a person with a history of gender dysphoria sought a waiver from these standards, the Mattis Policy provided that the gender dysphoria standards “are subject to the same procedures for waiver as any other [mental health] standards.' Id .

Secretary Mattis also made a genuine effort to address the reliance interests of those servicemembers who had, prior to the Mattis Policy, already been diagnosed with gender dysphoria and begun medical or surgical transition to their preferred gender. For those servicemembers, the Mattis Policy recognized the “reasonable expectation of these Service members that the Department would honor their service on the terms” that existed under the Carter Policy, and it allowed them to continue to receive all medically necessary care, change their gender marker in DEERS, and serve in their preferred gender. Id . at 76-77, 114. Also grounding this policy was the Department’s “commitment to these [s]ervice members, including the substantial investment it has made in them,” and the Department's judgment that those factors 'outweigh[ed] the risks identified in [its] report.” Id . at 114.⁵

When President Biden took office three years later, he directed the Secretaries of Defense and Homeland Security to ensure that all transgender individuals who could meet military standards should be able to serve. The Office of the Page 18Undersecretary of Defense for Personnel and Readiness created a working group, and based on the group’s conclusions, Defense Secretary Lloyd Austin released revised accession and retention transgender-related guidelines. The Austin Policy, released in April 2021, allowed transgender personnel to openly serve again, and it mirrored the Obama-era Carter Policy in all material aspects. Talbott , 775 F. Supp. 3d at 295– 96.

When President Trump returned to office in January 2025, military accession and retention policies changed again. But instead of reverting back to the Mattis Policy from his first term, President Trump went much further. On his first day in office, President Trump issued an Executive Order denouncing transgender people and the whole concept of transgender identity as inconsistent with “biological truth.” That framing formed the basis of the new military policy that emerged just one month later.

On Inauguration Day, President Trump issued Executive Order 14168 (“EO 14168”), which emphatically rejected the legitimacy of people who identify as transgender and withdrew federal recognition of transgender people.⁹⁰ Fed. Reg. 8615 (Jan. 20, 2025). The President decried 'gender ideology,' which he defined as “replac[ing] the biological category of sex with an ever-shifting concept of self-assessed gender identity, permitting the false claim that males can identify as and thus become women and vice versa . . . .” Id. at 8615 § 2(f). The President described transgender people as employing “gender ideology extremism,” because they are “ideologues who deny the biological reality of sex[.]” Id . at 8615 § 1. The central tenet of EO 14168 was that transgender people perpetrate a lie;Page 19 indeed, part of the title of the Executive Order was “Restoring Biological Truth to the Federal Government.' Id .

Just one week later, President Trump demanded that the military implement his vision of 'biological truth.' In EO 14183, issued January 27, 2025, the President adopted the definitions of EO 14168 and complained that 'the Armed Forces have been afflicted with radical gender ideology. . . .” 90 Fed. Reg. 8757 § 1. The President declared further that “ expressing a false ‘gender identity’ divergent from an individual ' s sex cannot satisfy the rigorous standards necessary for military service.'⁶ Id . (emphasis added). The President further proclaimed that “adoption of a gender identity inconsistent with an individual's sex' conflicts not only “with a soldier’s commitment to an honorable, truthful, and disciplined lifestyle,” but also with “the humility and selflessness required of a service member.' Id . (emphasis added).

In sum, the Commander-in-Chief declared transgender people as categorically unfit for military service explicitly because of their gender identity. To add insult, the President labeled transgender persons as dishonorable, undisciplined, arrogant, selfish liars. The President directed the Secretary of Defense to issue guidance implementing EO 14183 within 30 days. Id . at 8758 § 4(c)(i).

Page 20Less than two weeks later, Defense Secretary Pete Hegseth reiterated the President's view that '[e]xpressing a false 'gender identity' divergent from an individual's sex cannot satisfy the rigorous standards necessary for Military Service.” PRIORITIZING MIL. EXCELLENCE AND READINESS (Feb. 7, 2025), Talbott , No. 1:25-cv-00240, Dkt. 33-1, ('Feb.⁷ Memo'). Thus, once again, the purpose of the Policy was expressed in terms of expelling from the military all transgender persons (i.e., people who “express a false gender identity'). To begin implementing this purpose, Secretary Hegseth mandated that “all new accessions for individuals with a history of gender dysphoria are paused, and all unscheduled, scheduled, or planned medical procedures associated with affirming or facilitating a gender transition for Service members are paused.” Id .

The extirpation from the military of all persons with a “false gender identity” was Secretary Hegseth’s explicit goal. Targeting persons with a history of gender dysphoria was merely the means of achieving that end.

On February 26, 2025, Secretary Hegseth issued the Hegseth Policy, pursuant to EO 14183 and his February 7 Memo, stating that “the medical, surgical, and mental health constraints on individuals who have a current diagnosis or history of, or exhibit symptoms consistent with, gender dysphoria are incompatible with the high mental and physical standards necessary for military service.” J.A. 48. Under the terms of the Hegseth Policy, persons with a history of medical or surgical transition are disqualified from service. Id . at 5354. Applicants and servicemembers are required to self-report any past diagnosis of gender dysphoria or if they are experiencing symptoms “consistent with[] gender dysphoria,”Page 21 the latter of which leads to a medical examination.⁷ See id . at 1290; Id . at 1302–03. The Hegseth Policy declared that people with past or present gender dysphoria do not meet the “high standards for Service member readiness, lethality, cohesion, honesty, humility, uniformity, and integrity.' Id . at 50.

The Hegseth Policy’s perspective of gender dysphoria was completely different from that of the Mattis Policy during the first Trump Administration. Secretary Mattis had explained that because the “vast majority” of applicants aged 17 to 24 are ineligible to join the military for one reason or another, a disqualification based on gender dysphoria should not be seen as a stigma:

[N]othing in the policy should be viewed as reflecting poorly on transgender persons who suffer from gender dysphoria , or have had a history of gender dysphoria, and are accordingly disqualified from service. . . . Transgender persons with gender dysphoria are no less valued members of our Nation than all other categories of persons who are disqualified from military service.”

Id . at 77 (emphasis added). Secretary Hegseth, consistent with President Trump, proclaimed the opposite. The Hegseth Policy continued the character attacks asserted in EO 14183, namely that persons afflicted with gender dysphoria are unfit for military service because, among other things, the character of such persons (in the President's and Secretary's words) is 'inconsistent' with the 'high standards . . . [of] honesty, humility, . . . and integrity.” See id . at 50.

Page 22Accordingly, the Hegseth Policy did not treat gender dysphoria the same as all other mental health conditions, as the Mattis Policy had. While the Mattis Policy evaluated the fitness of those servicemembers with gender dysphoria by placing them in the Disability Evaluation System for a case-bycase assessment (the same as any other medical condition), see id . at 453, the Hegseth Policy mandated that all disqualified persons go through the administrative separation process, which is generally used for personnel accused of misconduct. Id . at 55–56; Id . at 1315; Id . at 1320; Id . at 543–44; see also Talbott , 775 F. Supp. 3d at 448. While the Mattis Policy disqualification for accession was based on a diagnosis of gender dysphoria within the past three years (the same as other depressive disorders), the Hegseth Policy disqualified all persons with any history of gender dysphoria, regardless of how long ago the person had been stable and symptom-free— even if they were diagnosed as a child. See J.A. 50; Id . at 203; Id . at 1092. And finally, while the Mattis Policy allowed waiver of its disqualifications on a case-by-case basis, the Hegseth Policy precludes anyone—regardless of whether they are currently serving in the military or not—from obtaining a waiver unless they can demonstrate that they have never 'attempted to transition,' see id . at 210, which appears to exclude anyone who has ever tried to socially transition in their lifetime (such as a biological woman wearing slacks and a tie). 8Page 23 Appellants have never articulated any basis, medical or otherwise, for such a broad waiver disqualification based on past experience with social transition.

The differences between the Mattis Policy and Hegseth Policy abound, but the divergence is most stark when considering how each Policy treated those diagnosed with gender dysphoria while serving. Unlike the Mattis Policy, the Hegseth Policy makes no real accommodation of the reliance interests of persons with gender dysphoria or persons who had transitioned while serving their country honorably and without incident. The Mattis Policy allowed persons with gender dysphoria who had not begun transitioning to remain in the military, so long as they were willing and able to serve in their biological sex. The Hegseth Policy categorically disqualifies them. The Mattis Policy allowed servicemembers who had transitioned or were in the process of transitioning to remain in the military. Again, the Hegseth Policy categorically excludes them. In response to questions about why these servicemembers were not 'grandfathered' as in the Mattis Policy, the Administration’s response in FAQs accompanying the Hegseth Policy was that:

While these individuals have volunteered to serve our country and will be treated with dignity and respect, expressing a false ‘gender identity' divergent from an individual's sex cannot satisfy the rigorous standards necessary for military service.

Page 24J.A. 207 (emphasis added).

While the Hegseth Policy elsewhere cited costs and readiness as reasons for not retaining servicemembers with gender dysphoria, see id . at 65, that statement cannot explain all of the broad disqualifications, such as expelling someone because they had gender dysphoria as a child or because the person once wore clothing or used pronouns associated with a gender other than their biological sex. From all of the circumstances, including the FAQ response, it is clear that the Hegseth Policy was based, at least in part, on an intent to exclude persons from the military due to their gender identity.

Under the standard set forth in Winter v. Natural Resources Defense Council, Inc. , a plaintiff seeking a preliminary injunction must establish: (1) that they are likely to succeed on the merits; (2) that they are likely to suffer irreparable harm in the absence of preliminary relief; (3) that the balance of equities tips in their favor; and (4) that an injunction is in the public interest. 555 U.S. 7, 20 (2008). This Court reviews the District Court's factual findings for clear error, see Serono Lab’ys, Inc. v. Shalala , 158 F.3d 1313, 1318 (D.C. Cir. 1998), '[its] legal conclusions de novo ,' and 'its weighing of the four relevant factors for abuse of discretion.” In re Fed. Bureau of Prisons’ Execution Protocol Cases , 955 F.3d 106, 111–12 (D.C. Cir. 2020) (per curiam).

To determine whether the Plaintiff-Appellees have demonstrated a likelihood of success on the merits of their equal protection challenge (and assuming we will not employ Page 25any form of heightened scrutiny), we must analyze whether: (1) the government has set forth a legitimate state interest in distinguishing the military eligibility of persons excluded from military service pursuant to the Hegseth Policy from the eligibility of those who do not face such exclusion, and (2) the Hegseth Policy is rationally related to a legitimate government interest. City of Cleburne, Tex. v. Cleburne Living Ctr., 473 U.S. 432, 446 (1985). “A bare … desire to harm a politically unpopular group” can never “constitute a legitimate governmental interest.' Moreno , 413 U.S. at 535; see also Cleburne , 473 U.S. at 447; Trump v. Hawaii , 585 U.S. 667, 705 (2018).

Given that this case concerns a policy decision by the military, we will provide “great deference to the professional judgment of military authorities concerning the relative importance of a particular military interest.' Goldman v. Weinberger , 475 U.S. 503, 507 (1986) (emphasis added). But that does not mean, as the Dissent seems to suggest, see Dissenting Op. at 29–44, that when an Article III court is tasked with determining the constitutionality of a military policy, our role is transformed into that of a judicial rubber stamp. While “the tests and limitations to be applied may differ because of the military context[,] [w]e of course do not abdicate our ultimate responsibility to decide the constitutional question . . . .” Rostker v. Goldberg , 453 U.S. 57, 67 (1981). F 9

Page 26In Goldman , the Court was called upon to determine whether an Air Force regulation mandating uniform dress violated the First Amendment because it precluded a Jewish officer from wearing a yarmulke while in uniform. 475 U.S. at 504. While the Court gave great deference to the military judgment about the 'relative importance' of the interest in uniform dress, the Court also made clear that deference to this military interest “d[id] not, of course, render entirely nugatory in the military context the guarantees of the First Amendment.” 475 U.S. at 507. Thus, the Court carefully reviewed the specific requirements of the regulation and evaluated its constitutionality using the test of whether the classification 'reasonably and evenhandedly regulated' military dress. Id . at 510 (“[W]e hold that those portions of the regulations challenged here reasonably and evenhandedly regulate dress in the interest of the military's perceived need for uniformity.”).

Something akin to Goldman’s “reasonable and evenhanded” test from the First Amendment context has been used to evaluate equal protection challenges to military policies. In Schlesinger v. Ballard , 419 U.S. 498 (1975), a male naval lieutenant who twice failed to be selected for promotion challenged a statute that mandated his separation from the military after nine years of service, where the statute applicable to women who twice failed to be promoted did not mandate their separation until after thirteen years of service. The Court gave great deference to the military’s interest in the 'up or out' policy, because '[i]n the absence of some mandatory attrition of naval officers, the result would be Page 27stagnation of promotion of younger officers and disincentive to naval service.' Id . at 502.

The Court was careful to note that the sex-based classification was not based on “archaic and overbroad generalizations” about the sexes, but rather on “the demonstrable fact that male and female line officers in the Navy are not similarly situated with respect to opportunities for professional service.' Id . at 508. Because women officers had restrictions on participation in combat and most sea duty (which were not challenged), the Court concluded that “Congress may thus quite rationally have believed that women line officers had less opportunity for promotion than did their male counterparts, and that a longer period of tenure for women officers would, therefore, be consistent with the goal to provide women officers with 'fair and equitable career advancement programs.’” Id . (quoting H.R. Rep. No. 216, 90th Cong., 1st Sess., 5). The Court thus upheld the classification as “rational,” observing that '[t]he complete rationality of this legislative classification is underscored by the fact that in corps where male and female lieutenants are similarly situated, Congress has not differentiated between them with respect to tenure.' Id . at 508–09. In sum, like in Goldman , the Court concluded that the military classification was reasonable and evenhanded.

Rostker followed a similar analytical path. There, the Court considered a due process challenge (essentially equivalent to an equal protection challenge) to a statute that authorized the President to require the registration of males for the draft, but not females. See 453 U.S. at 59. Once again, the Court gave great deference to Congress’s determination about the composition of the military, id . at 64-67, but the Court was careful to note that “[n]one of this is to say that Congress is free to disregard the Constitution when it acts in the area of military affairs.” Id . at 67. The Court observed that “[t]he Page 28reconciliation between the deference due Congress and our own constitutional responsibility is perhaps best instanced in Schlesinger v. Ballard ,” id . at 70, and the Court evaluated the statute in a manner similar to the analysis described above. The Court pointed out that Congress had extensively studied and debated the issue before enacting the classification, thus rejecting the challengers’ assertion that Congress acted 'unthinkingly' or 'reflexively and not for any considered reason.” Id . at 72 (internal citations omitted). Because of the extensive hearings and study, the Court reasoned that the classification “was not 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)) (internal citations omitted).

The Court declined to specify whether it was applying heightened scrutiny due to the sex-based nature of the classification or rational basis scrutiny due to deference owed Congressional judgments about the military, because “[a]nnounced degrees of ‘deference’ to legislative judgments, just as levels of ‘scrutiny’ which this Court announces that it applies to particular classifications made by a legislative body, may all too readily become facile abstractions used to justify a result.” Id . at 69–70. Instead, the Court focused on whether, when acting within its constitutional authority, Congress transgressed the guarantee of equal protection under the law. Id . Based on the fact that women were precluded from service in combat (a restriction that was not challenged), id . at 76–77, and the “purpose of registration was to prepare for a draft of combat troops,” the Court held that the sex-based classification was 'sufficiently . . . related' to the Congressional purpose. Id . at 79. In doing so, the Court observed that Congress is permitted to prioritize military need over equity, but that nonetheless, similarly situated persons must be treated Page 29similarly and invidious discrimination will not be countenanced. Id . at 79-80.

These precedents demonstrate that the Court has employed something akin to a 'reasonable and evenhanded' test to review constitutional challenges to classifications made in military statutes and regulations. In doing so, the Court considered military purposes and interests as presumptively legitimate and due great deference, but the Court nonetheless looked to whether the classification was sufficiently related to that governmental purpose or interest. While gender-based classifications were not subjected to heightened scrutiny, the Court did so only after concluding that (1) such classifications were based on detailed study rather than “archaic and overbroad generalizations,” (2) where possible, similarly situated persons were treated similarly, and (3) the classification was not based on invidious discrimination. I will employ the same analysis here.¹⁰

I note also that both sides ask us to review the likelihood of success of the equal protection challenge mindful of United States v. Skrmetti , 605 U.S. 495 (2025), which was decided after the District Court's ruling. Skrmetti considered a Tennessee law that banned doctors from performing certain medical treatments and surgical procedures on minors as Page 30treatment for gender dysphoria. Id . at 505–06. The Court held that a state law prohibiting medical interventions to treat gender dysphoria for minors did not classify based on sex or transgender status. Instead, the Court held it was a classification based on a medical condition, as the law did not “prohibit conduct for one sex that it permits for the other.” Id . at 510–12. The Court explained that the law “d[id] not mask sex-based classifications,' because these medical treatments were prohibited for both male and female minors, and thus there was no sex discrimination. Id . at 513-17. The Court further held that the law was not a classification based on transgender status-a healthcare provider could not provide puberty blockers or hormones to any minor suffering from gender dysphoria, whether the child was transgender or not. Id. at 517–19.

As described below, the classifications in the Hegseth Policy have many distinctions from the classifications in Skrmetti . 11 More to the point, the Hegseth Policy does not classify whether persons are eligible to serve in the military in a reasonable and evenhanded manner. I focus specifically on three provisions of the Policy: (1) the disqualification from Page 31service based on any history of gender dysphoria, however remote, (2) the disqualification from eligibility for a waiver based on any past “attempt to transition,” and (3) the requirement that all persons disqualified under the policy go through the administrative separation process, rather than the Disability Evaluation System.¹²

In Skrmetti , the Court observed that 'if a law neither burdens a fundamental right nor targets a suspect class, we will uphold the legislative classification so long as it bears a rational relation to some legitimate end.' 605 U.S. at 510 (quoting Romer v. Evans , 517 U.S. 620, 631 (1996)). The Court upheld the Tennessee law because it was rationally related to the state’s proclamation of a “legitimate, substantial, and compelling interest in protecting minors from physical and Page 32emotional harm.” Id . at 516-17 (quoting Tenn. Code. Ann. § 68-33-101(m)).

Here, Appellants have partly justified the Hegseth Policy based on their interests in minimizing health care costs associated with the treatment of gender dysphoria, as well as lessening the risks to readiness and unit cohesion posed by persons with gender dysphoria or persons who have transitioned to another gender. These are legitimate government interests, and they help to explain certain parts of the Policy, like why surgical or hormonal gender transition while serving in the military is prohibited. Yet none of them explain why all persons with a history of gender dysphoria are disqualified, even if the person has been asymptomatic for many years and even if their only diagnosis was as a child. Persons who do not currently suffer from gender dysphoria or with such a remote history of the condition pose no costs or risks to the military. Further, the Hegseth Policy did not even attempt to explain why this broad classification was necessary or appropriate, let alone why this classification is so much broader than the Mattis Policy (which looked back three years).

Similarly, while the Hegseth Policy contains a waiver for both retention and accession where 'there is a compelling Government interest in accessing the applicant that directly supports warfighting capabilities,” the waiver is inapplicable to anyone who cannot demonstrate they have never “attempted to transition.” J.A. 53–55; Id . at 210. As discussed above, since persons who have undertaken medical and surgical transition are disqualified by the Hegseth Policy, “attempt to transition” must include attempts to socially transition. To construe the phrase otherwise would strip it of any operative force. The Hegseth Policy never explained why this broad 'attempt to transition” language appears in the Policy. Nor does the Policy explain what legitimate government interest this classification Page 33furthers. And Appellants never explained this provision in the District Court. In their briefing to this Court, Appellants refer to “attempt to transition” as a “classification based on medical interventions (i.e., cross-sex hormone therapy and sexreassignment surgery).” See Reply Br. at 8. But that formulation conflicts with the plain text of the Hegseth Policy, because it already disqualifies persons who have attempted medical or surgical transition, so including “attempt to transition” in the waiver policy has no effect unless it means some other type of transition-i.e., social transition. The government has not told us what legitimate interest it has in preventing a woman from joining (or remaining in) the military because she used he/him or they/them pronouns at some point during her lifetime, and I cannot think of one.

Further to this point, to the extent that the Hegseth Policy disqualified people who had a current diagnosis of gender dysphoria, it justified the classification on the basis that persons suffering from gender dysphoria do not meet the 'high standards for Service member readiness, lethality, cohesion, honesty, humility, uniformity, and integrity.” J.A. 50. Requiring servicemembers to meet those high standards is a legitimate government interest, and the Hegseth Policy presented some evidence demonstrating why gender dysphoria was incompatible with readiness, lethality, cohesion and uniformity. But, Appellants conceded below that they had presented no evidence to establish that persons with gender dysphoria are not honest, humble, and full of integrity, see id . at 1212–1213; Id . at 1240; Id . at 1280, and Appellants never contended, in the District Court or in this Court, that this part of the justification for excluding persons with a history of gender dysphoria has any sufficient relation to a legitimate government interest.

Page 34As discussed in Part II, the Hegseth Policy also justified its disqualifying classifications on the grounds that persons “expressing a false ‘gender identity’ divergent from an individual’s sex cannot satisfy the rigorous standards necessary for military service.' Id . at 207. Again, setting rigorous standards for military service is of course a legitimate government interest. Yet also again, the Policy and guidance documents offer no explanation or evidentiary support for the statement that transgender persons cannot meet those rigorous standards. The Hegseth Policy FAQs basically say that whether transgender servicemembers can meet rigorous military standards at all is beside the point:

1. Q. The Secretary of Defense has said that the focus needs to be on 'lethality, meritocracy, accountability, standards, and readiness.” Specifically focusing on 'meritocracy,' will consideration be given to high performing transgender service members?
2. A. While these individuals have volunteered to serve our country and will be treated with dignity and respect, expressing a false “gender identity' divergent from an individual's sex cannot satisfy the rigorous standards necessary for military service.

* * * Q. Did the Department consider reinstating the Mattis Policy regarding a Service member diagnosed with gender dysphoria and allow them to be grandfathered?

Page 35A. While these individuals have volunteered to serve our country and will be treated with dignity and respect, expressing a false “gender identity' divergent from an individual's sex cannot satisfy the rigorous standards necessary for military service.

Id . at 206–07.

The government abandoned all pretext of trying to craft sex-based classifications that were “sufficiently . . . related,” see Rostker , 453 U.S. at 79, to the legitimate government interests in 'lethality, meritocracy, accountability, standards, and readiness.” Those interests were dismissed as irrelevant — having the correct gender identity is all that matters. The contrast with the Mattis Policy alone shows that the Hegseth Policy is far from the “reasonable and evenhanded” approach upheld in Goldman . See 475 U.S. at 510.

When asked by the District Court for evidence that persons with a “false gender identity” cannot satisfy rigorous military standards, Appellants offered none. See Tr. of Mar. 12, 2025 hearing , Talbott , No. 1:25-cv-00240, Dkt.⁹⁰ at 182-83. Indeed, in the proceeding below, Appellants did not contest that all of the currently-serving Plaintiff-Appellees, who have served for a combined 130 years and collectively earned more than 80 commendations, have “served honorably” and 'satisfied the rigorous standards' demanded of them. See Talbott , 775 F. Supp. 3d at 292, 298. In our Court, Appellants have made no attempt to explain how the classificationexpelling persons who express a 'false gender identity'connects with the government interest in rigorous standards for military service. Rather, Appellants’ strategy seems to be one of willful ignorance, treating this statement as if it was not repeatedly made in the Executive Order and policy documents, Page 36and to instead deflect by contending that this case is only about gender dysphoria and medical treatments, the same as in Skrmetti . But that is clearly not so.

In Skrmetti , the Court concluded that the Tennessee law was 'simply a prohibition on the prescription of puberty blockers and hormones to treat gender dysphoria,” and that “[a] law prohibiting the administration of specific drugs for particular medical uses” does not warrant heightened scrutiny. 605 U.S. at 515–16. The Court strongly implied that a law that seeks to “force conformity with sex” or “ evinc[ing] sex-based stereotyping' would be considered a classification based on sex. Id .; see also Bostock v. Clayton Cnty., Ga. , 590 U.S. 644 (2020).

The Mattis Report began with the “fundamental premise” that “any eligible individual who can meet the high standards for military service without special accommodations should be permitted to serve. This is no less true for transgender persons than for any other eligible individual.” J.A. 73. The Hegseth Policy was mandated by an Executive Order which declared that ' expressing a false 'gender identity' divergent from an individual ' s sex cannot satisfy the rigorous standards necessary for military service.” 90 Fed. Reg. 8615, 8615 § 1 (emphasis added). Thus, while the Mattis Policy was not premised on a belief that transgender persons are unfit for the military, the Hegseth Policy is undergirded by the opposite belief. The Hegseth Policy clearly seeks to disqualify persons who do not act in conformity with their sex and who do not adhere to sex- Page 37based stereotypes. As discussed above, Secretary Hegseth reiterated those statements in his February 7 Memo, and in the FAQs accompanying the Hegseth Policy. Accordingly, unlike in Skrmetti, the challenged policy here is based, at least in part, on sex.

Further, as described below, Appellants have proffered no argument regarding how basing a policy on disfavoring persons with a “false gender identity” constitutes a legitimate government interest. Nor have Appellants demonstrated that there is any factual basis for the statement that persons with a “false gender identity” cannot meet the rigorous standards of military service. Thus, the Hegseth Policy justifies its classifications, at least in part, on factually unsupported (and undefended) archaic and overbroad generalizations about sex, in direct contradiction to the circumstances surrounding the sex-based classifications in Schlesinger v. Ballard . Indeed, the Hegseth Policy is a prototypical example of a sex-based classification created “reflexively and not for any considered reason,” in direct contrast to the Congressional action that the Court upheld in Rostker . See 453 U.S at 72.

In Skrmetti , the Court upheld Tennessee's ban on certain treatments for gender dysphoric minors because there were “plausible reasons” for the government to take the action, given the ongoing debate among medical experts regarding the risks and benefits of medical and surgical transition treatments for minors. See 605 U.S. at 522-23 (quoting FCC v. Beach Page 38Commc’ns, Inc. , 508 U.S. 307, 313 (1993)). In other words, the classification was consistent with ordinary uses of agebased and medical-based standards.

As explained above, the Mattis Policy treated gender dysphoria as an “analog . . . to any other mental condition listed in DoDI 6130.03.” J.A. 113. In other words, under the Mattis Policy, persons with gender dysphoria were treated as similarly situated with persons with other mental health disorders, such as depressive disorder. Id .

Here, the Hegseth Policy did not treat persons with gender dysphoria as similarly situated to persons with other mental health conditions. While the Mattis Policy disqualification for accession was based on a diagnosis of gender dysphoria within the past three years (the same as other depressive disorders), the Hegseth Policy disqualifies all persons with any history of gender dysphoria, regardless of how long ago the person has been stable—even if they were diagnosed only once long ago as a child. See id . at 50; Id. at 203; Id. at 1092.

Further, while the Mattis Policy evaluated the fitness of those servicemembers with gender dysphoria by placing them in the Disability Evaluation System for a case-by-case assessment (the same as any other medical condition), see id . at 85; Id . at 453, the Hegseth Policy mandates that all disqualified persons go through the administrative separation process, which is generally used for personnel accused of misconduct. Id . at 53–54; Id . at 1297; Id . at 1315; Id . at 1320. Indeed, former Assistant Secretary of the Air Force Alex Wagner testified without contradiction that administrative separation is generally used for misconduct such as drug abuse or repeated failure to meet standards, while the Disability Evaluation System is ordinarily used for evaluating impacts from medical conditions. Id . at 543–44. And former Deputy Page 39Assistant Secretary of the Air Force Martha Soper similarly stated in an uncontested declaration that the Hegseth Policy departs from the military’s standard practice: “In the typical circumstance, when a service member presents with a medical condition, they go through the Medical Evaluation Board (MEB) process at the wing level. From there, they would be referred to the Disability Evaluation System (DES), which allows the military to consider how a person's medical condition impacts their service and potential deployability.' Id . at 756.

Thus, servicemembers with every other medical condition receive an individualized review of their circumstances to determine if they can continue to serve—except if they have gender dysphoria.¹³ As the Mattis Report explained, even prior to the time when transgender persons could first serve openly in the military, “transgender persons were not usually processed for administrative separation on account of gender dysphoria or gender identity itself, but rather on account of medical comorbidities (e.g., depression or suicidal ideation) or misconduct due to cross dressing and related behavior.” Id . at 82. Plaintiff-Appellees pointed out this abnormal treatment of gender dysphoria in their brief, see Appellees' Br. at 25, and Appellants made no response in reply.

Appellants have therefore conceded that persons with a history of gender dysphoria are not treated the same as similarly situated individuals.

In Skrmetti , the challengers did not argue that the Tennessee law was based on animus. See 605 U.S. at 516. In contrast, Plaintiff-Appellees make such an argument here, and the District Court agreed with them, finding that the Hegseth Policy is “soaked in animus” and that it “stigmatizes transgender persons as inherently unfit.” Talbott , 775 F. Supp. 3d at 326. The District Court's animus finding is well supported by the two Executive Orders, the February 7 Memo, the Hegseth Policy, and the FAQs, all of which contain numerous statements demeaning transgender people as lacking honesty, integrity, and humility because of their gender identity.¹⁴

Appellants do not meaningfully contest the animus finding; they instead ignore the Administration’s repeated statements in their briefing. As deflection, Appellants repeatedly emphasize that the Hegseth Policy classifies on the basis of a medical condition, and that the Policy is premised on legitimate purposes including reducing medical and readiness risks and costs associated with that condition, to proclaim that the Hegseth Policy “belies any suggestion that it is motivated by animus.' See Appellants’ Br. at 51. But Appellants cannot just Page 41ignore the disparaging statements repeatedly made by the decisionmakers. See Vill. of Arlington Heights v. Metro. Hous. Dev. Corp. , 429 U.S. 252, 266, 268 (1977) ('Determining whether invidious discriminatory purpose was a motivating factor demands a sensitive inquiry into such circumstantial and direct evidence of intent as may be available' including “contemporary statements by members of the decisionmaking body”).

Furthermore, as described above, some aspects of the Hegseth Policy disqualify persons who have no current or recent diagnosis of gender dysphoria, so those classifications cannot be justified by concerns about costs, medical and readiness risks, and high standards. In addition, some classifications, like excluding persons from obtaining a waiver because of past experience with social transitioning, have not been explained at all. Nor did Secretary Hegseth explain the reason for processing all disqualifications through the administrative separation process rather than the Disability Evaluation System.

What has been clearly and repeatedly explained are the foundational premises of the Hegseth Policy: persons with a “false gender identity” are unfit for the military, and persons with a history of gender dysphoria are also unfit because they lack “honesty, humility, . . . and integrity.” See Feb.⁷ Memo; J.A. 50. As described above, those animus-filled reasons were expressly given to justify aspects of the Hegseth Policy. See J.A. 206–07. As a result, this is not a case where we are left to speculate why the government drafted such broad, undifferentiated classifications. Unless we are going to fall for the old Groucho Marx line—“who are you going to believe, me or your lying eyes?”—we have direct evidence in this case that animus motivated the classifications in the Hegseth Policy.

Page 42But even if this were the run-of-the-mill case where the classifications are merely unexplained and we were left with only circumstantial evidence of animus, the Hegseth Policy contains classifications that are ''divorced from any factual context from which we c[an] discern a relationship to legitimate state interests[,]' and '[their] sheer breadth [is] so discontinuous with the reasons offered for [them]' that the [classifications] see[m] ‘inexplicable by anything but animus.'' See Hawaii , 585 U.S. at 706 (quoting Romer v. Evans , 517 U.S. 620, 632, 635 (1996)). These unexplained and unreasoned departures from standard protocols are the kind of 'unusual deviation' that provide 'strong evidence of a law having the purpose and effect of disapproval of a class' of individuals. See United States v. Windsor , 570 U.S. 744, 746 (2013).

All of these things, when taken together, demonstrate that the government’s stated reason for issuing the Hegseth Policy as based solely upon gender dysphoria was pretextual, and that instead, the Hegseth Policy was premised, at least in part, on a non-legitimate state interest to harm the politically unpopular group of transgender persons. See Kelo v. City of New London, Conn. , 545 U.S. 469, 491 (2005) (Kennedy J., concurring) ('[A] court applying rational-basis review under the Equal Protection Clause must strike down a government classification that is clearly intended to injure a particular class of private parties, with only incidental or pretextual public justifications.”).

For these reasons, I reject the Dissent’s conclusion that the Hegseth Policy is not at all based upon animus or that we should endeavor to imagine legitimate justifications for the problematic classifications. See Dissenting Op. at 39-44. The government is not allowed to fabricate a legitimate justification for a classification post hoc , when it has already articulated an Page 43impermissible justification ex ante . Nor is our dissenting colleague.

The “any conceivable reason” doctrine came about because neither Congress nor the President is required to give a reason for every classification at the time they make it. See Heller v. Doe by Doe , 509 U.S. 312 (1993) (“[A] classification ‘must be upheld against equal protection challenge if there is any reasonably conceivable state of facts that could provide a rational basis for the classification.’”) (quoting Beach Commc’ns , 508 U.S. at 313). Rational basis doctrine requires us to presume that a classification is based on any plausibly legitimate justification we can imagine when no justification was expressly stated contemporaneously; that follows from the presumption that when the legislature (or the President) did not give a reason for the classification at the time that they made it, they are presumed to have acted for any plausible constitutional reason. Id .

But the doctrine has never allowed the Court to stick its head in the sand, like the proverbial ostrich, and ignore the justification that was expressly articulated at the time the decision was made. Indeed, the law, like common sense, requires the Court to evaluate the validity of the justification that was proclaimed at the time the classification was made. See Minnesota v. Clover Leaf Creamery Co. , 449 U.S. 456, 463, n.7 (1981) (“In equal protection analysis, this Court will assume that the objectives articulated by the legislature are actual purposes of the statute, unless an examination of the circumstances forces us to conclude that they ‘could not have been a goal of the legislation.'') (quoting Weinberger v. Wiesenfeld , 420 U.S. 636, 648, n.16 (1975)). As such, “th[e] Court’s review does require that a purpose may conceivably or ‘ may reasonably have been the purpose and policy' of the relevant governmental decisionmaker.” Nordlinger v. Hahn ,Page 44 505 U.S. 1, 15-16 (1992) (emphasis added) (quoting Allied Stores of Ohio, Inc. v. Bowers , 358 U.S. 522, 528–29 (1959)).

It is neither conceivable nor reasonable to conclude that the President acted for reasons other than those he expressly gave in the Executive Order, nor is it conceivable or reasonable to conclude that the Secretary acted for reasons other than the ones he expressly gave when promulgating the Hegseth Policy. See Allied Stores of Ohio , 358 U.S. at 530 (in explaining the holding of Wheeling Steel Corp. v. Glander , 337 U.S. 562 (1949), the Court observed that “[h]aving themselves specifically declared their purpose, the Ohio statutes left no room to conceive of any other purpose for their existence'); see also Nordlinger , 505 U.S. at 16, n.7 ('After the Court in Wheeling Steel determined that the statutory scheme's stated purpose was not legitimate, the other purposes did not need to be considered[.]'). If we are at the point where invidious reasons that were expressly given for a classification can be completely ignored and replaced with our imagined noninvidious reasons, then equal protection jurisprudence has truly become bankrupt.

* * *

In sum, applying Skrmetti and the military precedents described above, Plaintiff-Appellees demonstrated a likelihood of success on the merits that the Hegseth Policy violates the Equal Protection Clause.

To demonstrate irreparable harm, the movant must show that the harm is “certain and great,” “actual and not theoretical,” and that it is “beyond remediation” by compensatory or other relief at the conclusion of litigation.

Page 45Chaplaincy of Full Gospel Churches v. England , 454 F.3d 290, 297 (D.C. Cir. 2006). Accordingly, 'loss of earnings or damage to reputation' cannot afford a basis for finding irreparable harm. Sampson v. Murray , 415 U.S. 61, 89-90 (1974). That being said, the “loss of constitutional freedoms, for even minimal periods of time, unquestionably constitutes irreparable injury.' Mills v. District of Columbia , 571 F.3d 1304, 1312 (D.C. Cir. 2009) (citation modified).

Here, Plaintiff-Appellees contend that they have been irreparably harmed because the Hegseth Policy deprives them of “constitutional freedoms,” namely the loss of their right to equal protection of the law under the Due Process Clause. Appellees’ Br. at 60. Since Plaintiff-Appellees have demonstrated a likelihood of success on the merits—i.e., that the Hegseth Policy violates the equal protection guaranteed to them by the Fifth Amendment—then it follows that they have been irreparably harmed. As we said in Singh v. Berger , where we held that the Marine Corps’ refusal to allow plaintiffs to adhere to certain grooming standards in accordance with their religion was a violation of the Religious Freedom Restoration Act, the 'indignity of being unable to serve for reasons that . . . bear no relationship to their ability to perform” constitutes irreparable harm.⁵⁶ F.4th 88, 110 (D.C. Cir. 2022) (citation modified); see also 11A WRIGHT & MILLER, FED. PRAC. & PROC. § 2948.1 (3d ed. 2021) ('[W]hen there is an alleged deprivation of a constitutional right . . . many courts find that no further showing of irreparable injury is necessary.”).

I now turn to the balance of equities and public interest, which merge when the government is the opposing party. See Karem v. Trump , 960 F.3d 656, 668 (D.C. Cir. 2020). Even when a plaintiff can show irreparable harm, she is not entitled Page 46to a preliminary injunction as a matter of right. See Winter , 555 U.S. at 24. We “must balance the competing claims of injury and . . . consider the effect on each party of the granting or withholding of the requested relief.' Id . (quoting Amoco Prod. Co. v. Vill. of Gambell, Alaska , 480 U.S. 531, 542 (1987)). We review the District Court’s weighing of these factors for abuse of discretion. Winter , 555 U.S. at 32.

I begin by reviewing the District Court’s weighing of these factors in granting the preliminary injunction of the Hegseth Policy as applied to current servicemembers.

Appellants argue that any irreparable harm demonstrated by the Plaintiff-Appellees is outweighed by the harm to the government. They state that the Austin Policy, the predecessor to the Hegseth Policy, and which allowed transgender people and persons with a recent diagnosis of gender dysphoria to serve in the military, 'poses substantial risks to an effective national defense,” which they classify as “specific predictive judgments by senior military officials” that must be afforded great deference. Appellants’ Br. at 53–54 (internal citations omitted).

But, as the District Court found, there was little “specific predictive judgment[] by senior military officials” involved in the formulation and issuance of the Hegseth Policy. See Talbott , 775 F. Supp. 3d at 333 (finding that Appellants had provided “no testimony” from military officers as to the burden a preliminary injunction would place on the government). Indeed, the Mattis Policy—adopted in the first Trump Administration—concluded that the benefits of retaining servicemembers who were transgender or who had suffered from gender dysphoria outweighed any risks they posed to the Page 47military. See J.A. 114. Moreover, the government put forward no evidence that the Austin Policy posed a substantial risk to national security, and the Plaintiff-Appellees proffered affidavits and declarations all pointing to the contrary. A reversal to the status quo that has been held for four years cannot possibly be deemed a major threat to national security when no such national security threats were raised in years prior. Cf. Ziglar v. Abbasi , 582 U.S. 120, 143 (2017) ('[N]ational-security concerns must not become a talisman used to ward off inconvenient claims—a label used to cover a multitude of sins.”) (internal citations omitted).

Specifically, Appellants have not shown how a preliminary injunction preventing them from removing current servicemembers will harm national security. As shown above, Appellants have not contested that all of the currently serving Plaintiff-Appellees, who dedicated a combined 130 years to military service and collectively earned more than 80 commendations, have served honorably and met all military standards during their service. Appellants have thus forfeited any argument that, for purposes of the balance of equities analysis, retaining these servicemembers will harm national security.

In sum, the balance of equities, as well as the public interest, tips in favor of the Plaintiff-Appellees who are currently serving.

However, the calculus is different for those PlaintiffAppellees who seek to join the military. While PlaintiffAppellees who seek admission to the military have been deprived of constitutional freedoms, the harms and equities of their situations are not identical to those of persons already Page 48serving. Plaintiff-Appellees seeking admission to the military can still obtain full relief later, following a final adjudication on the merits. For those servicemembers facing expulsion, it is not clear how easily they can be reinstated and made whole. But even if they can be reinstated after being separated, it appears to us to be a much greater hardship to end a military career than to delay the start of one.

Moreover, those Plaintiff-Appellees who have been serving for years have demonstrated that their presence in the military is not harming national security, but we can only make a predictive judgment in that regard for the Plaintiff-Appellees seeking admission. The point about relative risk to national security is particularly relevant because the effect of the preliminary injunction is to revert to the Austin Policy, which allows accession of persons with a history of gender dysphoria so long as they have 18 months of stability. As described above, the Mattis Policy required 36 months of stability for accession because that is the same stability standard for depressive disorder. The shorter stability standard potentially poses increased risk of return of the illness, whereas, the risks are less unknown for persons already serving and being evaluated by their superiors and military doctors. This was a relevant consideration for the District Court in determining whether to stay its hand in imposing a preliminary injunction that would revert to the Austin Policy accession standards. See Weinberger v. Romero-Barcelo , 456 U.S. 305, 312 (1982) (“Where plaintiff and defendant present competing claims of injury, the traditional function of equity has been to arrive at a 'nice adjustment and reconciliation' between the competing claims[.]”) (quoting Hecht Co. v. Bowles , 321 U.S. 321, 329 (1944)).

Additionally, we must also consider that this is only a preliminary determination of the merits. It is still possible that Page 49following complete factual development and briefing, Appellants could prevail on the merits. If that were to happen, and in the meantime new servicemembers were admitted who would have been excluded under the Hegseth Policy, that would have caused Appellants to admit persons they were actually legally entitled to exclude—a distinct harm. This is against the public interest. See Yakus v. United States , 321 U.S. 414, 440-41 (1944) ('[W]here an injunction is asked which will adversely affect a public interest for whose impairment, even temporarily, an injunction bond cannot compensate, the court may in the public interest withhold relief until a final determination of the rights of the parties, though the postponement may be burdensome to the plaintiff.”).

Given these differences, the District Court was obligated to consider separately the balance of equities and public interest for those who were already serving in the military as compared to those who are seeking to enlist. Rather, it analyzed-and balanced-both sets of injuries, equities and interests as one. See Talbott , 775 F. Supp. 3d at 333. In doing so, the District Court abused its discretion. See Kickapoo Tribe of Indians of Kickapoo Rsrv. in Kansas v. Babbitt , 43 F.3d 1491, 1497 (D.C. Cir. 1995) (District Court abuses its discretion when it 'failed to consider a relevant factor'); NextEra Energy Glob. Holdings B.V. v. Kingdom of Spain , 112 F.4th 1088, 1108–11 (D.C. Cir. 2024) (holding that the District Court had abused its discretion in granting a preliminary injunction when it “overlooked” relevant factors in its analysis).

Accordingly, I conclude that the District Court abused its discretion in failing to consider relevant issues when balancing the preliminary injunction factors with respect to the accession standards of the Hegseth Policy. As a result, and for our separate reasons, Judge Rogers and I vote to affirm the Page 50preliminary injunction with respect to the retention standards of the Hegseth Policy, while Judge Walker and I vote to vacate the preliminary injunction as to those Plaintiff-Appellees seeking admission to the military. Further, we hold that the District Court’s universal injunction should be narrowed to the named Plaintiff-Appellees who are currently serving in the military. See Trump v. CASA, Inc. , 606 U.S. 831, 851-52 (2025).

For the reasons expressed above and in the separate opinions, we affirm in part, and vacate in part, the District Court’s preliminary injunction, and we remand for proceedings consistent with this opinion.

Page 51ROGERS, Senior Circuit Judge , concurring in part and dissenting in part: On January 27, 2025, President Trump issued Executive Order 14183 that stated:

It is the policy of the United States Government to establish high standards for troop readiness, lethality, cohesion, honesty, humility, uniformity, and integrity. This policy is inconsistent with the medical, surgical, and mental health constraints on individuals with gender dysphoria. This policy is also inconsistent with shifting pronoun usage or use of pronouns that inaccurately reflect an individual’s sex.

90 Fed. Reg. 8757. The Secretary of Defense was to implement this policy. Id. at 8757-58. As relevant, the Hegseth Policy disqualifies from military service all current servicemembers 'who have a current diagnosis or history of, or exhibit symptoms consistent with, gender dysphoria” and “who have a history of cross-sex hormone therapy or a history of sex reassignment or genital reconstruction surgery as treatment for gender dysphoria or in pursuit of a sex transition.” Dep’t of Defense, Additional Guidance on Prioritizing Military Excellence and Readiness, at 6 (Feb. 26, 2025) (“the Hegseth Policy”). In addition to this retention policy, the same group of people are barred from accession by appointment, enlistment, or induction into the military. Id. at 4. Various waivers are provided under the Hegseth Policy, but they are exceedingly restrictive. See Wilkins Op. at 22-23, 32.

On appeal by the government, Judge Wilkins concludes that the district court properly enjoined the Hegseth Policy on retention and abused its discretion on accession upon weighing the equities under the preliminary injunction factors. See Wilkins Op. at 4. For the following reasons, I would affirm the order enjoining the policy on retention and accession, as applied to the named plaintiffs. Therefore, I concur in part, and Page 52I respectfully dissent in part on accession, see Wilkins Op. Part III.C.2.

A threshold question of law is what level of scrutiny applies in evaluating the plaintiffs-appellees' likelihood of success on the merits of their equal protection claim. In challenging the district court's order enjoining the Hegseth Policy, the government contends that the Policy is subject only to rational basis review because of the high level of deference due to military judgments by the courts. See Appellants Br. at 22–27 (citing e.g. , Rostker v. Goldberg , 453 U.S.⁵⁷ (1981), and Goldman v. Weinberger , 475 U.S. 503 (1986)). Additionally, the government contends that rational basis review applies independently of military deference because the Hegseth Policy 'draws lines based on a medical condition (gender dysphoria)” rather than “identity or status.” Id. at 26. It compares the Policy to a state law restricting certain medical interventions for minors that the Supreme Court upheld under rational basis review in United States v. Skrmetti , 605 U.S. 495 (2025). Further, the government contends that under rational basis review the Hegseth Policy must be upheld even if it was based in part on animus 'so long as it can reasonably be understood to result from a justification independent of unconstitutional grounds.' Appellants Br. at 52 (quoting Trump v. Hawaii , 585 U.S. 667, 705 (2018)).

These contentions warrant examination. It is undisputed that 'courts must give great deference to the professional judgment of military authorities concerning the relative importance of a particular military interest.” Goldman , 475 U.S. at 503; see Rostker , 453 U.S. at 64-65. But in neither Goldman nor Rostker did the Supreme Court hold that rational basis review automatically applies where the military is Page 53involved. To the contrary, in Rostker , the Court rejected the government's argument to that effect. 453 U.S. at 69-70. Although the Court deferred to Congress’ determination that only men needed to register for the draft — observing that the issue had been 'extensively considered' in 'hearings, floor debate, and in committee” including “extensive testimony and evidence,” id. at 72 - the Court explained that it did 'not think that the substantive guarantee of due process or certainty in the law will be advanced by any further 'refinement' in the applicable tests as suggested by the Government,” id. at 69-70. While focusing on whether Congress violated equal protection, the Court indicated that it would apply intermediate scrutiny as established in Craig v. Boren, 429 U.S. 190 (1976). See id . at 70.

Nor does the Supreme Court’s recent decision in Skrmetti necessarily require the Hegseth Policy be assessed under rational basis review. There, the Court upheld state restrictions on the use of certain medical treatments for gender dysphoria in minors. 605 U.S. at 511. It concluded that the restrictions were classified based on age and medical conditions rather than on a person's sex or transgender identity. Id. at 511-18. Notably, in so doing, the Court contrasted regulation of “a class of treatments or conditions' with regulation of 'a class of persons identified on the basis of a specified characteristic.' Id. at 519 n.3. As a result, a policy targeting women would not be sex-neutral merely because it was recast as one targeting “all individuals who have ever, or may someday, menstruate.” Id. (citation and internal quotation marks omitted). Arguably, the same distinction applies here: the Hegseth Policy does not just regulate a particular medical procedure but instead excludes “a class of persons identified on the basis of a specified characteristic,” id. , namely, transgender individuals, who may have a history or symptoms consistent with gender dysphoria.

Page 54And even if rational basis review is the proper standard, it is unclear whether the Supreme Court's holding in Trump v. Hawaii is properly understood as establishing the generalized principle about animus that the government espouses. See Appellants Br. at 52. There, the Court addressed the admission and exclusion of foreign nationals, an area in which the political branches exercise a “fundamental sovereign attribute” and the judiciary’s role “is highly constrained.” 585 U.S. at 702, 704 (citations omitted). The Court held that it would review a presidential proclamation barring entry of foreign nationals from six predominantly Muslim countries, including “extrinsic evidence” of animus, and that it would “uphold the policy so long as it can reasonably be understood to result from a justification independent of unconstitutional grounds.' Id. at 705. That conclusion does not require lower courts to presume that evidence of animus is immaterial under rational basis review beyond the immigration context in Trump v. Hawaii .

At this preliminary, pretrial stage of the proceedings, the court need not resolve complex questions involving the intersection of military deference under Skrmetti , and Trump v. Hawaii , nor delineate a universal test that would apply to all military contexts, see Wilkins’ Op. at 29 & n.10. Even under the highly deferential form of rational basis review espoused by the government, see Appellants Br. at 52, a policy is unconstitutional if it is 'inexplicable by anything other than animus.” Trump v. Hawaii , 585 U.S. at 706 (quoting Romer v. Evans, 517 U. S. 620, 632, 635 (1996)). Plaintiffs-appellees are likely to make that showing. Judge Wilkins identifies three features of the Hegseth Policy that are unexplained and irrational. See Wilkins Op. at 24-32, 37-39. At least two are illustrative of that problem, and our dissenting colleague does not suggest that a policy explainable only by animus is constitutional, see Walker Op. at 40–43.

Page 55First, consider the way in which the Hegseth Policy on retention differs from the standard procedures of the Department of Defense by calling for discharge before individualized assessments are made. See Wilkins Op. at 3839 (citing declaration of former Deputy Assistant Secretary of the Air Force). Government counsel appeared to concede during oral argument that gender dysphoria is the only identifiable medical condition that automatically triggers administrative separation without an individualized assessment of whether the condition is so burdensome on the service member as to preclude continued service. Oral Arg. Tr. at 133–34 (Jan. 22, 2026). In any event, the government has provided no explanation for why this medical condition must be treated differently than all others. Nor is there a basis in the record that enables the court to conclude that gender dysphoria poses a unique danger dwarfing other medical conditions. The district court found, based on record evidence, that the enlisted plaintiffs have collectively 'provided over 130 years of military service” and “earned more than 80 commendations” from the military. Talbott v. United States , 775 F. Supp. 3d 283, 291 (D.D.C. 2025).

The government’s concerns about the negative impacts on cohesion and privacy do not advance its position. Appellants Br. at 35-38. The government maintains that 'permitting individuals to serve inconsistently with applicable sex-specific standards ‘would invade the expectations of privacy’ of other servicemembers” who share “berthing, bathroom, and shower facilities.” Id. at 36 (quoting Department of Defense Report and Recommendations on Military Service by Transgender Persons, at 37 (Feb. 2018) (“Mattis Policy”)). That position is disconnected from the government’s repeated refrain that the Hegseth Policy, like the Mattis Policy, is not a “blanket ban” on transgender personnel in the military and is instead an exclusion of individuals who have the medical condition of Page 56gender dysphoria, see, e.g., Appellants Br. at 40, for 'not all trans-identifying individuals have gender dysphoria” and the Hegseth Policy “permits individuals with a history of gender dysphoria to be considered for waivers,' id. at 41. So, according to the government, some transgender individuals may be able to serve capably in the military, while no reason is offered or evident why current service members are barred from the standard individualized evaluation to determine if they can continue to serve.

Second, the Hegseth Policy waiver provision, as implemented, requires that both current servicemembers and applicants show that they “never attempted to transition to any sex” other than their own sex at birth. See Hegseth Policy at 4-5; Dep't of Defense, Memorandum for Senior Pentagon Leadership at 1 (March 4, 2025). The plain text can be understood to exclude anyone who has ever attempted to socially transition and the government, which declined to clarify this provision upon the district court’s request, belatedly and unpersuasively suggests that the reference is limited to medical interventions. See Wilkins Op. at 22-23 & n.8, 32-33 (citing Reply Br. at 8). It is unclear what interest the military would have in categorically excluding such individuals who otherwise meet the criteria for a waiver under the Hegseth Policy. Like the retention policy, this aspect of the accession policy appears designed to target a group of disfavored people, not to address a particular medical condition.

The Supreme Court’s recent stay order in United States v. Shilling , 145 S. Ct. 2695 (2025), does not require a different conclusion, but signals the need to narrow the injunction to apply only to the named plaintiffs-appellees. There, the Court stayed, without explanation, a nationwide preliminary injunction of the Hegseth Policy in the Western District of Washington. Shilling v. United States , 773 F.Supp.3d 1069,Page 57 1098–99 (W.D. Wash. 2025). The stay was issued shortly before Trump v. CASA, Inc. , 606 U.S. 831, 837 (2025), holding that such universal relief issued by the district court 'likely exceed[s] the equitable authority that Congress has granted to federal courts.' In Shilling , the district court determined that it was unnecessary to consider whether the Hegseth Policy had been fueled by animus. Compare Shilling , 773 F.Supp.3d at 1098–99, with Talbott , 775 F. Supp. 3d at 326–32.

The cases cited in the dissent do not require a different conclusion either. While the Supreme Court has emphasized the need for deference to military judgments, it has also repeatedly recognized that constitutional rights still apply in this realm. See, e.g. , Rostker , 453 U.S. at 67–68 (stating that Congress is not “free to disregard the Constitution when it acts in the area of military affairs,” because “[i]n that area, as any other, Congress remains subject to the limitations of the Due Process Clause'); Goldman , 475 U.S. at 507 (stating that unique aspects of military life 'do not, of course, render entirely nugatory in the military context the guarantees of the First Amendment”); Chappell v. Wallace , 462 U.S. 296, 304 (1983) (“This Court has never held, nor do we now hold, that military personnel are barred from all redress in civilian courts for constitutional wrongs suffered in the course of military service.'); cf. Weiss v. United States , 510 U.S. 163, 176 (1994) (“Congress, of course, is subject to the requirements of the Due Process Clause when legislating in the area of military affairs, and that Clause provides some measure of protection to defendants in military proceedings.”). See also Wilkins Op. at 25 & n.9. In looking past this principle, our dissenting colleague embraces a logic of military deference that has no clear end point. See, e.g., Walker Op. at 9. Notwithstanding the postulate that '[o]nly the Executive and Congress are responsible for system-wide military judgments about the composition of the armed forces,” id. at 45, the courts do have Page 58a role in evaluating military determinations that intrude upon individual constitutional rights, as the Supreme Court has repeatedly made clear in drawing the line between exercises of military judgment and constitutional rights held by members of the military.

Therefore, because features of the Hegseth Policy are inexplicable by anything other than animus, and the “loss of constitutional freedoms' is an irreparable injury, Mills v. District of Columbia , 571 F.3d 1304, 1312 (D.C. Cir. 2009), plaintiffs-appellees have shown irreparable harm by demonstrating a likelihood of success on their equal protection claim.

The remaining question is whether the district court erred by enjoining the Hegseth Policy on accession. In the district court, neither of the parties raised the issue of severability on the merits. On appeal, the government has not challenged the record evidence on which the district court relied in issuing the injunction on retention and accession. Evidence before the district court included declarations showing professional and economic harms by plaintiffs who were seeking to join the military and claimed that the Policy violated their constitutional rights. See, e.g ., Declarations of Koda Nature and Cael Neary.

So, the question is whether the district court abused its discretion in weighing the preliminary injunction factors. See Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 32 (2008); Clevinger v. Advoc. Holdings, Inc. , 134 F.4th 1230, 1233 (D.C. Cir. 2025). Judge Wilkins concludes that it has. In his opinion, the district court 'was obligated to consider separately the balance of equities and public interest for those who were Page 59already serving in the military as compared to those who are seeking to enlist” and overlooked a relevant factor by considering 'both sets of injuries, equities and interests as one.” Wilkins Op. at 49. He posits that the “relative risk to national security” may be more significant for the accession policy than the retention policy because the “PlaintiffAppellees who have been serving for years have demonstrated that their presence in the military is not harming national security, but we can only make a predictive judgment in that regard for [those] seeking admission.' Id. at 48.

Under the abuse of discretion standard, the question before an appellate court is not whether it 'would as an original matter' reach a particular result; instead, 'it is whether the District Court abused its discretion in doing so.” Ins. Corp. of Ireland v. Compagnie des Bauxites de Guinee , 456 U.S. 694, 707 (1982) (quoting National Hockey League v. Metropolitan Hockey Club, Inc. , 427 U.S. 639, 642 (1976) (citing C. WRIGHT & A. MILLER, FEDERAL PRACTICE & PROCEDURE: CIVIL §2284, p. 765 (1970), and cases of the Eighth and Second Circuits Courts of Appeal)). The abuse of discretion standard 'contemplates reasoned decision making on the basis of relevant and appropriate considerations to the task at hand.' Kickapoo Tribe of Indians of Kickapoo Rsrv. in Kansas v. Babbitt , 43 F.3d 1491, 1497 (D.C. Cir. 1995). In reviewing for abuse of discretion, the appellate court considers whether the district court “failed to consider a relevant factor,” whether it “relied on an improper factor,” and whether its “reasons given reasonably support the conclusion.” Id. (quoting Johnson v. United States , 398 A.2d 354, 365 (D.C. 1979), and citing Maurice Rosenberg, Judicial Discretion of the Trial Court, Viewed from Above , 22 SYRACUSE L. REV. 635 (1971)). A district court does not abuse its discretion simply because an alternative conclusion may also be permissible; the appellate court's inquiry is instead whether the determination upon Page 60review “was a permissible alternative available to the district court.' United States v. Kanu , 695 F.3d 74, 81 (D.C. Cir. 2012); see United States v. Volvo Powertrain Corp ., 758 F.3d 330, 345 (D.C. Cir. 2014) (citations omitted). This characteristic distinguishes review for abuse of discretion from review of factual findings for clear error and de novo review for questions of law. Rosenberg, 22 SYRACUSE L. REV. at 646 (analyzing federal and state cases).

The district court, upon concluding that the plaintiffs were likely to succeed on their equal protection claim, balanced the equities and the public interest, and determined that the Hegseth Policy should be enjoined. Talbott , 775 F. Supp 3d at 332–33. The court concluded that the plaintiffs and the public interest would be harmed by the enforcement of an unconstitutional policy, while the government had failed to demonstrate that leaving the prior policy in place would impede the military. Id . The prior accession policy had been in effect from 2021 to 2025, yet neither the Hegseth Policy nor the government provided 'any studies or declarations that explain why maintaining the status quo pending litigation would unfairly burden the military.” Id. at 333. Instead, the plaintiffs’ evidence indicated the opposite: summarizing “declarations from the military leaders responsible for integrating transgender persons into open military service,” the district court found that this evidence showed that “recruiting, unit cohesion, and military readiness have improved since 2021.' Id. at 292. That evidence, the district court concluded, indicated that accession under the Hegseth Policy would harm military interests. See id . at 333. An uncontested declaration of the Secretary of the Navy who served from 2021 to 2025 stated that “there is no evidence-based justification for excluding from service someone who meets all applicable standards merely because they are transgender,' and 'such exclusion would harm military readiness by depriving our force Page 61of qualified personnel who have proven their ability to serve.” Id. at 308 (citation and alterations omitted).

Given the uncontested record evidence and the nature of this court’s review for abuse of discretion, the district court did not abuse its discretion in enjoining the Hegseth Policy on accession. This is so even if based on the evidence and argument before it, the district court could have concluded that the retention policy imposes 'a much greater hardship' on those currently in the military than those who seek to join it, as Judge Wilkins concludes, Wilkins Op. at 48, and only enjoined the Hegseth Policy on retention due to national security concerns. But the existence of a permissible alternative is not the same as showing that the district court abused its discretion by choosing another permissible alternative. See, e.g., Volvo Powertrain Corp ., 758 F.3d at 345 ; Kanu , 695 F.3d at 81.

The district court recognized that deference was due to the military’s judgment, not “[b]lind [d]eference.” Talbott , 775 F. Supp. 3d at 291, 312–13 (citing, e.g. , Rostker , 453 U.S. at 72, 83). That court exhaustively reviewed the record evidence and the parties’ arguments. See id. at 300–33. That court reached conclusions based on the uncontested declarations by military officials about the very issue now of concern to Judge Wilkins: potential harm to the military if plaintiffs-appellees were permitted to join the military service. Id. at 333. No relevant factor was overlooked. Suggesting that the district court erred in balancing the equities by not framing its analysis of harm in terms of the “greater hardship” in “end[ing] a military career” as opposed 'to delay[ing] the start of one' and by not speculating that 'the shorter stability standard [of the prior policy] potentially poses increased risk of return of the illness,” Wilkins Op. at 48, “fail[s] to give the trial court the deference that is the hallmark of abuse-of-discretion review.” Gen. Elec. Co. v. Joiner , 522 U.S. 136, 143 (1997). That suggestion is at Page 62odds with the district court’s factual findings of harm to the military under the Hegseth Policy on accession by contrast with military improvements under the prior policy, based on uncontested declarations of military officials. See, e.g. , Wilkins Op. at 34–36, 41–42. No record evidence confirms the medical speculation offered by Judge Wilkins. So far, then, the government has not presented a basis for this court, upon review for abuse of discretion, to vacate the injunction on accession, and, therefore, I respectfully dissent and do not join Judge Wilkins’ opinion in Part III.C.2.

Accordingly, I would affirm the order of the district court enjoining the Hegseth Policy on retention and accession, as modified to apply to the named plaintiffs, see Trump v. CASA, Inc. , 606 U.S. at 837.

Page 63WALKER, Circuit Judge , dissenting:

Judge Stephen Williams, “one of the most distinguished jurists ever to serve on the D.C. Circuit Court of Appeals,” 1 wrote seven years ago that in the military 'there is no constitutional right for, say, biological males who identify as female to live, sleep, shower, and train with biological females. Whether allowing such flexibility in military service is a good idea or not is of no concern to the courts; that is a question for the people acting through their elected representatives.” 2

Now, as then, transgender plaintiffs object to a policy that excludes them from the military. Now, as then, the district court has enjoined the policy’s operation. And now, as then, the Government has appealed.

'Once we abstract away from the politically charged subject-matter — as we must — this is a straightforward legal case.” 3 We have neither the expertise nor the authority to decide whether the military can exclude the plaintiffs from its ranks. The Constitution assigns that authority to Congress and the Commander in Chief.

Because their claim flouts that division of power, the plaintiffs cannot identify a single binding precedent holding that the Constitution dictates a composition of forces contrary to the preferences of our nation’s political branches.⁴

Not one.

Page 64That black hole of futile claims has swallowed up assertions of free speech.⁵ And free exercise.⁶ And due process.⁷ And the right to counsel.⁸ And, as here, equal protection.⁹

In each of those cases, the Supreme Court said that the military can deprive its members of rights that the Constitution may well guarantee to civilians.¹⁰ Like today's majority, I cherish those rights, and so I understand the impulse behind the majority's unprecedented intervention into military affairs. But because the plaintiffs are service members not civilians, and because we are judges not generals, I respectfully dissent.

For a decade, every presidential administration has changed the U.S. military's personnel policies regarding gender dysphoria and transgender service members. Each new policy was announced as an explicit rejection of the prior approach.¹¹

6 Goldman v. WeiPage 65nberger , 475 U.S. 503, 504 (1986); see also Austin Before 2016, military regulations “effectively banned all transgender persons from either joining or remaining in the military.” 12 The military's medical standards excluded all whoPage 66 'did not identify with the gender assigned to them at birth' regardless of whether they had been diagnosed with gender dysphoria, regardless of whether they had attempted to “transition[] to their preferred gender,” and regardless of their 'willing[ness] to serve pursuant to the military standards applicable to the sex assigned to them at birth.” 13

That changed in 2016 when the military adopted a more permissive set of policies.¹⁴ The new policy for people entering the military allowed people with a history of gender dysphoria to join and serve in their biological sex if they had demonstrated 18 months of stability in their biological sex.¹⁵ The new retention policy allowed them to continue serving if they met deployability standards.¹⁶ Those who had not completed a gender transition were required to serve in their biological sex; those who had completed a transition could serve in their preferred gender.¹⁷

Then in 2018, the military returned to a more restrictive approach.¹⁸ It adopted a policy similar to the pre-2016 policy, Page 67except that it grandfathered in service members who had relied on the more permissive 2016 policy. So people could no longer join the military if they had begun or completed a gender transition.¹⁹ But service members who had already begun or completed a transition before the policy went into effect received a “reliance exception” and could continue serving.²⁰

Shortly after President Biden took office in 2021, he directed the military to return to a permissive policy regime “to ensure that all transgender individuals who wish to serve in the United States military and can meet the appropriate standards shall be able to do so openly and free from discrimination.” 21 The military formally adopted policies implementing that directive shortly thereafter.²² These were materially similar to the framework adopted in 2016, which allowed transgender service members to serve openly while being “subject to the same standards” as other service members. 23

Then, days after President Trump took office in 2025, he directed the military to adopt its most restrictive set of policies since 2016. 24 And once again, the military began to act on that directive shortly thereafter.

Page 68On February 26, 2025, the military adopted a new policy disqualifying from service 'individuals who have a current diagnosis or history of, or exhibit symptoms consistent with, gender dysphoria,” on the ground that “the medical, surgical, and mental health constraints” on these individuals do not meet “the high mental and physical standards necessary for military service.” 25 Existing service members whom the policy disqualified would “be processed for separation from military

The majority agrees that the policy at issue was announced on February 26, 2025, in the memorandum I cite here. See Majority Op. at 5. But the majority frequently refers to an earlier memorandum, issued on February 7, 2025, which ordered a pause on 'all new accessions for individuals with a history of gender dysphoria” and all 'medical procedures associated with affirming or facilitating a gender transition for Service members.” SecDef Memo, Talbott v. United States , No. 1:25-cv-240 (D.D.C. Mar. 4, 2025), Dkt. No. 331. That memo stated that the military 'must ensure it is building 'One Force' without subgroups defined by anything other than ability or mission adherence” and expressed concern about “[e]fforts to split our troops along lines of identity” that “weaken our Force and make us vulnerable.” It then quoted a portion of Executive Order 14183: ''Expressing a false 'gender identity' divergent from an individual’s sex cannot satisfy the rigorous standards necessary for Military Service.’” Id.

Aside from the pause on medical procedures, the memo did not include guidance regarding current service members. Instead, it stated that “[i]ndividuals with gender dysphoria have volunteered to serve our country and will be treated with dignity and respect” and authorized the Under Secretary of Defense for Personnel and Readiness to issue guidance for Service members “to implement this direction.” Id.

Page 69service' with an 'honorable' discharge unless 'the Service member's record otherwise warrant[ed].'²⁶ A disqualified member may seek a waiver “on a case-by-case basis” upon a showing of “a compelling Government interest . . . that directly supports warfighting capabilities,' along with 36 months of mental stability, no prior attempts to transition to another sex, and willingness to adhere to the military’s sex-specific standards.²⁷

An action memo issued the same day elaborated on the policy's rationale. Military service, the memo emphasized, requires personnel who are 'mentally and physically fit for duty” and able to “deploy, fight, and win, including in austere conditions and without the benefit of routine medical treatment or special provisions.” 28 The memo cited a 2021 Department of Defense study finding that “nearly 40% of Service members with gender dysphoria . . . were non-deployable over a 24 month period'²⁹ and a 2025 medical-literature review reporting significantly elevated rates of suicidal ideation, suicide attempts, and psychiatric diagnoses among transgender people.³⁰ Continued service by people with gender dysphoria, the memo concluded, was incompatible with the military's “rigorous standards and national security imperative to deliver a ready, deployable force.” 31

The plaintiffs filed this suit challenging the President's executive order the day after it was issued and moved for Page 70injunctive relief a week later.³² Soon after the military adopted the implementing policy, the plaintiffs amended their complaint and renewed their motion for injunctive relief, asking the court to enjoin enforcement of the policy. 33

The district court decided that this latest change in the decade-long back-and-forth of military policy is likely unconstitutional.³⁴ The court preliminarily enjoined the military from enforcing its new policies.³⁵ To remove any doubt, the district court specified that its order effectively requires the military “to maintain the status quo ” established by the previous administration.³⁶

Today's majority now concludes that a narrow set of disqualifications in the new military policy are “unexplained and unreasoned' and that some of the reasons the military provided for the policy reveal animus.³⁷ The majority finds those problems with some parts of the policy sufficient to override the longstanding tradition of deference to military decisionmaking and to affirm a preliminary injunction of the

33 Motion for Preliminary Injunction (Renewed),

States Talbott v. United , No. 1:25-cv-240 (D.D.C. Mar. 7, 2025), Dkt. No. 72.

Page 71entire policy as it relates to retention of the plaintiff service members.³⁸

The Constitution entrusts the political branches, not courts, with judgments about military readiness and effectiveness. Indeed, “[i]t would be difficult to think of a clearer example of the type of governmental action that was intended by the Constitution to be left to the political branches,' which are “directly responsible — as the Judicial Branch is not — to the electoral process.” 39

Constitutional authority aside, prudence points in the same direction: “it is difficult to conceive of an area of governmental activity in which the courts have less competence.” 40

As a result, 'courts traditionally have been reluctant to intrude upon the authority of the Executive in military and national security affairs” — that is, “unless Congress specifically has provided otherwise.” 41 The judiciary's lack of institutional competence explains why judges should defer to the political branches in matters bearing on military-wide force Page 72composition, training, discipline, and readiness; constitutional structure explains why we must .

The Constitution allocates military authority between Congress and the President. Article I empowers Congress to “raise and support Armies,” to “provide and maintain a Navy,” and to “make Rules for the Government and Regulation of the land and naval Forces.” 42 Article II vests the President with “[t]he executive Power” and confirms that he is “Commander in Chief” of the armed forces.⁴³

These provisions establish the core structural principle of military governance: The two political branches govern the military.

In Federalist No. 23, Alexander Hamilton defended this allocation as 'essential to the common defense.'⁴⁴ Because '[t]he circumstances that endanger the safety of nations are infinite,” the military powers vested in the political branches 'ought to exist without limitation.'⁴⁵ Whether to create a federal defense power at all was, he conceded, 'open to discussion.” 46 But once that question was answered, 'there can be no limitation of that authority . . . in any matter essential to Page 73the formation , direction , or support of the NATIONAL FORCES.'⁴⁷

For more than seventy years, the Supreme Court and the D.C. Circuit have looked to that constitutional structure and come to the same conclusion: Courts must not second-guess military judgments about how the armed forces are structured, trained, and operated. Instead, they “must give great deference' to the judgment of the political branches, which have been entrusted by the Constitution “with the care of the common defense.'⁴⁸

That principle does not just appear in a single test or a single line of cases. It appears everywhere. It's the reason courts resist interlocutory intervention in military commissions and courts-martial. 49 It's the reason courts do not issue preliminary injunctions that would interfere with military training or operations.⁵⁰ It's the reason courts sometimes Page 74refuse to review military decisions at all.⁵¹ And on the occasions courts do reach the merits, it’s the reason they apply an extraordinarily deferential standard of review for constitutional claims. 52

The courts’ foundational cases about the military’s structure and operation establish three related points, each demanding from judges one word — deference:

1. 1. Judgments about the military’s structure and operation belong to the political branches, not to the courts.
2. 2. The military is a distinct constitutional system, separate from civilian society. Its rules are necessarily different than those of civilian society. So courts do not apply the same standards to military rules as they do to the rules of civilian society.

The doctrine of military deference begins with a simple proposition: Judgments about the military belong to the political branches, not to the courts.

In Orloff v. Willoughby , the Supreme Court made the point directly. A doctor inducted into the Army asked the Court to intervene because the military had allegedly discriminated against him in his duty assignment based on his refusal to say if he had been a member of the Communist Party.⁵³ The Court did not treat that allegation as a basis for judicial intervention. It acknowledged that 'from top to bottom of the Army the complaint is often made, and sometimes with justification, that there is discrimination, favoritism or other objectionable handling of men.'⁵⁴ Nevertheless, the Court refused to intervene because “judges are not given the task of running the Army.'⁵⁵

To police the line between the rights of civilians and the rights of soldiers, Orloff said courts may “determine whether one has been lawfully inducted and is therefore within the jurisdiction of the Army and subject to its orders.” But once that inquiry is over, the Court suggested it has little further role: 'Nothing appears to convince us that [Orloff] is held in the Page 76Army unlawfully, and, that being the case, we cannot go into the discriminatory character of his orders.” 56

In other words, Orloff stopped at the Army’s gates. Why? Because crossing that threshold to review Orloff’s assignment would make the judiciary 'a disruptive force as to affairs peculiarly within the jurisdiction of the military authorities.” 57 And that disruption would imperil ordered liberty itself: “Orderly government requires that the judiciary be as scrupulous not to interfere with legitimate Army matters as the Army must be scrupulous not to intervene in judicial matters.” 58

Two decades later, the petitioners in Gilligan v. Morgan asked the Supreme Court to review the Ohio National Guard’s training standards in the wake of the Kent State shootings. The Court once again declined jurisdiction. It explained:

The complex, subtle, and professional decisions as to the composition, training, equipping, and control of a military force are essentially professional military judgments, subject always to civilian control of the Legislative and Executive Branches. The ultimate responsibility for these decisions is appropriately vested in branches of the government which are periodically subject to electoral accountability.⁵⁹

Page 77Those 55 words answer nearly every question we should ask today: 1) What are 'professional military judgments'? 2) Who controls them? 3) When ? 4) Why ?

What are “professional military judgments”? Whatever else they may include, they include system-wide 'decisions as to the composition, training, equipping, and control of a military force.” Such decisions are “ essentially ' - that is, by their very nature - 'professional military judgments.” 60

Who controls those decisions? '[T]he Legislative and Executive Branches.”

When are those decisions subject to the political branches’ control? Emphatically: ' always .”

Why those branches and not the judiciary? At least in part because the political branches “are periodically subject to electoral accountability.”

Taken together, Orloff and Gilligan establish the foundation for modern military deference. Courts do not step into the middle of military decisionmaking. They do not revise Page 78military assignments — even if the assignments are allegedly unfair or discriminatory. They do not supervise military training — even if that training has been questionable. They generally do not reweigh judgments about how the military should be organized and prepared. That is because courts are in a poor position to second-guess those decisions, even when they may look unfair or unreasoned from the bench. They lack the constitutional power to intrude into a system left to the political branches, and they lack the institutional competence to presume they should know better.

The basic rule — courts do not run the military — carries through the rest of the case law.

The cases that followed Orloff and Gilligan explained something deeper about the doctrine: Courts defer to military decisions because the military is a separate society altogether. It has different needs than civilian society. As a result, it has different rules than civilian society. The rights that prevail in civilian courts for civilian society do not operate in the same way for the military.⁶¹

The Supreme Court made that point plainly in Parker v. Levy when it rejected a First Amendment challenge to rules in the Uniform Code of Military Justice, explaining that 'the military is, by necessity, a specialized society separate from civilian society” that has “developed laws and traditions of its Page 79own during its long history.” 62 Those differences 'result from' the military's special purpose: 'to fight or be ready to fight wars should the occasion arise.'⁶³ And they result in 'a different application of [constitutional] protections” to military regulations - an application that 'may render permissible within the military that which would be constitutionally impermissible outside it.” 64

A year after Parker , the Supreme Court again invoked the military’s status as a specialized society, this time to caution lower courts against using their equitable powers to shortcircuit military processes. A district court had permanently enjoined the military from proceeding with a court-martial; the Court in Schlesinger v. Councilman reversed.⁶⁵

Councilman concluded that the military's status as a specialized society “counsels strongly against the exercise of equity power even where . . . intervention might be appropriate.” 66 That restraint reflects the 'unique military exigencies” that gave rise to the military’s distinct body of law and that are “as powerful now as in the past.” 67 The upshot: When courts must decide whether to use equitable authority to intervene in military decisions, the scales tip strongly against Page 80intervention from the start — even where “intervention might [otherwise] be appropriate.” 68

The next year, in Middendorf v. Henry , Justice Powell cited the military’s status as a specialized society once more, this time to explain why the right to counsel should not apply to summary courts-martial as it would in ordinary civilian criminal proceedings. Though “one’s constitutional rights are not surrendered upon entering the Armed Services,” the rights must be applied “in light of the ‘unique military exigencies’ that necessarily govern many aspects of military service.” 69

The rule established in Orloff and Gilligan — that courts do not run the military - thus grew into a full distinction between military society and civilian society, military rules and civilian rules. When courts review challenges to rules affecting the military, they must not begin as they would if they were considering cases concerning civilian society and civilian rules. Instead, they must consider the “unique military exigencies” that may require a different approach from the start.

Though service members leave the civilian ranks to become part of the military, they do not leave behind their citizenship and its rights. So it remains for us to consider how courts should handle claims that military rules violate individuals’ constitutional rights. The cases I discussed above have already suggested an answer. The cases I discuss below Page 81will make the point plain: Even for constitutional claims, courts must give “great deference” to the political branches.

Return to Parker for what it teaches about constitutional challenges. Military members 'are not excluded from the protection[s] granted by the” Constitution 70 - they are, after all, still citizens under the Constitution’s protection. But “the different character of the military community and of the military mission requires a different application of those protections.” 71 That's because '[t]he armed forces depend on a command structure that at times must commit men to combat, not only hazarding their lives but ultimately involving the security of the Nation itself.” 72 If anything might 'undermine the effectiveness of response to command,” it is 'constitutionally unprotected' within the military, even if it would be 'protected in the civil population.'⁷³ As a result, courts permit 'greater breadth' and 'greater flexibility' to military rulemaking than they do for rulemaking in civilian society.⁷⁴ And they give less weight to challenges concerning individual rights because “within the military community there is simply not the same autonomy as there is in the larger civilian community.” 75

Rostker v. Goldberg afforded the same latitude to rules that implicate equal-protection rights, holding that Congress may authorize the registration of men, but not women, for the Page 82draft.⁷⁶ In the process, the Supreme Court clarified that military deference applies not just to final judgments but to evaluations of evidence, holding that the district court 'was quite wrong in undertaking an independent evaluation of th[e] evidence, rather than adopting an appropriately deferential examination of Congress’ evaluation of that evidence.'⁷⁷

And then in Goldman v. Weinberger , the Supreme Court made the governing principle even more explicit: “courts must give great deference to the professional judgment of military authorities concerning the relative importance of a particular military interest.” 78

Goldman considered whether an Air Force regulation preventing an Orthodox Jew and ordained rabbi from wearing his yarmulke while in uniform violated the First Amendment. The Court deferred to the military’s judgment. When military officials make a decision based on a 'perceived need for uniformity' - or presumably for any other aspect of force readiness - 'they are under no constitutional mandate to abandon their considered professional judgment.” 79

How far does the “great deference” afforded to military judgments extend?⁸⁰ At least this far: The Constitution 'does not require the military to accommodate' a practice as seemingly innocuous as wearing a yarmulke if the military Page 83thinks it “would detract from the uniformity sought by the dress regulations.” 81

What evidence had the Air Force provided that such a small accommodation would imperil troop discipline, much less national security? According to Goldman: none. He said the Air Force “failed to prove that a specific exception for his practice of wearing an unobtrusive yarmulke would threaten discipline”; that the Air Force’s position was “mere ipse dixit ”; that there was “no support from actual experience or a scientific study in the record'; and that expert testimony showed religious exceptions would improve morale.⁸²

The Supreme Court found all of this 'quite beside the point.” 83 It concluded that 'the military's perceived need for uniformity' was sufficient for the Air Force to 'draw[] the line” that it drew. 84 The Air Force considered 'the necessary habits of discipline and unity” to be “as vital during peacetime as during war because its personnel must be ready to provide an effective defense on a moment’s notice.” 85 And that was enough.

Might you question just how detrimental to “uniformity” it would be to allow a rabbi to wear a yarmulke — especially given that the military provided no evidence to the contrary? I would. And might you believe that the military should make Page 84such an accommodation for someone whose faith is so important to him? I do. But those are my views as a judge sitting far away from a military base.⁸⁶

The aftermath of Goldman shows the proper path for change if citizens and their representatives find a military decision objectionable. The very next year, Congress passed a law allowing service members to “wear an item of religious apparel while wearing the uniform” so long as the item met certain standards.⁸⁷ The lesson: Objectionable military rules can be changed, and quickly. Our constitutional order makes that possible. But it does not give that role to the judiciary.

The cases since Goldman have continued to reinforce its “great deference” standard.

In Solorio , the Supreme Court deferred to Congress's choices about the 'scope of court-martial jurisdiction over offenses committed by servicemen.'⁸⁸ That was because Congress, and not the courts, is entrusted with “the delicate task of balancing the rights of servicemen against the needs of the military.” 89

In Winter , the Supreme Court concluded that injunctive relief against the military was improper because “the balance of equities and consideration of the overall public interest . . . tip strongly in favor of the Navy' when a judicial decision protecting marine mammals could “forc[e] the Navy to deploy Page 85an inadequately trained . . . force.” 90 The Court relied specifically on Goldman 's 'great deference' standard, and provided yet another reason for that standard: 'neither the Members of this Court nor most federal judges begin the day with briefings that may describe new and serious threats to our Nation and its people.” 91 Courts should therefore 'defer to' military officials’ “specific, predictive judgments” about how injunctive relief could impede military effectiveness.⁹²

Most recently, in Austin v. U.S. Navy Seals 1–26 , members of the Naval Special Warfare community challenged the Navy's denial of religious exemptions from its COVID-19 vaccination mandate under the Religious Freedom Restoration Act and the Free Exercise Clause.⁹³ The Supreme Court partially stayed a preliminary injunction that had barred the Navy from taking action against them based on their unvaccinated status, allowing the Navy to consider vaccination status in 'deployment, assignment, and other operational decisions.” 94 Justice Kavanaugh - no foe of religious liberty 95 - concurred, explaining that the district court had 'in Page 86effect inserted itself into the Navy's chain of command, overriding military commanders’ professional military judgments,” even though courts “‘should indulge the widest latitude’ to sustain the President’s ‘function to command the instruments of national force.’” 96

The D.C. Circuit’s decisions say nothing to the contrary. Kreis v. Secretary of Air Force held that military personnel decisions are nonjusticiable when they would require the court 'to second-guess' military decisions 'about how best to allocate military personnel in order to serve the security needs of the Nation” — a question “[t]his court is not competent” to answer.⁹⁷

Page 87This court’s opinion in Singh v. Berger likewise confirms that we must defer to professional military judgments “unless Congress specifically has provided otherwise.” 98 Singh (which was not a decision about constitutional rights) held that a military policy violated the Religious Freedom Restoration Act (which is Congress's express instruction that even military burdens on religious exercise must satisfy strict scrutiny).⁹⁹ So even though Singh informs our understanding of Congress' legislative judgment, it offers no basis for replacing professional military judgment with a court’s judicial judgment.

C. The Constitution + Caselaw = “Great Deference”

Constitutional structure and binding precedent yield a consistent rule: Absent a contrary statutory command, courts reviewing a constitutional challenge to a military policy must give “great deference” to the political branches’ professional military judgments. Those judgments include military-wide decisions about the composition, training, equipping, and control of the armed forces, as well as assessments of readiness, discipline, deployability, cohesion, uniformity, command effectiveness, and mission accomplishment.

If a challenged system-wide policy reflects a professional military judgment about a perceived military need, a court may Page 88not reweigh the evidence, 100 credit contrary expert views over military judgment, 101 require proof that the plaintiff's proposed exception to a rule would itself impair military effectiveness, 102 or substitute its own view of the policy’s desirability.¹⁰³ Judges Page 89therefore have only one option when we determine that the political branches have made a system-wide 'professional military judgment' based on a perceived military need: We must defer to the military’s judgment.

A 'professional military judgment' is not identified by how much or how little consideration military professionals devoted to making the decision. Nor is a “professional military judgment” identified by how well-supported a court believes the decision to be. The cases discussed above applied no tests of that sort. Rather, a 'professional military judgment' depends on the decisionmaker and the subject matter. It is a “professional judgment of military authorities concerning the relative importance of a particular military interest .” 104 That is why Gilligan categorized “decisions as to the composition, training, equipping, and control of a military force” as 'essentially professional military judgments.'¹⁰⁵ Only the political branches, which have been entrusted with such decisions, should make them.

That is a “great” deferential standard, indeed — one that is fundamental to our nation’s constitutional structure.

Page 90The Constitution demands this deference not to disparage any service members 106 or deprive anyone of any rights.¹⁰⁷ And when we discount that deference, we endanger 'every current servicemember'¹⁰⁸ and make it more difficult for the military to defend the rights it has protected from Fordow and Fallujah to the Philippine Sea, and from Little Round Top to Little Rock.¹⁰⁹

Admittedly, reasonable people disagree about whether those who wrote “all men” meant “all people” — in other words, whether “the architects of our Republic” were “signing a promissory note to which every American was to fall heir,” even if the promise was unfulfilled in their time. Martin Luther King, Jr., 'I Have a Dream' Speech (Aug. 28, 1963). The “contested meaning” of the Declaration’s most

Page 91The military's transgender policy is constitutional. The plaintiffs’ arguments against it lack legal support. So do the majority’s.

Sometimes a case will present an open question informed only by confusing and contradictory precedents. This is not one of those cases. Rather, as described above, an unbroken line of unambiguous authorities directs us to uphold a military policy (1) when it does not conflict with a statutory command, (2) when it reflects a “professional military judgment[],” and (3) when it addresses a perceived, military-wide need Page 92regarding the “composition, training, equipping, [or] control” of the armed forces.¹¹⁰

The policy challenged today meets those criteria.

First, no statute prohibits the military from implementing the policy, and the plaintiffs do not contend otherwise. The district court applied the reasoning of Bostock v. Clayton County , a case arising under Title VII of the Civil Rights Act of 1964.¹¹¹ But the district court did not - and could not - hold that Bostock directly controls here, given that “Title VII does not apply to uniformed members of the armed forces.” 112

Second, the policy reflects a professional military judgment. It was announced by the military authorities Page 93authorized to make such decisions.¹¹³ And its subject matter was indisputably about a particular military interest: qualifications for military service.¹¹⁴ Such “decisions as to the composition . . . of a military force” are “essentially professional military judgments.” 115

Third, the policy involves “the professional judgment of military authorities concerning the relative importance of a particular military interest.” 116 It explains how that judgment is shaped by 'unique military exigencies.'¹¹⁷ And it even provides evidence to support that connection between the policy and those exigencies.

The primary interest that informs the military’s policy is the need for deployable troops. The action memo accompanying the policy cites evidence that nearly 40% of service members with gender dysphoria in a studied cohort “were non-deployable over a 24 month period.” 118 It also cites the demand for ongoing treatment, the potential for interruption of care in deployed environments, and the incompatibility of certain medical requirements with austere conditions.¹¹⁹ And Page 94it cites a finding that “the suicide attempt rate is estimated to be 13 times higher among transgender individuals compared to their cisgender counterparts.” 120

Those kinds of military judgments demand deference because they are predictive.¹²¹ And because they are 'complex, subtle, and professional.'¹²² And because they involve the weighing of risks in a way that has no obvious judicial benchmark.¹²³

Even if that were not enough — and it is — there’s more. The memorandum relies on prior military policies, expertpanel findings, internal reviews, and medical literature. For example, it cites a 2018 Department assessment that concluded 'substantial risks' arise when people 'with a history or Page 95diagnosis of gender dysphoria” join and remain in the military.¹²⁴ It also cites a 2021 Department of Defense review that found higher rates of disability evaluation and substantial non-deployability.¹²⁵ In addition, it cites a 2025 literature review that identified elevated mental-health risks and limits in the strength of available evidence.¹²⁶ And it considers cost data reflecting tens of millions of dollars in treatment expenditures.¹²⁷ The policy thus reflects a synthesis of military, medical, and operational judgments of the sort courts have consistently declined to reweigh. One might disagree with the military's conclusions, but it's hard to look at the reasons and evidence the military provided in its action memo and say that the military acted “unthinkingly” or “reflexively and not for any considered reason.” 128

The majority seizes on this line from Rostker - along with Rostker 's observation that 'Congress had extensively studied and debated the issue” there — to craft a new test. That test requires the military to prove that any classification it makes was 'based on detailed study rather than ‘archaic and overbroad generalizations.’” Majority Op. at 29. Does even intermediate scrutiny demand such proof? And if the military’s action memo citing multiple studies and relating them to multiple military needs was insufficient, what more is needed? The majority does not even acknowledge the military’s reliance on those studies, much less defer to the military’s weighing of them, so we cannot know why they fall short. So where within Page 96To repeat, the political branches do not need to make such a thorough showing to earn judicial deference. We are now far beyond what “great deference” requires.¹²⁹ I go this far only to show that the military has met and surpassed its burden.

Note for a moment what I am not saying. I am not saying the evidence must be read as the military reads it. Nor am I saying the military’s conclusions are the best conclusions. I am in no way qualified to know who is sufficiently likely to be deployable on a moment's notice, what level of risk is acceptable, or how to balance readiness against other values.

Fortunately for me, it is not our job to answer those questions. In cases like today’s, our role requires deference to the judgment of the Executive and Congress. That is why, in every binding precedent identified by the plaintiffs regarding the constitutionality of internal military matters like militarywide preparedness, the military won.¹³⁰

Page 97B. The Plaintiffs’ Argument

In light of that uninterrupted line of adverse decisions, the plaintiffs were asked at oral argument to identify the best binding precedent for their claim. They cited Singh v. Berger 131 and Rostker v. Goldberg . 132 The first is mildly relevant but only in a way that is, for them, mildly unhelpful;Page 98 the second is more relevant and in a way that is even more unhelpful.

Unlike today’s constitutional case about the right to equal protection, Singh was a statutory case about the Religious Freedom Restoration Act. This court held that the military likely violated a congressional mandate by making a Sikh Marine shave his beard.¹³³ So to the degree that Singh is at all relevant here, it works against the plaintiffs, given that they can point to no congressional mandate that favors their position.

As for Rostker , the military won that case. In it, the Supreme Court rejected an equal-protection challenge to the single-sex military draft.¹³⁴ Rostker spared no words when criticizing courts that too closely scrutinize the political branches’ judgments about the system-wide composition of the military, calling the district court “quite wrong” for “undertaking an independent evaluation of th[e] evidence.” 135

You might think that the plaintiffs' inability to name a helpful precedent was the low point of their oral argument. It wasn't. That came when they refused to say whether the Constitution requires the military to let service members depart from the standards for their birth sex.

When asked if the Constitution allows the military to impose men's physical fitness requirements on transgender women, the plaintiffs said, in effect, no answer. What about making them wear a man’s uniform? No answer. What about making them cut their hair like a man? No answer. What about making them use men’s showers? No answer. What about Page 99making them use men’s bathrooms? No answer. What about imposing on them men’s deployability standards? No answer. What about imposing on them men’s retention standards? No answer.¹³⁶

In lieu of responsive answers, the plaintiffs relied on generalities like “whatever policy exists, it has to be rational and not fueled by animus.” 137 But at least as a general matter, concern about medical conditions’ effects on deployability is neither irrational nor fueled by animus. And the plaintiffs don’t deny that gender dysphoria is a medical condition — one that can cause clinically significant distress.¹³⁸

That makes the plaintiffs' non-answers a bit puzzling. They say people suffer from gender dysphoria whenever they are “prevented from being able to live congruently with their gender identity.” 139 And they say the military cannot force all transgender people out of the military.¹⁴⁰ So doesn't that mean the military must let people “live congruently with their gender identity”?¹⁴¹ And if it does, why not say so?

Consider also the question that logically arises from the following four propositions, which the plaintiffs either endorsed or did not disclaim. First, the military might be able to make transgender people live congruently with their birth Page 100sex.¹⁴² Second, making them do so causes gender dysphoria.¹⁴³ Third, gender dysphoria is a medical condition that can cause significant distress.¹⁴⁴ Fourth, concern about the distress caused by a medical condition can at least sometimes be a rational basis for exclusion from the military.¹⁴⁵

Why, then, is that concern not a rational basis for the plaintiffs’ exclusion?

If you ask them, don’t expect an answer more responsive than “whatever policy exists, it has to be rational and not fueled by animus.'¹⁴⁶

The majority discounts the need for deference, weighs the evidence, and concludes that the military’s policy is arbitrary and fueled by animus.

Page 101Rather than deferring to the military's national-security assessment, the majority boldly declares that a “reversal to the status quo that has been held for four years cannot possibly be deemed a major threat to national security when no such national security threats were raised in years prior.” 147 In my view, that blanket proposition is hardly self-evident. But here again, it’s not my view (or the majority’s) that should matter. Judges this side of Midian and Moab do not “begin the day with briefings that may describe new and serious threats to our Nation and its people,” 148 and we 'have less competence' here than in almost any “area of governmental activity.” 149 So the Supreme Court and this court, in opinion after opinion, have afforded 'great deference' to the military's judgment about what is “a major threat to national security.” 150

The majority says the military “put forward no evidence” to rebut the plaintiffs’ experts 151 and has “not contested that all of the currently serving [plaintiffs] . . . have served honorably Page 102and met all military standards during their service.” 152 True. But that does not mean that the military has 'forfeited any argument that . . . retaining these servicemembers will harm national security.” 153 For 'complex, subtle, and professional decisions” like the standards for military service and whether a group of service members who have served well in the past are sufficiently likely to deploy adequately in the future, the military is the only expert that matters. And whether the plaintiffs’ “expert witnesses” agree with the military’s national-security assessment “is quite beside the point.” 154

In addition, the majority calls the policy “arbitrary” 155 for disqualifying anyone “with a history of gender dysphoria” 156 and denying waivers to anyone who has 'attempted to transition' from one gender to another.¹⁵⁷ But it's at least conceivable that a history of gender dysphoria, or a history of symptoms like an attempt to transition, makes a recurrence more likely. And according to the rational-basis standard the majority purports to apply, a 'conceivable' rationale is all Page 103that's needed.¹⁵⁸ That standard gives policymakers room to make rules without having to connect every dot in the inferential chain - for example, to explain why a person's medical history may indicate future medical vulnerability.

Several other military policies seem to make the same short inferential leap by excluding people with a history of medical and psychiatric conditions, even though the conditions may have come and gone long ago. For example, the military excludes people with a history of asthma or asthma-like symptoms “after the 13th birthday.” 159 Is that “inexplicable”?¹⁶⁰ It also excludes people with a history ofPage 104 'any eating disorder.'¹⁶¹ Is that a disqualification with 'no reasonable justification”?¹⁶²

The plaintiffs’ own assertions at oral argument provide a reasonable justification for why the military might choose to disqualify anyone prone to gender dysphoria - even if that condition has been treated. According to the plaintiffs, gender dysphoria results from someone being “unable to or prevented from being able to live congruently with their gender identity.” 163 So even people who have “undergone treatment” 164 may be expected to experience gender dysphoria again if required to serve in their birth sex. Indeed, the plaintiffs here say that none of them are able to comfortably 'adhere to the military's standards for their birth sex with respect to grooming, uniform, bathrooms, [and] where they sleep.” 165 Yet that is precisely what the new military policy requires. And (as you’ll recall from above) the plaintiffs take no position on whether such a requirement is constitutional.¹⁶⁶

Finally, for the same reasons that the policy does not otherwise fail the rational-basis standard the majority applies, Page 105the policy is not “based upon animus.” 167 It 'has a legitimate grounding in national security concerns'¹⁶⁸ — as do all military readiness policies. And courts must 'uphold the policy so long as it can reasonably be understood to result from a justification independent of unconstitutional grounds.” 169

I need not again enumerate the reasons this policy rests on a permissible justification, 170 and the majority's opinion does not explain which part of the military’s justification is unreasonable. Is it because the Constitution prohibits the military from using past diagnoses as a proxy for current or future risk of gender dysphoria — a proxy the military uses for many other medical conditions?¹⁷¹ Or is it because the Constitution prohibits the military from treating prior attempts to transition as evidence that a person may have once had gender dysphoria - even though the plaintiffs describe attempts to transition as a symptom of gender dysphoria?¹⁷² Or is it because the Constitution prohibits the military from making people serve in their preferred gender — an argument that the plaintiffs refused to make?¹⁷³

The first of those arguments conflicts with the military’s common practice in areas undoubtedly free of animus. The other two arguments conflict with the plaintiffs’ own words. None of them proves animus. Nor do they warrant the Page 106majority’s unprecedented intrusion into the internal operations of the armed forces.¹⁷⁴

To be sure, this is a different case in a different procedural posture. It is an appeal asking us to vacate a preliminary injunction, see Winter , 555 U.S. at 12, rather than an application asking the Supreme Court to stay a preliminary injunction, see Hollingsworth v. Perry , 558 U.S. 183, 190 (2010). But “the parties’ legal arguments and relative harms generally” are the same. Cf. Noem v. National TPS Alliance , 146 S. Ct. 23, 24 (2025). So there is more than a little reason to think that '[t]he same result that [the Supreme Court] reached [last] May is appropriate here.' Id.

The plaintiffs suggest that this case is distinguishable because the district court here found animus. Appellees' Br. at 48. That is unpersuasive. While the district court and the majority have made the role of animus more explicit than in Shilling and have treated it as independently dispositive, Shilling relied on the same underlying defects in the Government's justification. In Shilling , the district court concluded that the challenged policy was unsupported by evidence, rested on stigmatizing and derogatory characterizations, and bore no genuine connection to the government's asserted interests. See Shilling v. United States , 773 F. Supp. 3d 1069, 1094– 97, 1101 (W.D. Wash. 2025). The district court there rejected the Government’s reliance on speculative “predictions” that ignored “the reality of years of open service,” and emphasized the Government’s

Page 107'We know that from top to bottom of the Army the complaint is often made, and sometimes with justification, that there is discrimination, favoritism or other objectionable handling of men. But judges are not given the task of running the Army.'¹⁷⁵ Only the Executive and Congress are responsible for system-wide military judgments about the composition of the armed forces.

The Supreme Court has never assumed that role for itself.

Neither has the D.C. Circuit.

Not until today.