USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 1 of 59
FOR PUBLICATION
In the
United States Court of Appeals
For the Eleventh Circuit
No. 25-11881
COMPUTER & COMMUNICATIONS INDUSTRY
ASSOCIATION,
NETCHOICE,
Plaintiffs-Appellees,
versus
JAMES UTHMEIER,
Defendant,
ATTORNEY GENERAL, STATE OF FLORIDA,
Defendant-Appellant.
Appeal from the United States District Court
for the Northern District of Florida
D.C. Docket No. 4:24-cv-00438-MW-MAF
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 2 of 59
2 Opinion of the Court 25-11881
Before NEWSOM, LUCK, and TJOFLAT, Circuit Judges.
NEWSOM, Circuit Judge:
For better or worse, the proliferation of social media has
transformed American life. Unsurprisingly, it has also generated “a
raft of public policy issues.” Moody v. NetChoice, LLC, 603 U.S. 707,
716 (2024). One of the most pressing is how and in what ways social-media use affects minors. According to the State of Florida,
social-media companies seed their platforms with “addictive features”—push notifications, auto-play, infinite scrolling, etc.—in order to make them, well, addictive. These features, the state says,
are “psychological hooks” that “overstimulate the reward center in
the brain” and “foster compulsive overuse.” Br. of Appellant at 4
(citations omitted). Florida insists that children are particularly susceptible to these “addictive features” because their brains are still
developing. In particular, the state contends, the features contribute to “[c]ompulsive or uncontrollable” social-media use among
children and threaten a “youth mental health crisis.” Id. at 1, 5.
Children face “‘sleep problems, attention problems, and feelings of
exclusion,’ all of which put [them] at risk of ‘depression, anxiety,
and neuroticism.’” Id. at 5 (citations omitted).
In an effort to protect children from social-media addiction
and its ensuing effects, Florida passed HB3, which imposes age-restriction requirements on platforms that use “addictive features.”
Fla. Stat. § 501.1736(1)(e). In particular, HB3 requires platforms
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 3 of 59
25-11881 Opinion of the Court 3
within its ambit to prohibit anyone under the age of 14 from creating an account, and to prohibit anyone else under 16 from doing so
without parental consent. Id. § 501.1736(2)–(3).
The Computer & Communications Industry Association
and NetChoice, both Internet trade organizations—to which, collectively, we will hereinafter refer as “CCIA”—brought facial and
as-applied First Amendment challenges to HB3. The district court
concluded that HB3 was likely unconstitutional and preliminarily
enjoined Florida from enforcing it. We stayed the injunction pending appeal. Before us, Florida contends (1) that CCIA lacks standing to assert the First Amendment rights of minors, (2) that the district court should have abstained from deciding the case under
Younger v. Harris, 401 U.S. 37 (1971), and (3) that HB3 satisfies First
Amendment scrutiny.
Our analysis proceeds in three parts. First, as threshold matters, we hold that CCIA has standing to challenge HB3 and that
Younger abstention doesn’t foreclose its claim. Second, we address
the merits of CCIA’s as-applied challenge, and we hold that HB3
likely passes First Amendment muster because it is a content-neutral speech regulation that satisfies intermediate scrutiny. Finally,
we reject CCIA’s facial challenge to HB3 under the principles articulated in Moody.1
1 CCIA and NetChoice have challenged the constitutionality of similar laws in
other states restricting minors’ use of social-media platforms. See, e.g.,
NetChoice v. Carr, 789 F. Supp. 3d 1200 (N.D. Ga. 2025) (Georgia), appeal docketed, No. 25-12436 (11th Cir. July 16, 2025); NetChoice, LLC v. Skrmetti, No. 25-USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 4 of 59
4 Opinion of the Court 25-11881
I
A
We begin with a primer on HB3. The law defines a “social
media platform” as “an online forum, website, or application”
(1) that “[a]llows users to upload content or view the content or
activity of other users”; (2) on which at least 10% of daily active
users under 16 years old spend an average of at least two hours per
day; (3) that “[e]mploys algorithms that analyze user data or information on users to select content for” them; and (4) that “[h]as any
of the following addictive features”: “[i]nfinite scrolling,” “[p]ush
notifications,” displays of “personal interactive metrics,” “[a]utoplay,” or “[l]ive-streaming.” Fla. Stat. § 501.1736(1)(e). HB3 defines each “addictive feature” as follows:
• “Infinite scrolling” refers to either (1) the “[c]ontinuous[]
loading” of content without the need to open a separate
page, or (2) the “[s]eamless” loading of content “with no visible or apparent end or page breaks.”
5660, 2026 WL 2546423, at *2 (6th Cir. Aug. 28, 2026) (Tennessee); Comput. &
Commc’ns Indus. Ass’n v. Paxton, 184 F.4th 420, 437 (5th Cir. 2026) (Texas);
NetChoice, LLC v. Yost, 180 F.4th 268 (6th Cir. 2026) (Ohio); NetChoice, LLC v.
Bonta, 152 F.4th 1002 (9th Cir. 2025) (California); NetChoice, L.L.C. v. Fitch, 134
F.4th 799 (5th Cir. 2025) (Mississippi); NetChoice v. Murrill, 812 F. Supp. 3d 594
(M.D. La. 2025) (Louisiana); NetChoice, LLC v. Griffin, No. 5:23-CV-5105, 2025
WL 978607 (W.D. Ark. Mar. 31, 2025) (Arkansas); NetChoice, LLC v. Reyes, 748
F. Supp. 3d 1105 (D. Utah 2024) (Utah).
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 5 of 59
25-11881 Opinion of the Court 5
• “Push notifications” are “alerts” sent by the platform “to inform a user about specific activities or events related to the
user’s account.”
• “Personal interactive metrics” refer to “[d]isplays . . . indicat[ing] the number of times other users have clicked a
button to indicate their reaction to content or have shared
or reposted the content”—think “likes,” “comments,” “reposts,” “shares,” etc.
• “Auto-play” is a feature that enables a video to “play without
the user first clicking on the video or on a play button for
that video.”
• “Live-streaming” is a “function that allows a user or advertiser to broadcast live video content in real-time.”
Id. § 501.1736(1)(e)(4)(a)–(e). HB3 exempts from its definition of
“social media platform” any “online service, website, or application
where the exclusive function is e-mail or direct messaging consisting of text, photographs, pictures, images, or videos shared only
between the sender and the recipients.” Id.
HB3 requires covered platforms (1) to prohibit minors under
the age of 14 from creating accounts and (2) to prohibit 14- and 15-year-olds from creating accounts without parental consent. Id.
§ 501.1736(2)(a), (3)(a). If an account holder is under 14, or is 14 to
15 but doesn’t have parental consent, the platform must terminate
her account. Id. § 501.1736(2)(b)(1), (3)(b).
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 6 of 59
6 Opinion of the Court 25-11881
B
On October 28, 2024, shortly before HB3 took effect, CCIA
brought facial and as-applied First Amendment challenges to HB3’s
pertinent provisions. The next day, CCIA moved for a preliminary
injunction to prevent HB3’s enforcement. A flurry of docket activity ensued. (The procedural recitation that follows may seem
pretty tedious, but it’s relevant to the Younger issue before us.)
Lawyers filed notices of appearance and withdrawal, as well as pro
hac vice motions, on which the district court ruled. Florida filed a
motion to dismiss, to which CCIA responded. The state also filed
an opposition to CCIA’s preliminary-injunction motion, and CCIA
filed a reply in support. In addition, Florida deposed four industry
witnesses. See Cleland Dep., Dkt. No. 51-4 (NetChoice’s General
Counsel); Schruers Dep., Dkt. No. 51-5 (CCIA’s President and
CEO); Boyle Dep., Dkt. No. 51-6 (Snap’s Senior Director of Product Management); Veitch Dep., Dkt. No. 51-7 (YouTube’s Director
of Government Affairs and Public Policy across the Americas).
Following a hearing, the district court denied CCIA’s preliminary-injunction motion on associational standing grounds. The
court held that CCIA had failed to show that any of its members’
asserted injuries were “actual or imminent” and “fairly traceable”
to the state because it didn’t allege “that at least one specific member of [CCIA] [was] likely regulated by the law.” Dist. Ct. Order,
Mar. 13, 2025, at 6. The court also granted Florida’s motion to dismiss CCIA’s complaint. CCIA promptly filed an amended complaint that primarily sought to remedy the standing defects that the
district court had identified and renewed its preliminary-injunction
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 7 of 59
25-11881 Opinion of the Court 7
motion. The court held a scheduling conference and set a briefing
schedule. In the meantime, CCIA deposed three of the state’s expert witnesses. See Twenge Dep. 32–34, 53–57, Dkt. No. 63-1 (opining that using social media causes, or at least correlates with, harmful outcomes for children); Allen Dep. 12–16, 58–139, Dkt. No. 63-2 (explaining how age-verification systems function); Alter Dep.
27–54, Dkt. No. 63-3 (describing a theory of how design features of
social-media platforms trigger behavioral addiction).
On April 21, 2025, before the district court had ruled on
CCIA’s renewed preliminary-injunction motion, Florida filed an
enforcement action in state court against Snap, a social-media company and member of CCIA’s co-plaintiff NetChoice, alleging violations of HB3. Snap removed the enforcement action to federal district court. 2
The district court then granted CCIA’s renewed preliminary-injunction motion. As relevant here, the court determined
(1) that CCIA had standing, (2) that Younger didn’t require dismissal
of CCIA’s suit, and (3) that although HB3 was likely a content-neutral speech regulation, it was likely insufficiently tailored to survive
2 Notice of Removal, Off. of the Att’y Gen., State of Fla., Dep’t of Legal Affs. v. Snap
Inc., No. 3:25-cv-676 (N.D. Fla. May 21, 2025), Dkt. No. 1. On the same day
that we heard oral argument in this case, we heard argument in the appeal
from the district court’s order declining to remand the enforcement action to
state court. See Off. of the Att’y Gen., State of Fla., Dep’t of Legal Affs. v. Snap Inc.,
No. 25-12814.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 8 of 59
8 Opinion of the Court 25-11881
intermediate scrutiny. The court thus enjoined Florida from enforcing the challenged provisions of HB3.
We stayed the preliminary injunction pending appeal. In
our stay order, we agreed with the district court (albeit tentatively)
that CCIA had standing, that Younger abstention didn’t apply, and
that the statute was likely content neutral. Nevertheless, we concluded (again, tentatively) that HB3 was sufficiently tailored to satisfy intermediate scrutiny. We therefore granted the stay on the
ground that Florida was likely to succeed on the merits.
Following oral argument of the appeal of the district court’s
preliminary-injunction order, we directed the parties to submit
supplemental briefing on the questions (1) whether CCIA has
standing and (2) whether the district court decided CCIA’s as-applied challenge.
II
Before diving into the merits, we need to address two
threshold questions: (1) whether CCIA has standing to pursue its
First Amendment challenge, and (2) whether the district court
should have abstained from deciding this case under Younger. For
reasons we’ll explain, we answer those questions yes and no, respectively.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 9 of 59
25-11881 Opinion of the Court 9
A
First, standing.3 “Article III standing is a ‘bedrock constitutional requirement.’” Food & Drug Admin. v. All. for Hippocratic
Med., 602 U.S. 367, 378 (2024) (quoting United States v. Texas, 599
U.S. 670, 675 (2023)). To establish standing, “a plaintiff must
demonstrate (i) that she has suffered or likely will suffer an injury
in fact, (ii) that the injury likely was caused or will be caused by the
defendant, and (iii) that the injury likely would be redressed by the
requested judicial relief.” Id. at 380.
CCIA asserts standing to vindicate the First Amendment
rights of its members—i.e., the social-media platforms. The Supreme Court has recognized that “an organization may sue to redress its members’ injuries” under what is known as “associational
standing.” United Food & Com. Workers Union Loc. 751 v. Brown Grp.,
517 U.S. 544, 552 (1996). It’s a “form of third-party standing” that
arguably “relax[es] both the injury and redressability requirements
for Article III standing.” All. for Hippocratic Med., 602 U.S. at 398–
99 (Thomas, J., concurring).
Under associational-standing principles, an organization
may sue on its members’ behalf when (1) “its members would otherwise have standing to sue in their own right,” (2) “the interests it
seeks to protect are germane to the organization’s purpose,” and
(3) “neither the claim asserted nor the relief requested requires the
3 “We review the legal question of standing de novo.” Charles H. Wesley Educ.
Found. v. Cox, 408 F.3d 1349, 1351 (11th Cir. 2005).
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 10 of 59
10 Opinion of the Court 25-11881
participation of individual members in the lawsuit.” Brown Grp.,
517 U.S. at 553 (quoting Hunt v. Wash. State Apple Advert. Comm’n,
432 U.S. 333, 343 (1977)).
1
To satisfy the associational-standing test’s first requirement,
at least one of CCIA’s members must itself meet the three Article III standing elements: (1) injury in fact, (2) causation, and (3) redressability. See id.
An “injury in fact” is “an invasion of a legally protected interest which is (a) concrete and particularized, and (b) actual or imminent, not conjectural or hypothetical.” Lujan v. Defs. of Wildlife,
504 U.S. 555, 560 (1992) (citation modified). In the pre-enforcement-challenge context, a “plaintiff satisfies the injury-in-fact requirement where he alleges an intention to engage in a course of
conduct arguably affected with a constitutional interest, but proscribed by a statute, and there exists a credible threat of prosecution
thereunder.” Susan B. Anthony List v. Driehaus, 573 U.S. 149, 159
(2014) (citation modified). In this circuit, “[t]he injury requirement
is most loosely applied—particularly in terms of how directly the
injury must result from the challenged government action—where
[F]irst [A]mendment rights are involved, because of the fear that
free speech will be chilled even before the law, regulation, or policy
is enforced.” Speech First, Inc. v. Cartwright, 32 F.4th 1110, 1120
(11th Cir. 2022). A First Amendment plaintiff has standing if “the
operation or enforcement of the government policy would cause a
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 11 of 59
25-11881 Opinion of the Court 11
reasonable would-be speaker to self-censor—even where the policy falls short of a direct prohibition against the exercise of First
Amendment rights.” Id. (citation modified). “The fundamental
question under our precedent . . . is whether the challenged policy
‘objectively chills’ protected expression.” Id.
At least a handful of CCIA’s members satisfy the injury-infact requirement. The alleged conduct in which Snap, YouTube,
and Meta engage—and intend to continue to engage—is “arguably
affected with a constitutional interest” that is “proscribed by” HB3.
Susan B. Anthony List, 573 U.S. at 159. CCIA contends that its member platforms are likely to suffer First Amendment injuries because
HB3 will hinder their ability to communicate with their users
through “curated compilations of third-party speech,” Supp. Br. of
Appellees at 3–4, and features such as “[p]ush notifications” and
“personal interactive metrics,” Am. Compl. ¶¶ 23, 36, 49. “[C]ompiling and curating others’ speech” is an “expressive activity” protected by the First Amendment. Moody, 603 U.S. at 731. The Moody
Court concluded, based on the record in that case, that Facebook’s
and YouTube’s decisions about “which third-party content those
feeds will display, or how the display will be ordered and organized,” are “expressive choices” meriting First Amendment protection. Id. at 740. The same reasoning applies here. CCIA has alleged
that Snap, YouTube, and Meta (through Facebook and Instagram)
curate content “that they think will be particularly relevant to”
their users and choose that content for “a particular user based in
part on the content that the user has interacted with in the past.”
Am. Compl. ¶ 48; see also id. ¶¶ 22, 35. Therefore, by restricting
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 12 of 59
12 Opinion of the Court 25-11881
some users’ ability to create accounts, HB3 would limit the platforms’ ability to engage in the expressive activity of delivering curated content to those users.
Moreover, at least some of CCIA’s members face a “credible
threat of prosecution.” Susan B. Anthony List, 573 U.S. at 159 (citation omitted). We’ve said that “when a plaintiff challenges a law
soon after it was enacted and the state vigorously defends the law
in court, we can infer a credible threat of prosecution exists.” Shen
v. Comm’r, Fla. Dep’t of Agric. & Consumer Servs., 158 F.4th 1227,
1242 (11th Cir. 2025) (citation modified). That’s what happened
here: The state has been nothing if not vigorous. And the record
shows that Snap and YouTube, for example, meet HB3’s coverage
criteria. The enforcement of HB3 would not only objectively chill
these platforms from delivering curated content to the younger users HB3 covers, see Speech First, Inc., 32 F.4th at 1120, but would
also pressure them into instituting “significant and costly compliance measures” to implement age-verification and parental-verification systems. Am. Compl. ¶ 21 (Snap); id. ¶ 34 (YouTube); id.
¶ 47 (Meta). Cf. Virginia v. Am. Booksellers Ass’n, 484 U.S. 383, 392
(1988) (holding that the plaintiffs met the injury-in-fact requirement because, “if their interpretation of the statute is correct, [they]
will have to take significant and costly compliance measures or risk
criminal prosecution”). CCIA has thus shown that at least one of
its members is likely to suffer an injury in fact.
The remaining requirements for Article III standing—causation and redressability—are clearly satisfied by at least one CCIA
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 13 of 59
25-11881 Opinion of the Court 13
member. The “defendant’s conduct”—here, Florida’s enactment
of HB3—is the cause of the platform members’ First Amendment
injuries, and relief in the form of an injunction prohibiting HB3’s
enforcement would doubtlessly redress them.
Accordingly, CCIA has met the first prong of the associational-standing test by showing that at least one of its members has
Article III standing.
2
As for the second prong of the associational-standing test,
we agree with the district court—and Florida doesn’t contest—that
the interests that CCIA “aims to protect are germane to [its] purposes.” Brown Grp., 517 U.S. at 553.
3
CCIA also satisfies the third and final prong—namely, that
its claim can be resolved and requested relief can be granted without the participation of its individual members. Brown Grp., 517
U.S. at 553. The Supreme Court has called the individual-participation element a “judicially self-imposed limit” that aims to serve
“administrative convenience and efficiency.” Id. at 555–57. Unlike
the first two prongs, which are “constitutional necessit[ies],” this
third prong is “prudential.” Id. at 555–56. Meeting it rebuts the
“background presumption” that a litigant (here, CCIA) may not assert the rights of absent third parties (here, its member platforms).
Id. at 557.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 14 of 59
14 Opinion of the Court 25-11881
In similar circumstances, a number of other courts have held
that CCIA’s co-plaintiff, NetChoice, satisfies the associationalstanding test’s individual-participation prong. NetChoice L.L.C. v.
Fitch, 134 F.4th 799, 805 (5th Cir. 2025) (“NetChoice’s ‘claims can
be proven by evidence from representative injured members’ and
‘the participation of [certain] individual members does not thwart
associational standing.’”); CCIA v. Paxton, 747 F. Supp. 3d 1011,
1030 (W.D. Tex. 2024), aff’d in relevant part, 184 F.4th 420, 445 (5th
Cir. 2026); NetChoice v. Murrill, 812 F. Supp. 3d 594, 629–31 (M.D.
La. 2025); NetChoice, LLC v. Reyes, 748 F. Supp. 3d 1105, 1118–19 (D.
Utah 2024); NetChoice, LLC v. Griffin, No. 5:23-CV-5105, 2023 WL
5660155, at *10 (W.D. Ark. Aug. 31, 2023). We agree with them.
The state argues that CCIA doesn’t have associational standing because, on the merits, its First Amendment claim requires “numerous factual determinations about individual platforms.” Supp
Br. of Appellant at 6. For support, it cites Harris v. McRae, 448 U.S.
297, 321 (1980), and Georgia Cemetery Ass’n v. Cox, 353 F.3d 1319,
1322 (11th Cir. 2003). But both cases are distinguishable. In Harris,
a pro-choice division of a religious organization asserted that federal abortion-funding restrictions violated the First Amendment’s
Free Exercise Clause. 448 U.S. at 321. The Supreme Court held
that the group couldn’t proceed under an associational-standing
theory because adjudicating the free-exercise issues would necessarily entail an inquiry into whether and to what extent the restrictions operated coercively on particular members. Id. In Cox, a
cemetery association brought a Takings Clause challenge to a law
capping fees for the transfer of burial rights. 353 F.3d at 1322. We
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 15 of 59
25-11881 Opinion of the Court 15
held that the group couldn’t assert associational standing because
resolving its takings claim required the participation of its members. “[T]he economic impact,” we reasoned, “will vary depending
upon the economic circumstances of” particular cemeteries. Id.
Generally, whether a violation occurred in the regulatory-takings
context requires an individualized inquiry into whether each affected party was denied the compensation it was due. Tahoe-Sierra
Pres. Council v. Tahoe Reg’l Plan. Agency, 535 U.S. 302, 322 (2002)
(characterizing regulatory-takings jurisprudence as entailing “‘essentially ad hoc, factual inquiries,’ designed to allow ‘careful examination and weighing of all the relevant circumstances’” (citation
omitted)).
This case is different. While free-exercise and takings claims
require consideration of the law’s effects on individual parties,
CCIA’s free-speech claim, which it brings on behalf of its member
platforms, turns on whether HB3 implicates the First Amendment,
whether it is content based or content neutral, and whether it is
appropriately tailored to the asserted state interest—inquiries that
don’t depend on fact-intensive determinations regarding the member platforms. Nor does CCIA’s requested remedy require the individual members’ participation—for the simple reason that it
seeks declaratory and injunctive relief, rather than damages. See
Brown Grp., 517 U.S. at 553–54 (“If in a proper case the association
seeks a declaration, injunction or some other form of prospective
relief, it can reasonably be supposed that the remedy, if granted,
will inure to the benefit of those members of the association actually injured.” (citation modified)).
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 16 of 59
16 Opinion of the Court 25-11881
The state separately argues that First Amendment “claim[s]
related to algorithmic speech” are “fact intensive” and therefore
will “surely vary from platform to platform.” Supp. Br. of Appellant at 10 (quoting NetChoice, LLC v. Bonta, 152 F.4th 1002, 1014 (9th
Cir. 2025)). Not necessarily. To be sure, in Bonta, on which the
state relies, the Ninth Circuit held that a challenge brought by
NetChoice required the participation of individual member platforms. But the law and the challenge there don’t resemble those
here. In Bonta, NetChoice sued to enjoin the enforcement of a law
that restricted minors’ access to “personalized feeds” generated by
the platforms’ content-recommendation algorithms. 152 F.4th at
1010, 1014. Based on NetChoice’s acknowledgment that “each of
its members” and their algorithms were “unique,” the Ninth Circuit concluded that the law would implicate the First Amendment
only with respect to platforms whose algorithms were “expressive.” Id. at 1014. The determination whether each platform’s algorithm was expressive, the court concluded, was “fact intensive”
and required individual members’ participation. Id. HB3 is very
different; its restrictions are tethered not to each platform’s unique
algorithmic design but to so-called addictive features that are present across platforms.
We are satisfied that CCIA has satisfied the third and final
prong of associational standing.
* * *
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 17 of 59
25-11881 Opinion of the Court 17
For these reasons, we hold that CCIA has associational
standing and may bring this action on behalf of its members. 4
B
We next address whether Younger v. Harris, 401 U.S. 37
(1971), required the district court to abstain from ruling on CCIA’s
4 In both its complaint and its initial brief on appeal, CCIA separately contended that it also has third-party standing to vindicate the First Amendment
rights of its members’ users. See, e.g., Am. Compl. ¶ 50 (“CCIA and NetChoice
have standing to assert . . . the First Amendment rights of their members’ current and prospective users.”); id. ¶¶ 7, 8 (same); Br. of Appellees at 43 (arguing
that CCIA and NetChoice have “‘prudential standing’ to assert the First
Amendment rights of their members’ users”). In its supplemental brief,
though, CCIA seems to have retreated from its original, standing-based position, and now it says instead that its contentions regarding users’ rights were
meant only to supplement the merits inquiry. See Supp. Br. of Appellees at 15
(calling the assertion of prudential standing a “misunderstanding”); id. (“Plaintiffs invoke users’ rights not to get into court or expand their universe of
claims, but because users are a necessary component of the requisite First
Amendment inquiry[.]”). In any event, CCIA’s assertion of third-party standing to pursue its members’ users’ rights runs more to its facial overbreadth
claim than its as-applied claim. But for reasons we explain in Part IV, we conclude that, at least on the record as it currently exists, CCIA hasn’t made the
showing necessary to proceed with any facial challenge. If CCIA wishes to
continue to assert the rights of its members’ users in connection with a facial
overbreadth challenge, the district court may address those issues then.
One more standing-related loose end: CCIA separately asserts thirdparty standing to pursue its members’ users’ rights on jus tertii grounds, which
could apply to its as-applied claim. We needn’t address this argument because
we find that CCIA has associational standing.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 18 of 59
18 Opinion of the Court 25-11881
preliminary-injunction motion in light of Florida’s state-court enforcement action against Snap, a member of NetChoice.5
“[A]bstaining from the ‘exercise of federal jurisdiction is the
exception, not the rule.’” New Ga. Project, Inc. v. Att’y Gen. of Ga.,
106 F.4th 1237, 1241 (11th Cir. 2024) (quoting Colo. River Water Conservation Dist. v. United States, 424 U.S. 800, 813 (1976)). Accordingly, only the “clearest of justifications merits abstention.” Id.
(quoting Jackson-Platts v. Gen. Elec. Cap. Corp., 727 F.3d 1127, 1140
(11th Cir. 2013)). In Younger, the Supreme Court recognized a “limited exception to the usual rule.” Id.
Under Younger, a federal court must abstain if (1) “there is an
ongoing state judicial proceeding,” (2) “the state proceeding implicates important state interests,” and (3) “there is an adequate opportunity in the state proceeding to raise constitutional challenges.” Id. at 1242 (citation modified) (quoting Middlesex Cnty. Ethics Comm. v. Garden State Bar Ass’n, 457 U.S. 423, 432 (1982)). The
question before us involves the first requirement—whether Florida’s enforcement action against Snap was “ongoing” in the relevant sense. Id.
Perhaps counterintuitively, a state proceeding is deemed to
be “ongoing” for Younger purposes “even if it’s not actually pending
when the federal suit is filed—provided that it’s instituted . . . ‘before any proceedings of substance on the merits have taken place
5 “We review a Younger abstention decision for an abuse of discretion.” Hughes
v. Att’y Gen. of Fla., 377 F.3d 1258, 1262 (11th Cir. 2004).
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 19 of 59
25-11881 Opinion of the Court 19
in federal court.’” Id. at 1243 (quoting Hicks v. Miranda, 422 U.S.
332, 349 (1975)) (emphasis in original). That means that if the federal suit is still in an “embryonic stage” and “no contested matter
has been decided” by the federal court by the time the state brings
its enforcement action, the latter will be treated as if it were “ongoing”—meaning that the federal court must abstain from further litigation. Tokyo Gwinnett v. Gwinnett Cnty., 940 F.3d 1254, 1268 (11th
Cir. 2019).
Whether “proceedings of substance on the merits” have occurred in a federal suit entails a “fact-specific assessment of the circumstances.” Id. at 1272. That assessment depends on “the time
that the district court has spent considering the case, any motions
ruled on, any discovery, the number of conferences [or hearings]
held, and any change in the parties’ position as a result of the federal litigation.” Id. “[T]he filing of motions” is also relevant, “even
if the district court did not rule” on them. Id. And even “significant
and substantial procedural steps” can qualify as “proceedings of
substance on the merits” sufficient to preclude abstention. Id. at
1271–72.
Florida brought its enforcement action against Snap nearly
six months after CCIA initially filed its lawsuit, and during that
span, “proceedings of substance on the merits” undoubtedly occurred in federal court. Let’s recap: (1) Florida moved to dismiss
CCIA’s action, and (2) CCIA responded; (3) Florida filed an opposition to CCIA’s preliminary-injunction motion, and (4) CCIA reUSCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 20 of 59
20 Opinion of the Court 25-11881
plied; (5) the district court held a hearing on CCIA’s motion, a substantial portion of which was dedicated to a discussion of the merits
of CCIA’s claim, see Prelim. Inj. Hr’g Tr. 21–41, 68–92, 101–114,
123–135, Dkt. No. 71; (6) the court then denied CCIA’s motion (albeit on standing grounds); (7) CCIA filed an amended complaint
and (8) renewed its preliminary-injunction motion; and (9) the district court convened a status conference and (10) issued a briefing
schedule.
Along the way, (11) the parties conducted limited discovery,
including several depositions, all of which bore on the merits to
some extent. In particular, the state deposed four industry witnesses, see Cleland Dep., Dkt. No. 51-4 (testifying about
NetChoice’s members, their business practices, and their age-verification systems); Schruers Dep., Dkt. No. 51-5 (testifying about
CCIA’s members, their business practices, and their parental controls); Boyle Dep., Dkt. No. 51-6 (testifying about Snapchat’s features and parental controls); Veitch Dep., Dkt. No. 51-7 (testifying
about YouTube’s features and age-based policies and systems). So
too, CCIA deposed three of Florida’s expert witnesses, see Twenge
Dep. 32–34, 53–57, Dkt. No. 63-1 (testifying about how social-media features trigger harmful behavior in children); Allen Dep. 12–
16, 58–139, Dkt. No. 63-2 (testifying about how age-verification systems function); Alter Dep. 27–54, Dkt. No. 63-3 (testifying about
research on behavioral addiction and features of digital platforms
that make them difficult to resist).
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 21 of 59
25-11881 Opinion of the Court 21
To recount this history, we think, is to answer the Younger
question: By the time Florida brought its enforcement action
against Snap, CCIA’s federal case had passed the “embryonic
stage.” Tokyo Gwinnett, 940 F.3d at 1268. The district court did not
abuse its discretion in declining to abstain under Younger.
III
On, then, to the merits. 6 As already explained, CCIA presented, and the district court addressed, both facial and as-applied
challenges to HB3.7 We will consider CCIA’s as-applied challenge
6 “We review the district court’s grant of a preliminary injunction for abuse of
discretion” and the underlying questions of law de novo. Noble Prestige Ltd. v.
Galle, 83 F.4th 1366, 1374 (11th Cir. 2023).
7 Candidly, the district court’s order is less than pellucid. On one reading, the
court’s preliminary injunction might be understood as invalidating HB3 only
on its face. See Dist. Ct. Order, June 3, 2025, at 57 n.33 (“[A]lthough Plaintiffs’
amended complaint and motion for preliminary injunction do not expressly
state that they are raising a facial challenge, their briefing indicates that they
are . . . . Regardless, even if Plaintiffs had not brought a facial challenge, this
Court would be permitted to evaluate the law’s constitutionality on its face.”);
id. at 58 (“Defendant Uthmeier must take no steps to enforce [the challenged
provisions of HB3] until otherwise ordered.”).
Notably, though, in their supplemental briefs before us, the parties
agree that the district court addressed CCIA’s as-applied claim, as well. See
Supp. Br. of Appellant at 20–21; Supp. Br. of Appellees at 19–20. We agree
with them. Understood in the context of the district court’s entire order—and
even more so in the light of the parties’ competing contentions before that
court—we think it clear that the court preliminarily enjoined HB3 both on its
face and as applied to CCIA’s members. See, e.g., Dist. Ct. Order, June 3, 2025,
at 25 (“This Court begins by noting that Plaintiffs bring both facial and as-applied challenges to Florida’s law . . . .”); id. at 28 (observing that much of the USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 22 of 59
22 Opinion of the Court 25-11881
first, and then deal with its facial challenge—which presents unique
issues—in Part IV. Our evaluation of CCIA’s as-applied challenge
proceeds in three parts. First, we ask whether HB3 implicates the
First Amendment at all. Second, if it does, we must determine
whether HB3 is content based or content neutral. And finally, we
must decide whether HB3 survives the requisite level of scrutiny.
A
A law implicates the First Amendment—and thus triggers
constitutional scrutiny of some sort—if it either (1) “directly regulate[s] protected expressive activity” or (2) regulates even non-expressive conduct in a way that “impose[s] a disproportionate burden upon those engaged in protected First Amendment activities.”
TikTok, Inc. v. Garland, 604 U.S. 56, 68 (2025) (per curiam) (quoting
Arcara v. Cloud Books, Inc., 478 U.S. 697, 703–04 (1986)).
Florida contends that HB3 “regulates product design, not
children’s speech,” and, accordingly, doesn’t even implicate the
First Amendment. Br. of Appellant at 27. In particular, the state
says, HB3 restricts only the creation of accounts, not social-media
First Amendment analysis “is the same for both the facial and as-applied challenges,” and explaining the district court’s view that only at the tailoring stage
does a court “separately consider[] whether Plaintiffs have carried their burden
to sustain a facial challenge”); id. at 52–53 (“[T]his Court concludes that Plaintiffs are substantially likely to succeed not just in their as-applied challenge but
in their facial challenge as well.”).
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 23 of 59
25-11881 Opinion of the Court 23
access writ large, and thus regulates only “commercial transaction[s]”—namely, a platform’s ability to “enter[] into a contract”
with children of a certain age. Id. at 28.
Even accepting the state’s characterization, it seems clear
enough to us that HB3 “disproportionate[ly] burden[s]” CCIA’s
member platforms’ “protected First Amendment activities,” TikTok, 604 U.S. at 68—namely, “compiling and curating others’
speech” for their users, Moody, 603 U.S. at 731. In particular, HB3
burdens platforms’ ability to communicate with their users by requiring them (1) to prohibit minors under 14 years of age from creating or maintaining accounts and (2) to secure parental consent
before communicating with 14- and 15-year-olds.
We have little trouble concluding that HB3’s restrictions implicate the First Amendment—and thus trigger constitutional scrutiny. The trickier question, to which we turn next: What kind of
scrutiny?
B
First Amendment doctrine distinguishes between speech restrictions that are “content based” and those that are “content neutral.” Free Speech Coal. v. Paxton, 606 U.S. 461, 471 (2025). A content-based regulation triggers “strict” constitutional scrutiny,
meaning that it must represent “the least restrictive means” of accomplishing a compelling governmental interest. Id. (quoting
McCullen v. Coakley, 573 U.S. 464, 478 (2014)). A content-neutral
regulation, by contrast, triggers only “intermediate” scrutiny,
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 24 of 59
24 Opinion of the Court 25-11881
meaning (although the formulations vary) that it need only “advance[] important governmental interests unrelated to the suppression of free speech and . . . not burden substantially more speech
than necessary to further those interests.” TikTok, 604 U.S. at 70
(quoting Turner Broad. Sys., Inc. v. FCC, 520 U.S. 180, 189 (1997)
(Turner II)). For reasons we will explain, we hold that HB3 is a content-neutral speech regulation. Accordingly, the law is subject to
intermediate, rather than strict, scrutiny.
1
A law is content based if it “target[s] speech based on its communicative content.” TikTok, 604 U.S. at 70 (quoting Reed v. Town
of Gilbert, 576 U.S. 155, 163 (2015)). To be clear, a law can be content based either on its face or in its justification. A “law is content
based on its face if it ‘applies to particular speech because of the
topic discussed or the idea or message expressed.’” Id. (quoting
Reed, 576 U.S. at 163); see also Free Speech Coal., 606 U.S. at 504 (Kagan, J., dissenting) (observing that a law is content based if it “draws
distinctions based on the topic, subject matter, idea, or message expressed” (citation modified)). Alternatively, even a facially content-neutral law can be content-based if “it cannot be justified without reference to the content of the regulated speech or was adopted
by the government because of disagreement with the message the
speech conveys.” TikTok, 604 U.S. at 70–71 (quoting Reed, 576 U.S.
at 164 (citation modified)).
And the flip side? A law is facially content neutral if it doesn’t
“target particular speech based upon its content”—or, put another
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 25 of 59
25-11881 Opinion of the Court 25
way, doesn’t “impose a ‘restriction, penalty, or burden’ by reason
of content.” Id. at 71 (quoting Turner Broad. Sys., Inc. v. FCC, 512
U.S. 622, 644 (1994) (Turner I)). A good proxy for content neutrality: Can the speaker “‘avoid or mitigate’ the effects of the [law] by
altering” the substance of his or her expression? Id. (quoting Turner
I, 512 U.S. at 644). If so, the law is likely content based; if not, it’s
probably content neutral. A law has a content-neutral justification
if it “neither references the content of speech . . . nor reflects disagreement with the message such speech conveys.” Id. at 72. As is
probably evident, content-neutral laws “pose a less substantial risk
of excising certain ideas or viewpoints from the public dialogue”
than those that are content based. Id. at 70 (quoting Turner I, 512
U.S. at 642).
A pair of examples illustrate the distinction between content-based and content-neutral speech restrictions. In Brown v. Entertainment Merchants Ass’n, the Supreme Court held that a California law banning vendors from selling “violent video games” to minors was an impermissible content-based speech restriction. 564
U.S. 786, 789 (2011). “Violent” speech, the Court reasoned, is a
particular kind of speech—a particular kind of content—and California’s law singled it out for unfavorable treatment vis-à-vis
speech that didn’t depict violence. Id. at 799. Citing gory examples
from Grimm’s Fairy Tales, The Odyssey, Inferno, and Lord of the Flies,
the Court reasoned that violent content has historically been protected for minors. Id. at 795–96. So, regardless of form—whether
in print, television, or video games—California’s attempt to draw
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 26 of 59
26 Opinion of the Court 25-11881
a distinction between violent and nonviolent speech was impermissibly content based. Id. at 799. Even as to minors, who enjoy more
limited First Amendment protections, the government doesn’t possess “a free-floating power to restrict the ideas to which children
may be exposed.” Id. at 794 (emphasis added). Accordingly, the
Court concluded, speech that is protected for minors—including
violent speech—“cannot be suppressed solely to protect the young
from ideas or images that a legislative body thinks unsuitable for
them.” Id. at 795 (quoting Erznoznik v. City of Jacksonville, 422 U.S.
205, 213 (1975)).
In TikTok, by contrast, the Supreme Court held that a federal
statute that required a Chinese social-media company either to sell
its U.S.-based operations to U.S. owners or to cease operations in
this country was content neutral. TikTok, 604 U.S. at 66, 71. The
Court reasoned that the law was facially content neutral because
its prohibitions didn’t “target particular speech based upon its content,” “regulate speech based on its function or purpose,” or “impose a ‘restriction, penalty, or burden’ by reason of content” on the
company’s service. Id. at 71. Put simply, it didn’t “facially regulate
‘particular speech because of the topic discussed or the idea or message expressed.’” Id. (quoting Reed, 576 U.S. at 163). And, the
Court concluded, the statute’s justification was also content neutral: to prevent China, “a foreign adversary[],” “from collecting vast
amounts of sensitive data from” millions of American users. Id. at
71–72. That purpose, the Court said, was “unrelated to the content
of expression,” and it “neither reference[d] the content of speech
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 27 of 59
25-11881 Opinion of the Court 27
on TikTok nor reflect[ed] disagreement with the message such
speech convey[ed].” Id. at 72 (citing Ward, 491 U.S. at 792–93).
HB3 more closely resembles the law at issue in TikTok than
the one in Brown. HB3 is facially content neutral: It regulates social-media platforms based on their functional capabilities and features, not the content they communicate. The law neither references the substance of the speech on any platform nor draws distinctions based on it. Put differently, HB3 applies to platforms
wholly without regard to the topics discussed or the ideas or messages conveyed on them. Truth Social is just as likely as Instagram
to be affected by HB3 so long as (1) a proportion of minors under
16 spend an average of at least two hours a day on it; (2) it uses
infinite scrolling, push notifications, auto-play, or livestreaming, or
displays user-engagement metrics; and (3) it selects content algorithmically for users. Fla. Stat. § 501.1736(1)(e). And tellingly, neither a platform nor a user can “avoid or mitigate” the effects of HB3
by “altering” the substance of any posted expression. TikTok, 604
U.S. at 71 (quoting Turner I, 512 U.S. at 644).
HB3’s justification is also content neutral: to “protect[] children from [social-media] addiction and associated mental-health
harms.” Br. of Appellant at 36. That purpose is “unrelated to the
content of expression” on any given platform, and it doesn’t “reflect[] disagreement with the message such speech conveys.” TikTok, 604 U.S. at 72. There’s no indication, for instance, that Florida
enacted HB3 because it disagreed with social-media companies’ decisions to “deplatform” users who “advocate for terrorism,” Moody,
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 28 of 59
28 Opinion of the Court 25-11881
603 U.S. at 736–37, deprioritize posts on “conspiracy theories about
the origin of the [COVID] virus,” Murthy v. Missouri, 603 U.S. 43,
51 (2024), or fact-check “misleading claims about elections,” id. at
50. 8
By contrast, HB3 is easily distinguished from the violentvideo-games restriction at issue in Brown. That law was doomed
because it separated video games with violent content from video
games without it. HB3 makes no comparable distinction.
2
CCIA has counterarguments, of course, and they deserve
careful consideration.
First, CCIA contends that HB3 abridges its member platforms’ First Amendment rights by restricting “a website’s choices
about how to organize and present collections of third-party
speech,” which, as already noted, Moody held to be “protected expressive activity.” Br. of Appellees at 38 (citing Moody, 603 U.S. at
731–32). Maybe so, but the fact that a law regulates expressive activity (or more accurately here, “disproportionate[ly] burden[s]” it
in a way that we’ve found HB3 does, see supra at 22–23) merely puts
8 To be clear, HB3’s content neutrality is evident not only in its rules but also
in its exceptions. For instance, the term “social-media platform” is defined not
to include “an online service, website, or application where the exclusive function is e-mail or direct messaging consisting of text, photographs, pictures, images, or videos shared only between the sender and the recipients . . . .” Fla.
Stat. § 501.1736(1)(e). Whether HB3 applies to an email or messaging service
has nothing to do with the subject matter or topic of any correspondence.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 29 of 59
25-11881 Opinion of the Court 29
the First Amendment on the table—it doesn’t ipso facto render the
law unconstitutional. The next question in the decision tree—and
in practical effect, often the most important—is whether the law
regulates based on content. And on that score, HB3 bears scant
resemblance to the law at issue in Moody. The Supreme Court
treated that law as content based because it restricted social-media
platforms’ freedom to “delet[e], alter[], label[], or deprioritiz[e]”
certain posts—which is to say, it specifically hindered their ability
to curate and moderate their content as they saw fit. Moody, 603
U.S. at 720. Put simply, the law “interfere[d] with [platforms’] editorial choices . . . by ordering the excluded to be included.” Id. at
732. HB3 doesn’t restrict platforms’ content-moderation determinations or tools, but rather it regulates only front-end access and
website mechanics (e.g., so-called “addictive features”).
Second, CCIA compares HB3 to a hypothetical statute that
“restrict[s] all minors from accessing libraries just because some
spend (in the state’s view) too much time reading page-turning
novels,” or a law that limits minors “from using streaming services
because some spend too much time watching engaging cartoons.”
Br. of Appellees at 17. Surely, CCIA says, such restrictions would
be unconstitutional—and so, the argument goes, must be HB3. Id.
Perhaps the laws that CCIA imagines would be unconstitutional—
we needn’t say, and so we don’t. But if those laws did violate the
First Amendment, it wouldn’t be because they were content
based—but rather because, for whatever reason, they failed the
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 30 of 59
30 Opinion of the Court 25-11881
scrutiny applicable to content-neutral regulations, a topic we’ll address soon enough. To be clear, though, those laws would be content neutral.
Third, CCIA contends that HB3 is content based because it
applies only to so-called “social” speech: The law, CCIA asserts,
targets websites “based on the ‘social’ subject matter ‘of the material they disseminate.’” Br. of Appellees at 26 (quoting NetChoice v.
Reyes, 748 F. Supp. 3d 1105, 1122–23 (D. Utah 2024)) (citation modified). Dissenting from our order staying the preliminary injunction, Judge Rosenbaum similarly contended that HB3’s restrictions
are a “proxy for a content-based regulation directly prohibiting
‘public, interactive, user-to-user social speech.’” Comput. & Commc’ns
Indus. Ass’n v. Uthmeier, No. 25-11881, 2025 WL 3458571, at *15
(11th Cir. Nov. 25, 2025) (Rosenbaum, J., dissenting) (emphasis in
original). Respectfully, we disagree. The “social”-ness of speech is
a function of its form, not its substance. Put simply, “social”
speech, as CCIA seems to be using the term, is simply speech that
occurs between users on social-media platforms, wholly without
regard to “the topic discussed or the idea or message expressed.”
TikTok, 604 U.S. at 70. Users can engage in so-called “social” speech
about any topic, from the sublime (religion, philosophy, etc.) to the
ridiculous (celebrity gossip, cat videos, etc.).
Finally, CCIA asserts that HB3’s focus on the “addictive features” of social-media platforms, rather than, say, streaming services, shows that Florida is “plainly focused on particular content.”
Br. of Appellees at 16. But that’s not quite right, either. Regardless
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 31 of 59
25-11881 Opinion of the Court 31
of whether the content on a social-media platform is a clip from
The Bachelor or a political meme, the platform is subject to HB3. It
can’t “avoid or mitigate” the effects of HB3 “by altering [its]
speech.” TikTok, 604 U.S. at 71.
* * *
For the foregoing reasons, we conclude that while HB3 restricts First Amendment-protected activity, it likely does so in a
content-neutral manner.
C
A content-neutral law like HB3 is subject to intermediate
scrutiny, meaning that it is valid so long as it “advances important
governmental interests unrelated to the suppression of free speech
and does not burden substantially more speech than necessary to
further those interests.” TikTok, 604 U.S. at 70 (quoting Turner II,
520 U.S. at 189). A regulation “advances” the government interest
so long as it “would be achieved less effectively absent the regulation.” Free Speech Coal., 606 U.S. at 496 (quoting TikTok, 604 U.S. at
70).
A few intermediate-scrutiny corollaries: First, to satisfy intermediate scrutiny—unlike strict scrutiny—a speech regulation
need not be the “least restrictive means of serving the State’s interest.” Id. (quoting Ward, 491 U.S. at 798) (citation modified). Even
if other “approaches are equally or more effective, under intermediate scrutiny a ‘regulation will not be invalid simply because a
court concludes that the government’s interest could be adequately served by some less-speech-restrictive alternative.’” Id. at
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 32 of 59
32 Opinion of the Court 25-11881
497–98 (quoting Ward, 491 U.S. at 800). Second, “[t]he regulation’s
validity does not turn on our agreement with the legislature concerning the most appropriate method for promoting significant
government interests or the degree to which those interests should
be promoted.” Id. at 496 (citation modified) (quoting Ward, 491
U.S. at 800). Finally, intermediate scrutiny does not impose on
speech regulations a “‘freestanding underinclusiveness limitation,’
and the Government ‘need not address all aspects of a problem in
one fell swoop.’” TikTok, 604 U.S. at 76 (quoting Williams-Yulee v.
Fla. Bar, 575 U.S. 433, 449 (2015)).
Let’s evaluate HB3 against each of intermediate scrutiny’s
two core requirements—again, that it (1) advance an “important
governmental interest[] unrelated to the suppression of free
speech” and (2) “not burden substantially more speech than necessary” to further that interest. TikTok, 604 U.S. at 70.
1
HB3 easily satisfies the first requirement. CCIA doesn’t dispute that HB3 advances an important state interest; it acknowledges that “protecting minors is certainly a laudable goal.” Br. of
Appellees at 28. Even so, it says, HB3 fails intermediate scrutiny’s
first element because Florida’s interest “is plainly related to the suppression of expression.” Id. In particular, CCIA emphasizes that
HB3 prohibits addictive features only on websites, not on other “services that disseminate speech”—like, for instance, “books that end
in cliffhangers.” Id. at 28–29. Florida singles out social-media platUSCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 33 of 59
25-11881 Opinion of the Court 33
forms, CCIA insists, because they host speech that Florida “distrusts.” Id. at 17. CCIA’s argument doesn’t necessarily follow. It
would hold up only if websites have an inherent expressive value
that differs from the expressive value inherent in all books. But
both books and websites communicate content that spans the full
range of expressive values; they just do so in different formats. Absent a restriction on the substance of the topics, ideas, or messages
communicated by websites as compared to those communicated
by books, HB3’s “singling out” of websites is not, as CCIA argues,
“plainly related to the suppression of expression.” 9 Br. of Appellees
at 28.
2
HB3 also likely satisfies intermediate scrutiny’s second requirement—that it be appropriately tailored to advancing Florida’s
interest in protecting minors. See TikTok, 604 U.S. at 70. In partic9 CCIA’s argument is better understood as a critique of HB3’s tailoring—that
to the extent it seeks to protect minors from “addictive” speech features, HB3
is underinclusive inasmuch as it doesn’t also cover literary devices. But setting
aside the question whether literary devices are addictive, even if the state believed them to be an issue, it “need not address all aspects of a problem in one
fell swoop” because intermediate scrutiny imposes no “freestanding underinclusiveness limitation.” TikTok, 604 U.S. at 76 (quoting Williams-Yulee, 575
U.S. at 449). Further, the analogy to literary devices that make books more
engaging—like cliffhangers, shorter chapters, or serialization—simply reinforces Florida’s contention that it’s not focused on speech and the accompanying ideas, topics, and messages, but rather on features that in and of themselves
are agnostic as to content.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 34 of 59
34 Opinion of the Court 25-11881
ular, HB3 doesn’t “burden substantially more speech than necessary” because it (1) calibrates its restrictions by age, (2) regulates
minors’ creation of social-media accounts, not necessarily their access to the platforms themselves, (3) covers only those platforms
that have been determined to induce heavy use by children, and
(4) applies only to platforms that use both algorithms and one of
several addictive features. Id.; see Fla. Stat. § 501.1736(1)(e).
First, HB3 is age delimited. It prohibits minors under 14
from creating accounts, id. § 501.1736(2)(a), and it requires 14- and
15-year-olds to obtain parental consent before doing so, id.
§ 501.1736(3)(a). Notably, it imposes no limitations for those 16
years of age and older. The law thus applies only to younger minors and recognizes the fine-grained distinctions even among those
to whom it applies—namely, that 13-year-olds (in general, middleschoolers) are different from 14- and 15-year-olds (those starting
high school), who, in turn, are different from those 16 and up (budding adults).
Second, even among those to whom HB3 does apply, the
law restricts only their ability to create their own accounts, not
their ability to access social-media platforms per se. The creation
of a social-media account is a commercial act—a business transaction. Even a minor—indeed, even a young child—may continue
to use the platforms provided he either (1) uses his parents’ (or
someone else’s) account or (2) gains access without an account.
Third, HB3 targets only the platforms to which many minors are already addicted. As a reminder, a platform comes within
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 35 of 59
25-11881 Opinion of the Court 35
HB3’s purview only if at least 10% of daily active users under 16
years old spend an average of two hours or more per day on it. Fla.
Stat. § 501.1736(1)(e). Together with its other limitations, HB3’s
tailoring efforts fall within the broad range of potential legislative
solutions that intermediate scrutiny permits.
And fourth, HB3 operates not as an outright prohibition of
platform usage but rather as a “conditional ban” of sorts—namely,
on platforms’ use of addictive features and algorithms. See TikTok,
604 U.S. at 74, 77. The law applies only to platforms that use a
combination of algorithms and addictive features and that succeed
in keeping many minors on them for an average of at least two
hours a day. See Fla. Stat. § 501.1736(1)(e). Refraining from using
either algorithms or addictive features, or remaining below the
temporal threshold, exempts a platform from the law’s application.
Now, none of that is necessarily to say that HB3 embodies
the best policy choice. Maybe, maybe not. But intermediate scrutiny affords governments “latitude . . . to design regulatory solutions to address content-neutral interests.” TikTok, 604 U.S. at 77
(quoting Turner II, 520 U.S. at 213). Accordingly, it is within Florida’s discretion to draw the lines it has—many of which, of course,
narrow the law’s scope. So, for instance, Florida may validly opt to
prohibit minors under 14, rather than 16, from creating social-media accounts, and to require 16-year-olds, rather than all those under 18, to obtain their parents’ consent before creating accounts.
So too, it may validly train its attention on platforms that a segment
of children use more than two hours per day, rather than one. And
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 36 of 59
36 Opinion of the Court 25-11881
finally, it is free to focus on sites that use both algorithms and “addictive features,” rather than one or the other of those.
The numerous limitations on HB3’s sweep—that it applies
only to users under the age of 16, operates differently on 14- and
15-year-olds than on younger children, restricts only account creation and not access per se, and affects only those platforms using
features most likely to addict minors—differentiate it from the law
at issue in Packingham v. North Carolina, 582 U.S. 98 (2017), on
which CCIA relies. There, applying intermediate scrutiny, the Supreme Court invalidated a North Carolina statute making it a felony for any registered sex offender to access any “commercial social networking Web site” that allowed any minors to use it. Id. at
101. That law operated with “extraordinary breadth.” Id. at 109
(Alito, J., concurring in the judgment). Unlike HB3, which applies
only to a subset of minors, the statute there covered all registered
sex offenders. And unlike HB3, which applies only to a subset of
social-media platforms, the law there defined the covered sites exceedingly broadly; it included “commonplace social media websites like Facebook and Twitter,” as well as “websites as varied as
Amazon.com, Washingtonpost.com, and Webmd.com.” Id. at
101, 106 (majority opinion). And unlike HB3, which restricts account creation but doesn’t limit access to the platforms per se, the
statute there forbade any usage of the covered sites. Id. at 101. HB3
is tailored in a way that the law at issue in Packingham was not.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 37 of 59
25-11881 Opinion of the Court 37
CCIA complains that Florida failed to adopt social-media restrictions that would have restricted less speech. Citing the Supreme Court’s decision in an abortion-clinic-buffer-zone case,
McCullen v. Coakley, 573 U.S. 464 (2014), CCIA contends that the
state didn’t “demonstrate that alternative measures that burden
substantially less speech would fail to achieve the government’s interests.” Br. of Appellees at 32 (quoting McCullen, 573 U.S. at 495).
One such alternative, CCIA says, would have been to defer the decision of whether children can open social-media accounts entirely
to their parents, who “already have many tools to protect their minors on the Internet.” Id.
But again, as the Supreme Court recently reiterated in Free
Speech Coalition, under intermediate scrutiny—at least as conventionally applied, and as applied in Internet-related cases—a regulation needn’t be the “least restrictive means of serving the State’s
interest.” 606 U.S. at 496 (quoting Ward, 491 U.S. at 798) (citation
modified). Even if other “approaches are equally or more effective,” HB3 isn’t “‘invalid simply because a court concludes that the
government’s interest could be adequately served by some lessspeech-restrictive alternative.’” Id. at 497–98 (quoting Ward, 491
U.S. at 800). Instead, the legislature has some discretion to choose
“the most appropriate method for promoting” its interest in curbing minors’ social-media addiction or “the degree to which th[at]
interest[] should be promoted.” Id. at 496 (quoting Ward, 491 U.S.
at 800). Florida chose to prohibit the creation of social-media accounts for those under 14 because it found that relying exclusively
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 38 of 59
38 Opinion of the Court 25-11881
on parental supervision and consent was “ineffective.” Br. of Appellant at 8; see also States’ Amicus Br. at 7 (“Parental controls have
already proven insufficient to stem the tide of harm kids are suffering. And besides, kids inevitably find workarounds.”). 10
CCIA seeks to distinguish Free Speech Coalition on the ground
that the law at issue there was less speech-restrictive than HB3. In
that case, the Supreme Court upheld under intermediate scrutiny
a Texas statute requiring pornographic websites to verify users’
10 We acknowledge that McCullen could be read to impose something that
looks like a least-restrictive-means requirement, even in the context of intermediate scrutiny. There, in the course of invalidating a Massachusetts statute
that made it a crime to stand on a public street or sidewalk within 35 feet of
an abortion clinic, the Supreme Court said that, for intermediate-scrutiny purposes, the law burdened substantially more speech than necessary because the
state hadn’t “seriously undert[aken] to address the problem with less intrusive
tools readily available to it.” McCullen, 573 U.S. at 494. The Court faulted
Massachusetts for failing to use other tools at its disposal—such as existing
general criminal statutes prohibiting assault, harassment, or breach of the
peace—to “ensur[e] public safety outside abortion clinics” and “prevent[] harassment and intimidation of patients and clinic staff.” Id. at 490, 492; see also
Fla. Preborn Rescue, Inc. v. City of Clearwater, 161 F.4th 732, 745 (11th Cir. 2025)
(applying McCullen to hold that a city ordinance prohibiting people from standing on a public sidewalk five feet away from an abortion-clinic driveway burdened substantially more speech than necessary because the city could have
enforced the state’s existing anti-obstruction statute to achieve the same interest).
The simple response to McCullen—and its arguable imposition of a
least-restrictive-means-like requirement as part of an intermediate-scrutiny
analysis—is that Free Speech Coalition is both (1) more recent and (2) as an Internet-related case, more closely on point.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 39 of 59
25-11881 Opinion of the Court 39
ages. 606 U.S. at 495. In the course of so holding, the Court observed that the Texas law prohibited minors from accessing
speech—namely, pornography—that was unprotected as to them.
Id. CCIA contends that, by contrast, HB3 prohibits certain minors
from accessing protected speech. That may be so, but Free Speech
Coalition didn’t purport to establish a constitutional outer limit;
nothing in the Court’s reasoning there precludes a state from imposing an age-verification requirement that operates to limit minors’ access to some protected speech—provided, of course, that it
does so in a content-neutral manner.
To the contrary, everything the Court said in Free Speech Coalition seems to point in the other direction. The Court emphasized, for instance, that “[r]equiring age verification is common
when a law draws lines based on age.” Id. at 479. It noted that
many states require proof of age to buy alcohol and tobacco, as
well as to obtain licenses to carry a handgun, to drive, and to marry.
Id. And in a similar vein, it observed that brick-and-mortar stores
have long engaged in the “widespread” practice of verifying a purchaser’s age in person for “age-restricted services, such as online
gambling, alcohol and tobacco sales, and car rentals.” Id. at 479,
497. And, importantly, the Court stressed that in the online world,
“the need for age verification . . . is even greater” because “[u]nlike
a store clerk, a website operator cannot look at its visitors and estimate their ages.” Id. at 481.
* * *
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 40 of 59
40 Opinion of the Court 25-11881
For the foregoing reasons, we hold that HB3 is likely constitutional as applied to CCIA’s members. Yes, it implicates the First
Amendment and thus triggers constitutional scrutiny. But it likely
does so in a content-neutral manner, and it likely survives intermediate scrutiny because it doesn’t burden substantially more speech
than necessary to advance the important interest in protecting minors from social-media addiction.
IV
Last but not least, we must separately assess CCIA’s facial
challenge in light of the Supreme Court’s recent decision in Moody
v. NetChoice, LLC, 603 U.S. 707 (2024). As we understand it, CCIA’s
facial challenge takes two forms: what we’ll call a “traditional” facial challenge, in which it seeks to vindicate its members’ First
Amendment rights; and a facial “overbreadth” challenge, in which
it seeks to vindicate its members’ users’ rights. For reasons previously explained, we decline to examine the user-focused overbreadth challenge, as CCIA seems (at least for now) to have walked
away from it. See supra at 17 n.4. That leaves the more traditional,
member-focused facial challenge.
As the district court observed, CCIA raised both facial and
as-applied First Amendment challenges to HB3. And despite some
initial uncertainty, the parties seem to agree that the district court
considered and adjudicated both claims. See Supp. Br. of Appellant
at 20–21; Supp. Br. of Appellees at 19–20. For instance, in granting
CCIA’s preliminary-injunction motion, the district court held that
CCIA was “substantially likely to succeed not just in [its] as-applied
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 41 of 59
25-11881 Opinion of the Court 41
challenge but in [its] facial challenge as well.” Dist. Ct. Order, June 3,
2025, at 52–53 (emphasis added).
We hold that CCIA has not carried its facial-challenge “burden” under Moody. 603 U.S. at 744. As the Supreme Court warned
there, the decision to mount a facial challenge “comes at a cost.”
Id. at 723. Because “facial challenges threaten to short circuit the
democratic process,” they are, by design, “hard to win.” Id. (citation modified); see also id. at 745 (Barrett, J., concurring) (“[D]ealing
with a broad swath of varied platforms and functions in a facial
challenge strikes me as a daunting, if not impossible, task.”); id. at
748 (Jackson, J., concurring in part and concurring in the judgment)
(“[P]laintiffs bringing a facial challenge must clear a high bar.”); id.
at 778 (Alito, J., concurring in the judgment) (“[P]arties mounting
facial attacks [must] satisfy demanding requirements.”). Even in
First Amendment cases, where the facial-challenge bar is lower
than it otherwise might be, the plaintiff must show that “a substantial number of [the law’s] applications are unconstitutional, judged
in relation to the statute’s plainly legitimate sweep.” Id. at 723 (majority opinion) (quoting Ams. for Prosperity Found. v. Bonta, 594 U.S.
595, 615 (2021)).
To conduct the required Moody analysis, a reviewing court
proceeds in two steps. First, it must “assess the state laws’ scope,”
asking, “What activities, by what actors, do the laws prohibit or
otherwise regulate?” Id. at 724. Having done that, the court must
then “decide which of the laws’ applications violate the First
Amendment, and . . . measure them against the rest.” Id. at 725.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 42 of 59
42 Opinion of the Court 25-11881
To succeed on a facial challenge, a plaintiff must show that “the
law’s unconstitutional applications substantially outweigh its constitutional ones.” Id. at 724.
CCIA asserts that HB3 is invalid on its face because it “is unconstitutional in all its applications.” Br. of Appellees at 43 n.7 (emphasis added). Echoing the district court, CCIA reasons that HB3
“applies to any service where people upload content and view the
content of others” and thus “implicates the First Amendment in
every application except for a service that only has unconstitutional
speech.” Oral Arg. 28:44–28:56. And Florida, CCIA says, hasn’t
shown “that there’s any service that’s covered by this law that is all
about only unconstitutional speech.” Id. 28:57–29:03; accord Dist.
Ct. Order, June 3, 2025, at 27–28 (observing that HB3 covers only
social media platforms “on which protected speech . . . proliferate[s]” and that Florida hadn’t “hypothesize[d]” that the law could
cover sites whose “content is all or mostly unprotected speech”).
Respectfully, we don’t think that’s quite right. To be sure,
HB3 applies to websites that enable users to “‘upload content or
view the content or activity of other users’—in other words, engage in speech.” Dist. Ct. Order, June 3, 2025, at 43 n.25 (citing Fla.
Stat. § 501.1736(1)(e)). But those features—uploading and viewing
content—aren’t necessarily limited to websites that host only fully
protected speech. Put differently, just because users are “engag[ing] in speech” on a platform doesn’t mean that their expression is fully protected by the First Amendment. Id. There are all
kinds of social-media platforms featuring—or even if not featuring,
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 43 of 59
25-11881 Opinion of the Court 43
permitting—all kinds of expression. To be sure, much of that expression is fully protected. But some isn’t. The Supreme Court has
held, for instance, that certain kinds of speech enjoy no First
Amendment protection at all—e.g., incitement, true threats. See
Brandenburg v. Ohio, 395 U.S. 444 (1969) (incitement); Counterman v.
Colorado, 600 U.S. 66 (2023) (true threats). Other kinds of speech
enjoy constitutional protection when exhibited to adults but not
when exhibited to children—pornography being the most obvious
example. See Free Speech Coal., 606 U.S. at 496.
It may well be true that most expression on most social-media
platforms is indeed protected. But not all of it is, at least not with
respect to all users. So, for instance, imagine a platform that hosts
only unprotected speech—say, Incitement.com or TrueThreats.com. A law restricting children’s (or anyone’s) access to it
wouldn’t implicate the First Amendment at all and would trigger
only rational-basis review. A law restricting access to a site that
hosts speech that is protected for adults but not children—like pornography—would implicate the First Amendment, but a contentneutral restriction—like an age-verification requirement—would
be subject only to intermediate scrutiny. See Free Speech Coal., 606
U.S. at 495. Accordingly, even if HB3 “imposes the same onerous
restriction” on every platform, Dist. Ct. Order, June 3, 2025, at 43,
that restriction could well be subject to different levels of scrutiny,
potentially leading to different outcomes.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 44 of 59
44 Opinion of the Court 25-11881
It’s no answer to say, as both CCIA and the district court
have, that Florida has failed to demonstrate constitutional applications of HB3. See Oral Arg. 24:20–24:32 (“Under the First Amendment, it’s the state’s burden to prove that the law has more constitutional applications than unconstitutional applications.”); Dist. Ct.
Order, June 3, 2025, at 28. In a facial-challenge posture, we don’t
think it’s the state’s obligation to do so. Rather, the Supreme Court
seems to have put the “burden” squarely on the plaintiff: As one
“price” of its “decision to challenge the law[] as a whole,” the challenger must “carry its burden” of demonstrating that the invalid
applications substantially outweigh the valid ones. Moody, 603 U.S.
at 744.
One might interpose a “come on” objection at this point:
Surely, the argument would go, the overwhelming majority of
platforms to which HB3 applies overwhelmingly host protected
speech. And if, as CCIA insists, the law’s application to those websites would be unconstitutional, then the law’s unconstitutional applications would probably substantially outweigh the constitutional ones. But however intuitive, that sort of impressionistic
judgment won’t satisfy Moody’s stringent demands. Whether or
not it’s likely that the collective content of the platforms within
HB3’s ambit is overwhelmingly protected, CCIA bears the “burden” of demonstrating both (1) the universe of the law’s applications and (2) that the invalid applications substantially outweigh
the valid ones. See id.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 45 of 59
25-11881 Opinion of the Court 45
To discharge its facial-challenge burden, CCIA must provide
the reviewing court with the data—either hard facts or, at the very
least, sufficient briefing—necessary to conduct the Moody-required
comparative analysis. It hasn’t done so here. Indeed, CCIA’s briefing doesn’t mention even a single platform that isn’t one of its
members—let alone any nonmember platform that (in whole or in
part) hosts unprotected speech. In that respect, we think, CCIA
makes the same mistake that it and NetChoice made in Moody:
treating the case “more like [an] as-applied claim[] than like [a] facial one[].” Moody, 603 U.S. at 724; cf. NetChoice, LLC v. Bonta, 152
F.4th 1002, 1020 (9th Cir. 2025) (“[F]acial challengers cannot focus
on a law’s effect on their own platforms—they cannot even limit
the inquiry to the tech giants.”); NetChoice, LLC v. Yost, 180 F.4th
268, 312 (6th Cir. 2026) (Batchelder, J., concurring in the judgment)
(refraining from addressing NetChoice’s facial challenge “without
a developed record”).
We’ve learned the facial-challenge lesson the hard way. In
Moody, the Supreme Court, even while largely agreeing with our
merits analysis, faulted us for “confining [our] analysis” to “certain
heartland applications” instead of “address[ing] the full range of activities the law[] cover[ed].” Moody, 603 U.S. at 724; see NetChoice,
LLC v. Att’y Gen., Fla., 34 F.4th 1196 (11th Cir. 2022), abrogated by
Moody, 603 U.S. at 724. The Court went on to observe that even if
it were to attempt the facial analysis itself, it couldn’t “undertake
the needed inquiries” because the “parties ha[d] not briefed the critical issues [t]here, and the record [wa]s underdeveloped.” 603 U.S.
at 726. The Court therefore vacated and remanded our case (and
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 46 of 59
46 Opinion of the Court 25-11881
its companion, NetChoice, L.L.C. v. Paxton, 49 F.4th 439 (5th Cir.
2022), abrogated by Moody, 603 U.S. at 724), “to consider the scope
of the laws’ applications, and weigh the unconstitutional as against
the constitutional ones.” Id. at 726. In turn, both we and our colleagues on the Fifth Circuit remanded our respective cases to the
district courts for further development of the issues pertinent to the
facial-challenge analysis. See Order, NetChoice, LLC v. Att’y Gen.,
State of Fla., No. 21-12355, (11th Cir. Aug. 16, 2024), Dkt. No. 178;
NetChoice, L.L.C. v. Paxton, 121 F.4th 494, 500 (5th Cir. 2024) (remanding to the district court to “develop a factual record” in response to Moody).
We are similarly stymied here. For the foregoing reasons,
we conclude that CCIA has failed to satisfy Moody because it hasn’t
adequately addressed the “full range” of HB3’s applications, nor
does any part of the record bear on the constitutionality of those
applications.
V
To recap: We hold (1) that CCIA has associational standing
to assert the First Amendment rights of its social-media platform
members; (2) that the district court didn’t abuse its discretion in
declining to abstain under Younger; and (3) that although HB3 implicates First Amendment interests, it likely passes constitutional
muster, as it is a content-neutral speech restriction that satisfies intermediate scrutiny. Finally, (4) we reject the facial challenge to
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 47 of 59
25-11881 Opinion of the Court 47
HB3 on the ground that CCIA hasn’t carried its burden of demonstrating, as required by Moody, that the law’s unconstitutional applications substantially outweigh its valid applications.
Accordingly, we AFFIRM the district court’s determinations
that CCIA has associational standing and its decision not to abstain
under Younger, REVERSE the preliminary injunction with respect
to CCIA’s First Amendment claim as applied to its members,
VACATE the preliminary injunction with respect to CCIA’s facial
First Amendment challenge, and REMAND to the district court for
proceedings consistent with this opinion.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 48 of 59
NEWSOM, Circuit Judge, concurring:
Needless to say, I join the Court’s opinion in full. I write
separately to flag an issue that implicates CCIA’s as-applied challenge and that arises from (of all places) the Supreme Court’s recent
decision in West Virginia v. B.P.J. ex rel. Jackson, 146 S. Ct. 2356
(2026). There, in the course of holding that states may, consistently
with the Equal Protection Clause, limit participation in women’s
and girls’ sports to biological females, the Court sort of said—well
no, expressly said, in so many words—that “as-applied challenges
are not entertained when courts apply intermediate scrutiny or rational basis review.” Id. at 2377 n.5. Now, to be sure, for that proposition, the Court cited another equal-protection case. See id. (citing Tuan Anh Nguyen v. INS, 533 U.S. 53, 70 (2001)). But just as
surely, the Court emphasized that “the intermediate scrutiny
test . . . applies in a variety of contexts” and, for the stock formulation of the standard, it cited a prominent First Amendment decision. See id. at 2376 (citing Ward v. Rock Against Racism, 491 U.S.
781, 801 (1989)).
So, the question: Is B.P.J. properly read to foreclose as-applied challenges for all claims—including First Amendment
claims—that trigger intermediate scrutiny? The answer doesn’t so
much matter here—because, for reasons explained in the main
opinion, we’ve concluded that CCIA’s as-applied challenge fails
“on the merits,” so to speak, on the ground that HB3 survives a
proper intermediate-scrutiny analysis. See Maj. Op. at 31–40. But
if B.P.J. were read for all it might be worth, perhaps we could have
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 49 of 59
2 Opinion of the Court 25-11881
gotten away with a much shorter opinion. And in any event, the
answer may very well matter in the next First Amendment case, in
which an as-applied plaintiff could otherwise show that a law
would fail intermediate scrutiny.
For reasons I’ll explain, I don’t think we (yet) have a clear
(enough) indication from the Supreme Court that it actually meant
in B.P.J. to cut off all as-applied challenges to laws reviewable under
an intermediate-scrutiny framework. But I suppose it might have.
In any event, I thought the question worth noting.
I
I should begin by briefly unpacking B.P.J. Again, the Supreme Court held there that a state doesn’t “violate[] the Equal
Protection Clause by maintaining female sports teams for biological females.” 146 S. Ct. at 2373. The Court began by articulating
its topline premises and conclusion: (1) Laws that segregate sports
teams based on sex are “sex-based classification[s] that trigger[] intermediate scrutiny”; (2) the sex-based classification that such laws
draw is “substantially related” to the states’ “important” interests
in “safety and competitive fairness”; and therefore, (3) “schools
may determine eligibility for women’s and girls’ sports based on
biological sex.” Id. at 2374–75. Straightforward enough.
The Court then shifted gears to address the plaintiffs’ “narrower” argument—namely, that even if the states’ interests in
safety and competitive fairness are “sufficient reasons to exclude
most biological males,” they “do not justify excluding [the] relatively small subclass of biological males . . . who identify as female
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 50 of 59
25-11881 Opinion of the Court 3
and have taken puberty blockers or hormones.” Id. at 2375 (second
emphasis in original). The Court observed that the plaintiffs “characterize[d] their equal protection argument” in several “overlapping ways,” two of which are relevant for our purposes. Id. First,
the plaintiffs asserted “a lack of substantial relationship between
the laws’ biological-sex classification” and the states’ asserted interests. Id. And second, the plaintiffs’ framed their attack “as an asapplied challenge to the biological-sex classification.” Id. The
Court rejected both approaches.
With respect to the plaintiffs’ first tack, the Court emphasized that its “equal protection precedents allow general [sexbased] classifications like those” in the laws before it “so long as
there is at least a substantial relationship between the classification
and the State’s interests.” Id. at 2376. Reiterating its earlier conclusion that “‘[p]erfection in making the necessary classifications is
neither possible nor necessary,’” id. (quoting Mass. Bd. of Ret. v.
Murgia, 427 U.S. 307, 314 (1976)), the Court observed that “none of
[its] ‘gender-based classification equal protection cases have required that the statue under consideration must be capable of
achieving its ultimate objective in every instance,’” id. (quoting
Nguyen, 533 U.S. at 70). In further support of its explanation, the
Court looked beyond the equal-protection context to First Amendment cases, noting that “under the intermediate scrutiny test
that . . . applies in a variety of contexts,” the “‘validity of the regulation depends on the relation it bears to the overall problem the
government seeks to correct, not on the extent to which it furthers
the government’s interests in an individual case.’” Id. (quoting
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 51 of 59
4 Opinion of the Court 25-11881
Ward, 491 U.S. at 801); see also id. (citing United States v. Edge Broadcasting Co., 509 U.S. 418, 430–31 (1993), for the proposition that the
intermediate-scrutiny analysis is focused on the relationship between the restriction and “the general problem” rather than “an
individual case”). Accordingly, the Court held that “States are not
required to conduct an individual-by-individual comparison of the
physical and athletic capabilities of all biological males in order to
satisfy intermediate scrutiny.” Id. at 2377.
The Supreme Court also rebuffed the plaintiffs’ second
framing of their position—namely, “as an as-applied equal protection argument.” Id. The Court found this second argument to be
“basically the same as the[ir] first” one. Id. As the Court described
the theory, the plaintiffs contended that even if the states’ sex-based
classification was “generally permissible,” it violated the Equal Protection Clause “as applied to those biological males such as [the
plaintiffs,] who identify as female and have taken puberty blockers
or hormones.” Id. The Court held, to the contrary, that “as long
as the relationship” between a state’s interest and the classification
“is sufficient as a general matter, the State is not constitutionally required to grant individualized exemptions to specific athletes or
subclasses.” Id. It was at this point in its analysis that the Court
dropped footnote five, in which it said the thing: Distinguishing its
earlier decisions in Caban v. Mohammed, 441 U.S. 380 (1979), and
Lehr v. Robertson, 463 U.S. 248 (1983), the Court emphasized that a
more recent decision had “clarified that as-applied challenges are not
entertained when courts apply intermediate scrutiny or rational basis reUSCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 52 of 59
25-11881 Opinion of the Court 5
view.” B.P.J., 146 S. Ct. at 2377 n.5 (emphasis added) (citing Nguyen, 533 U.S. at 70). Back up above the line, the Court then rejected
“[t]he plaintiffs’ as-applied argument [as] in essence a backdoor argument for strict scrutiny, which requires a much tighter relationship between the State’s classification and the State’s asserted interests.” Id. at 2378.
II
So, what to make of B.P.J.—and in particular the (potentially) portentous footnote five? Did the Court there really mean to
foreclose all as-applied challenges to laws that (like HB3) trigger intermediate scrutiny, including those brought under the First
Amendment? Candidly, I’m not quite sure. I can see arguments
going both ways.
A
Here’s the case, as I see it, for what I’ll call the “broad” reading. First, and (by far) most significant, is the starkness of the
Court’s chosen language. Once again, as relevant here: “[A]s-applied challenges are not entertained when courts apply intermediate scrutiny.” Id. at 2377 n.5. Second, and in a similar vein, the
Court referred to the intermediate-scrutiny standard generally, emphasizing that it “applies in a variety of contexts.” Id. Third, as
already noted, when explaining what it took to be the proper application of that “test,” the Court looked, not to an equal-protection
case, but rather to a pair of First Amendment decisions: “[T]he ‘validity of the regulation depends on the relation it bears to the overall problem the government seeks to correct, not on the extent to
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 53 of 59
6 Opinion of the Court 25-11881
which it furthers the government’s interests in an individual case.’”
Id. at 2376 (quoting Ward, 491 U.S. at 801, and citing Edge Broadcasting, 509 U.S. at 430–31).
Finally, and perhaps most fundamentally, the logic underlying the B.P.J. Court’s ultimate rejection of the plaintiffs’ contention
would seem to apply just as surely to First Amendment cases as to
equal-protection cases. The “as-applied argument” that the plaintiffs there were making, the Court warned, was “in essence a backdoor argument for strict scrutiny, which requires a much tighter
relationship between the State’s classification and the State’s asserted interest.” Id. at 2378; see also Adams v. Sch. Bd. of St. Johns
Cnty., 3 F.4th 1299, 1329 (11th Cir. 2021) (Pryor, C.J., dissenting)
(requiring the government to prove tailoring “as to each plaintiff
who brings an as-applied challenge would disregard intermediate
scrutiny by demanding a perfect fit between the sex-based classification and the government interest at issue.”), rev’d en banc, 57
F.4th 791 (11th Cir. 2022). That critique is at least arguably generalizable across contexts because, by their very nature, as-applied
challenges exist in some tension with less rigorous forms of scrutiny. On the one hand, plaintiff-by-plaintiff analysis is of course the
very sine qua non of as-applied litigation. On the other, intermediate scrutiny (to say nothing of rational-basis review) “affords the
Government latitude in designing a regulatory solution,” Turner
Broad. Sys., Inc. v. FCC, 520 U.S. 180, 213 (1997), and, famously, does
not require a perfect fit between ends and means, see Free Speech
Coal. v. Paxton, 606 U.S. 461, 471 (2025). But—and this is the
point—that “latitude” will essentially evanesce if the government
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 54 of 59
25-11881 Opinion of the Court 7
has to prove that its interests apply to each and every successive
individual who happens to file suit.
It’s perhaps particularly easy to see that dynamic at work in
a context like sex-segregated sports, where an insistence that the
government repeatedly justify wholesale distinctions case by case
at the retail level would risk swallowing those distinctions whole.
Though biological men are on average more athletically capable
than women, “[n]ot every biological male athlete is bigger, stronger,
faster, or otherwise more athletically able than every biological female athlete.” B.P.J., 146 S. Ct. at 2376 (emphases in original). But
if every less capable male athlete could seek an as-applied exemption from general rules requiring sex-segregated sports, the distinction between men’s and women’s sports would die a death of a
thousand cuts. The same phenomenon, though, may well exist in
the First Amendment context. Imagine, for instance, a contentneutral ordinance prohibiting large late-night gatherings in a suburban park. On its face, the restriction would probably survive intermediate scrutiny as a sufficiently narrowly tailored means of
preventing after-dark disruptions. Cf. Ward, 491 U.S. at 796. But
as-applied challenges introduce a wrinkle: Must the city repeatedly
prove—in every successive case—that each prospective gathering
is likely to be so loud that it would disturb neighborhood peace? If
it turns out that the city can’t always do so, then the as-applied exemptions will make Swiss cheese out of an otherwise facially valid
noise-prevention ordinance.
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 55 of 59
8 Opinion of the Court 25-11881
Bottom line: It may just be that if “the intermediate scrutiny
test that the Court applies in a variety of contexts”—equal protection, free-speech, etc.—is indeed focused on the relationship between the law at issue and “‘the overall problem the government
seeks to correct, not on the extent to which it furthers the government’s interests in an individual case,’” B.P.J., 146 S. Ct. at 2376
(quoting Ward, 491 U.S. at 801), then it’s a test built largely—if not
entirely—for facial challenges.
B
There is, though, another side to the story. Here’s the case,
as I see it, for the “narrow” reading of B.P.J. First, and most obviously, there’s the fact that the Court was deciding an equal-protection case. And at times, it seemed to frame its analysis specifically
in those terms. See, e.g., id. at 2376 (“[T]his Court’s equal protection precedents allow general classifications like those made in the
[laws before us] so long as there is at least a substantial relationship
between the classification and the State’s interests.”); id. (“[N]one
of this Court’s ‘gender-based classification equal protection cases
have required that the statute under consideration must be capable
of achieving its ultimate objective in every instance.’”); id. at 2377
(“Intermediate scrutiny permits a sex-based classification that, as
here, is ‘not invidious, but rather realistically reflects the fact that
the sexes are not similarly situated in certain circumstances.’”).1
1 Indeed, there are suggestions in the Court’s opinion that it might have meant
to focus its analysis even more narrowly on equal-protection cases that arise in
the sports context. See, e.g., B.P.J., 146 S. Ct. at 2376 (“the sports context USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 56 of 59
25-11881 Opinion of the Court 9
Second, although the Court cited its First Amendment decisions in Ward and Edge Broadcasting for support, it did so only in
addressing the first of the plaintiffs’ two “narrower” arguments—
namely, to say that the plaintiffs’ just misunderstood the level of
generality at which a proper intermediate-scrutiny analysis should
be conducted. See id. at 2376. The Court didn’t (at least explicitly)
circle back to those cases in the course of rejecting what it called
the plaintiffs’ second, “as-applied” argument. See id. at 2377.
One final reason for circumspection: Recall that, in support
of what is for our purposes the key statement—that “as-applied
challenges are not entertained when courts apply intermediate
scrutiny,” the Court cited its earlier decision in Nguyen. Id. at 2377
n.5. Not only is Nguyen an equal-protection case, but I’m not sure
I read it to say everything that the Court in B.P.J. seemed to attribute
to it. I certainly don’t find there anything that purports to expressly
foreclose as-applied challenges for all claims that trigger intermediate scrutiny.
III
So, what here? This case, of course, presents an as-applied
challenge to a content-neutral speech regulation that triggers intermediate scrutiny. See Maj. Op. at 31. Accordingly, if B.P.J.’s rule is
indeed generalizable—if, in fact, “as-applied challenges are not entertained when courts apply intermediate scrutiny or rational basis
is . . . crucial”); id. at 2377 (“the distinctive sports context”); id. at 2378
(“[e]specially in the sports context”); id. (“[p]articularly in the sports context”). USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 57 of 59
10 Opinion of the Court 25-11881
review,” full stop, 146 S. Ct. at 2356 n.5—then CCIA’s claim failed
the moment that we concluded that intermediate scrutiny applied.
I’m reluctant to go there—at least yet.
First, there’s precedent and consistent practice—a lot of it.
We have long (and seemingly without qualm) entertained as-applied First Amendment challenges to laws that trigger intermediate
scrutiny. See, e.g., United States v. Romain, --- F.4th ----, 2026 WL
2250001, at *4 (11th Cir. Aug. 5, 2026); Moms for Liberty v. Brevard
Pub. Schs., 118 F.4th 1324, 1336–37 (11th Cir. 2024); Fort Lauderdale
Food Not Bombs v. City of Fort Lauderdale, 11 F.4th 1266, 1291–92
(11th Cir. 2021); cf. Ranch House, Inc. v. Amerson, 238 F.3d 1273,
1284–85 (11th Cir. 2001) (remanding to the district court to consider an as-applied First Amendment challenge under intermediate
scrutiny). As have our sister circuits, in cases too numerous to
count. See, e.g., Green v. U.S. Dep’t of Just., 54 F.4th 738, 745–46
(D.C. Cir. 2024); Brokamp v. James, 66 F.4th 374, 397–403 (2d Cir.
2023); 360 Virtual Drone Servs. LLC v. Ritter, 102 F.4th 263, 278–80
(4th Cir. 2024); Hines v. Pardue, 117 F.4th 769, 779–85 (5th Cir.
2024); People for the Ethical Treatment of Animals, Inc. v. Reynolds, 173
F.4th 959, 966–69 (8th Cir. 2026); Project Veritas v. Schmidt, 125 F.4th
929, 952–60 (9th Cir. 2025) (en banc). I’m hesitant to conclude that
the B.P.J. Court abrogated all of that law and convention without
a clearer indication that it really meant to extend its reasoning to
the First Amendment context. And I’d be doubly hesitant given
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 58 of 59
25-11881 Opinion of the Court 11
that, just last year, in TikTok Inc. v. Garland, the Supreme Court itself entertained an as-applied First Amendment challenge while applying intermediate scrutiny. See 604 U.S. 56, 74–80 (2025).
Second, there’s Moody v. NetChoice, LLC, 603 U.S. 707 (2024).
That decision, too, it seems to me, stands in some tension with a
prohibition on as-applied First Amendment challenges to laws that
trigger intermediate scrutiny. The Supreme Court was crystal
clear there: As-applied challenges are the rule, and facial challenges
the rare, “hard to win” exception. Id. at 723. And for good reason.
Facial challenges “‘threaten to short circuit the democratic process’
by preventing duly enacted laws from being implemented in constitutional ways.” Id. (quoting Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 451 (2008)). But if the Supreme Court’s
point in Moody was to make as-applied challenges the new normal
in First Amendment litigation, it seems strange that the same
Court—the same nine people—would have then turned around
less than two years later and nixed as-applied challenges for the
whole range of First Amendment claims that call for intermediate
scrutiny.2 And I don’t think it makes me a result-oriented hack to
worry that a combination of Moody’s new (or newly emphasized)
2 See, e.g., Ward, 491 U.S. at 791 (content-neutral, time, place, or manner restrictions); United States v. O’Brien, 391 U.S. 367, 376 (1968) (expressive conduct); Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 562–
65 (1980) (commercial speech); City of Renton v. Playtime Theatres, Inc., 475 U.S.
41, 47–49 (1986) (regulation of secondary effects); Turner Broad., 520 U.S. at
186 (content-neutral regulation of the airwaves); Paxton, 606 U.S. at 478 (regulations that only incidentally burden protected speech).
USCA11 Case: 25-11881 Document: 94-1 Date Filed: 10/14/2026 Page: 59 of 59
12 Opinion of the Court 25-11881
limitations on facial First Amendment challenges and a prohibition
on as-applied challenges in a host of contexts would lead to, well, a
lot less protection for traditionally safeguarded speech.
* * *
In the end, I’ll confess that I’m not quite sure what to make
of B.P.J.’s statement casting doubt on the viability of as-applied
challenges to laws that trigger less than strict scrutiny. But before
exporting it to the First Amendment context, I think it’s wise to
await a clearer signal from the Supreme Court.