������ýapp�ٷ�

NOTICE: This opinion is subject to formal revision before publication in the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, pio@supremecourt.gov, of any typographical or other formal errors.

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 22–277 and 22–555

_________________

ASHLEY MOODY, ATTORNEY GENERAL OF FLORIDA, et al., PETITIONERS (22–277)

v.

NETCHOICE, LLC, dba NETCHOICE, et al.

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE ELEVENTH CIRCUIT

NETCHOICE, LLC, dba NETCHOICE, et al., PETITIONERS (22–555)

v.

KEN PAXTON, ATTORNEY GENERAL OF TEXAS

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT

[July 1, 2024]

Justice Kagan delivered the opinion of the Court.[1]*

Not even thirty years ago, this Court felt the need to explain to the opinion-reading public that the “Internet is an international network of interconnected computers.” Reno v. American Civil Liberties Union, 521 U.S. 844, 849 (1997). Things have changed since then. At the time, only 40 million people used the internet. See id., at 850. Today, Facebook and YouTube alone have over two billion users each. See App. in No. 22–555, p. 67a. And the public likely no longer needs this Court to define the internet.

These years have brought a dizzying transformation in how people communicate, and with it a raft of public policy issues. Social-media platforms, as well as other websites, have gone from unheard-of to inescapable. They structure how we relate to family and friends, as well as to businesses, civic organizations, and governments. The novel services they offer make our lives better, and make them worse — create unparalleled opportunities and unprecedented dangers. The questions of whether, when, and how to regulate online entities, and in particular the social-media giants, are understandably on the front-burner of many legislatures and agencies. And those government actors will generally be better positioned than courts to respond to the emerging challenges social-media entities pose.

But courts still have a necessary role in protecting those entities’ rights of speech, as courts have historically protected traditional media’s rights. To the extent that social-media platforms create expressive products, they receive the First Amendment’s protection. And although these cases are here in a preliminary posture, the current record suggests that some platforms, in at least some functions, are indeed engaged in expression. In constructing certain feeds, those platforms make choices about what third-party speech to display and how to display it. They include and exclude, organize and prioritize — and in making millions of those decisions each day, produce their own distinctive compilations of expression. And while much about social media is new, the essence of that project is something this Court has seen before. Traditional publishers and editors also select and shape other parties’ expression into their own curated speech products. And we have repeatedly held that laws curtailing their editorial choices must meet the First Amendment’s requirements. The principle does not change because the curated compilation has gone from the physical to the virtual world. In the latter, as in the former, government efforts to alter an edited compilation of third-party expression are subject to judicial review for compliance with the First Amendment.

Today, we consider whether two state laws regulating social-media platforms and other websites facially violate the First Amendment. The laws, from Florida and Texas, restrict the ability of social-media platforms to control whether and how third-party posts are presented to other users. Or otherwise put, the laws limit the platforms’ capacity to engage in content moderation — to filter, prioritize, and label the varied messages, videos, and other content their users wish to post. In addition, though far less addressed in this Court, the laws require a platform to provide an individualized explanation to a user if it removes or alters her posts. NetChoice, an internet trade association, challenged both laws on their face — as a whole, rather than as to particular applications. The cases come to us at an early stage, on review of preliminary injunctions. The Court of Appeals for the Eleventh Circuit upheld such an injunction, finding that the Florida law was not likely to survive First Amendment review. The Court of Appeals for the Fifth Circuit reversed a similar injunction, primarily reasoning that the Texas law does not regulate any speech and so does not implicate the First Amendment.

Today, we vacate both decisions for reasons separate from the First Amendment merits, because neither Court of Appeals properly considered the facial nature of NetChoice’s challenge. The courts mainly addressed what the parties had focused on. And the parties mainly argued these cases as if the laws applied only to the curated feeds offered by the largest and most paradigmatic social-media platforms — as if, say, each case presented an as-applied challenge brought by Facebook protesting its loss of control over the content of its News Feed. But argument in this Court revealed that the laws might apply to, and differently affect, other kinds of websites and apps. In a facial challenge, that could well matter, even when the challenge is brought under the First Amendment. As explained below, the question in such a case is whether a law’s unconstitutional applications are substantial compared to its constitutional ones. To make that judgment, a court must determine a law’s full set of applications, evaluate which are constitutional and which are not, and compare the one to the other. Neither court performed that necessary inquiry.

To do that right, of course, a court must understand what kind of government actions the First Amendment prohibits. We therefore set out the relevant constitutional principles, and explain how one of the Courts of Appeals failed to follow them. Contrary to what the Fifth Circuit thought, the current record indicates that the Texas law does regulate speech when applied in the way the parties focused on below — when applied, that is, to prevent Facebook (or YouTube) from using its content-moderation standards to remove, alter, organize, prioritize, or disclaim posts in its News Feed (or homepage). The law then prevents exactly the kind of editorial judgments this Court has previously held to receive First Amendment protection. It prevents a platform from compiling the third-party speech it wants in the way it wants, and thus from offering the expressive product that most reflects its own views and priorities. Still more, the law — again, in that specific application — is unlikely to withstand First Amendment scrutiny. Texas has thus far justified the law as necessary to balance the mix of speech on Facebook’s News Feed and similar platforms; and the record reflects that Texas officials passed it because they thought those feeds skewed against politically conservative voices. But this Court has many times held, in many contexts, that it is no job for government to decide what counts as the right balance of private expression — to “un-bias” what it thinks biased, rather than to leave such judgments to speakers and their audiences. That principle works for social-media platforms as it does for others.

In sum, there is much work to do below on both these cases, given the facial nature of NetChoice’s challenges. But that work must be done consistent with the First Amendment, which does not go on leave when social media are involved.

I

As commonly understood, the term “social media platforms” typically refers to websites and mobile apps that allow users to upload content — messages, pictures, videos, and so on — to share with others. Those viewing the content can then react to it, comment on it, or share it themselves. The biggest social-media companies — entities like Facebook and YouTube — host a staggering amount of content. Facebook users, for example, share more than 100 billion messages every day. See App. in No. 22–555, at 67a. And YouTube sees more than 500 hours of video uploaded every minute. See ibid.

In the face of that deluge, the major platforms cull and organize uploaded posts in a variety of ways. A user does not see everything — even everything from the people she follows — in reverse-chronological order. The platforms will have removed some content entirely; ranked or otherwise prioritized what remains; and sometimes added warnings or labels. Of particular relevance here, Facebook and YouTube make some of those decisions in conformity with content-moderation policies they call Community Standards and Community Guidelines. Those rules list the subjects or messages the platform prohibits or discourages — say, pornography, hate speech, or misinformation on select topics. The rules thus lead Facebook and YouTube to remove, disfavor, or label various posts based on their content.

In 2021, Florida and Texas enacted statutes regulating internet platforms, including the large social-media companies just mentioned. The States’ laws differ in the entities they cover and the activities they limit. But both contain content-moderation provisions, restricting covered platforms’ choices about whether and how to display user- generated content to the public. And both include individualized-explanation provisions, requiring platforms to give reasons for particular content-moderation choices.

Florida’s law regulates “social media platforms,” as defined expansively, that have annual gross revenue of over $100 million or more than 100 million monthly active users. Fla. Stat. §501.2041(1)(g) (2023).[2] The statute restricts varied ways of “censor[ing]” or otherwise disfavoring posts — including deleting, altering, labeling, or deprioritizing them — based on their content or source. §501.2041(1)(b). For example, the law prohibits a platform from taking those actions against “a journalistic enterprise based on the content of its publication or broadcast.” §501.2041(2)(j). Similarly, the law prevents deprioritizing posts by or about political candidates. See §501.2041(2)(h). And the law requires platforms to apply their content-moderation practices to users “in a consistent manner.” §501.2041(2)(b).

In addition, the Florida law mandates that a platform provide an explanation to a user any time it removes or alters any of her posts. See §501.2041(2)(d)(1). The requisite notice must be delivered within seven days, and contain both a “thorough rationale” for the action and an account of how the platform became aware of the targeted material. §501.2041(3).

The Texas law regulates any social-media platform, having over 50 million monthly active users, that allows its users “to communicate with other users for the primary purpose of posting information, comments, messages, or images.” Tex. Bus. & Com. Code Ann. §§120.001(1), 120.002(b) (West Cum. Supp. 2023).[3] With several exceptions, the statute prevents platforms from “censor[ing]” a user or a user’s expression based on viewpoint. Tex. Civ. Prac. & Rem. Code Ann. §§143A.002(a), 143 A. 006 (West Cum. Supp. 2023). That ban on “censor[ing]” covers any action to “block, ban, remove, deplatform, demonetize, de-boost, restrict, deny equal access or visibility to, or otherwise discriminate against expression.” §143A.001(1). The statute also requires that “concurrently with the removal” of user content, the platform shall “notify the user” and “explain the reason the content was removed.” §120.103(a)(1). The user gets a right of appeal, and the platform must address an appeal within 14 days. See §§120.103(a)(2), 120.104.

Soon after Florida and Texas enacted those statutes, NetChoice LLC and the Computer & Communications Industry Association (collectively, NetChoice) — trade associations whose members include Facebook and YouTube — brought facial First Amendment challenges against the two laws. District courts in both States entered preliminary injunctions, halting the laws’ enforcement. See 546 F. Supp. 3d 1082, 1096 (ND Fla. 2021); 573 F. Supp. 3d 1092, 1117 (WD Tex. 2021). Each court held that the suit before it is likely to succeed because the statute infringes on the constitutionally protected “editorial judgment” of NetChoice’s members about what material they will display. See 546 F. Supp. 3d, at 1090; 573 F. Supp. 3d, at 1107.

The Eleventh Circuit upheld the injunction of Florida’s law, as to all provisions relevant here. The court held that the State’s restrictions on content moderation trigger First Amendment scrutiny under this Court’s cases protecting “editorial discretion.” 34 F. 4th 1196, 1209, 1216 (2022). When a social-media platform “removes or deprioritizes a user or post,” the court explained, it makes a “judgment rooted in the platform’s own views about the sorts of content and viewpoints that are valuable and appropriate for dissemination.” Id., at 1210. The court concluded that the content-moderation provisions are unlikely to survive “intermediate — let alone strict — scrutiny,” because a State has no legitimate interest in counteracting “private ‘censorship’ ” by “tilt[ing] public debate in a preferred direction.” Id., at 1227–1228. Similarly, the Eleventh Circuit thought the statute’s individualized-explanation requirements likely to fall. Applying the standard from Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U.S. 626 (1985), the court held that the obligation to explain “millions of [decisions] per day” is “unduly burdensome and likely to chill platforms’ protected speech.” 34 F. 4th, at 1230.

The Fifth Circuit disagreed across the board, and so reversed the preliminary injunction before it. In that court’s view, the platforms’ content-moderation activities are “not speech” at all, and so do not implicate the First Amendment. 49 F. 4th 439, 466, 494 (2022). But even if those activities were expressive, the court continued, the State could regulate them to advance its interest in “protecting a diversity of ideas.” Id., at 482 (emphasis deleted). The court further held that the statute’s individualized- explanation provisions would likely survive, again even assuming that the platforms were engaged in speech. Those requirements, the court maintained, are not unduly burdensome under Zauderer because the platforms needed only to “scale up” a “complaint-and-appeal process” they already used. 49 F. 4th, at 487.

We granted certiorari to resolve the split between the Fifth and Eleventh Circuits. 600 U. S. ___ (2023).

II

NetChoice chose to litigate these cases as facial challenges, and that decision comes at a cost. For a host of good reasons, courts usually handle constitutional claims case by case, not en masse. See Washington State Grange v. Washington State Republican Party, 552 U.S. 442, 450–451 (2008). “Claims of facial invalidity often rest on speculation” about the law’s coverage and its future enforcement. Id., at 450. And “facial challenges threaten to short circuit the democratic process” by preventing duly enacted laws from being implemented in constitutional ways. Id., at 451. This Court has therefore made facial challenges hard to win.

That is true even when a facial suit is based on the First Amendment, although then a different standard applies. In other cases, a plaintiff cannot succeed on a facial challenge unless he “establish[es] that no set of circumstances exists under which the [law] would be valid,” or he shows that the law lacks a “plainly legitimate sweep.” United States v. Salerno, 481 U.S. 739, 745 (1987); Washington State Grange, 552 U. S., at 449. In First Amendment cases, however, this Court has lowered that very high bar. To “provide[ ] breathing room for free expression,” we have substituted a less demanding though still rigorous standard. United States v. Hansen, 599 U.S. 762, 769 (2023). The question is whether “a substantial number of [the law’s] applications are unconstitutional, judged in relation to the statute’s plainly legitimate sweep.” Americans for Prosperity Foundation v. Bonta, 594 U.S. 595, 615 (2021); see Hansen, 599 U. S., at 770 (likewise asking whether the law “prohibits a substantial amount of protected speech relative to its plainly legitimate sweep”). So in this singular context, even a law with “a plainly legitimate sweep” may be struck down in its entirety. But that is so only if the law’s unconstitutional applications substantially outweigh its constitutional ones.

So far in these cases, no one has paid much attention to that issue. In the lower courts, NetChoice and the States alike treated the laws as having certain heartland applications, and mostly confined their battle to that terrain. More specifically, the focus was on how the laws applied to the content-moderation practices that giant social-media platforms use on their best-known services to filter, alter, or label their users’ posts. Or more specifically still, the focus was on how the laws applied to Facebook’s News Feed and YouTube’s homepage. Reflecting the parties’ arguments, the Eleventh and Fifth Circuits also mostly confined their analysis in that way. See 34 F. 4th, at 1210, 1213 (considering “platforms like Facebook, Twitter, YouTube, and TikTok” and content moderation in “viewers’ feeds”); 49 F. 4th, at 445, 460, 478, 492 (considering platforms “such as Facebook, Twitter, and YouTube” and referencing users’ feeds); see also id., at 501 (Southwick, J., concurring in part and dissenting in part) (analyzing a curated feed). On their way to opposing conclusions, they concentrated on the same issue: whether a state law can regulate the content-moderation practices used in Facebook’s News Feed (or near equivalents). They did not address the full range of activities the laws cover, and measure the constitutional against the unconstitutional applications. In short, they treated these cases more like as-applied claims than like facial ones.

The first step in the proper facial analysis is to assess the state laws’ scope. What activities, by what actors, do the laws prohibit or otherwise regulate? The laws of course differ one from the other. But both, at least on their face, appear to apply beyond Facebook’s News Feed and its ilk. Members of this Court asked some of the relevant questions at oral argument. Starting with Facebook and the other giants: To what extent, if at all, do the laws affect their other services, like direct messaging or events management? See Tr. of Oral Arg. in No. 22–555, pp. 62–63; Tr. of Oral Arg. in No. 22–277, pp. 24–25; App. in No. 22–277, pp. 129, 159. And beyond those social-media entities, what do the laws have to say, if anything, about how an email provider like Gmail filters incoming messages, how an online marketplace like Etsy displays customer reviews, how a payment service like Venmo manages friends’ financial exchanges, or how a ride-sharing service like Uber runs? See Tr. of Oral Arg. in No. 22–277, at 74–79, 95–98; see also id., at 153 (Solicitor General) (“I have some sympathy [for the Court] here. In preparation for this argument, I’ve been working with my team to say, does this even cover direct messaging? Does this even cover Gmail?”). Those are examples only. The online world is variegated and complex, encompassing an ever-growing number of apps, services, functionalities, and methods for communication and connection. Each might (or might not) have to change because of the provisions, as to either content moderation or individualized explanation, in Florida’s or Texas’s law. Before a court can do anything else with these facial challenges, it must address that set of issues — in short, must “determine what [the law] covers.” Hansen, 599 U. S., at 770.

The next order of business is to decide which of the laws’ applications violate the First Amendment, and to measure them against the rest. For the content-moderation provisions, that means asking, as to every covered platform or function, whether there is an intrusion on protected editorial discretion. See infra, at 13–19. And for the individualized-explanation provisions, it means asking, again as to each thing covered, whether the required disclosures unduly burden expression. See Zauderer, 471 U. S., at 651. Even on a preliminary record, it is not hard to see how the answers might differ as between regulation of Facebook’s News Feed (considered in the courts below) and, say, its direct messaging service (not so considered). Curating a feed and transmitting direct messages, one might think, involve different levels of editorial choice, so that the one creates an expressive product and the other does not. If so, regulation of those diverse activities could well fall on different sides of the constitutional line. To decide the facial challenges here, the courts below must explore the laws’ full range of applications — the constitutionally impermissible and permissible both — and compare the two sets. Maybe the parties treated the content-moderation choices reflected in Facebook’s News Feed and YouTube’s homepage as the laws’ heartland applications because they are the principal things regulated, and should have just that weight in the facial analysis. Or maybe not: Maybe the parties’ focus had all to do with litigation strategy, and there is a sphere of other applications — and constitutional ones — that would prevent the laws’ facial invalidation.

The problem for this Court is that it cannot undertake the needed inquiries. “[W]e are a court of review, not of first view.” Cutter v. Wilkinson, 544 U.S. 709, 718, n. 7 (2005). Neither the Eleventh Circuit nor the Fifth Circuit performed the facial analysis in the way just described. And even were we to ignore the value of other courts going first, we could not proceed very far. The parties have not briefed the critical issues here, and the record is underdeveloped. So we vacate the decisions below and remand these cases. That will enable the lower courts to consider the scope of the laws’ applications, and weigh the unconstitutional as against the constitutional ones.

III

But it is necessary to say more about how the First Amendment relates to the laws’ content-moderation provisions, to ensure that the facial analysis proceeds on the right path in the courts below. That need is especially stark for the Fifth Circuit. Recall that it held that the content choices the major platforms make for their main feeds are “not speech” at all, so States may regulate them free of the First Amendment’s restraints. 49 F. 4th, at 494; see supra, at 8. And even if those activities were expressive, the court held, Texas’s interest in better balancing the marketplace of ideas would satisfy First Amendment scrutiny. See 49 F. 4th, at 482. If we said nothing about those views, the court presumably would repeat them when it next considers NetChoice’s challenge. It would thus find that significant applications of the Texas law — and so significant inputs into the appropriate facial analysis — raise no First Amendment difficulties. But that conclusion would rest on a serious misunderstanding of First Amendment precedent and principle. The Fifth Circuit was wrong in concluding that Texas’s restrictions on the platforms’ selection, ordering, and labeling of third-party posts do not interfere with expression. And the court was wrong to treat as valid Texas’s interest in changing the content of the platforms’ feeds. Explaining why that is so will prevent the Fifth Circuit from repeating its errors as to Facebook’s and YouTube’s main feeds. (And our analysis of Texas’s law may also aid the Eleventh Circuit, which saw the First Amendment issues much as we do, when next considering NetChoice’s facial challenge.) But a caveat: Nothing said here addresses any of the laws’ other applications, which may or may not share the First Amendment problems described below.[4]

A

Despite the relative novelty of the technology before us, the main problem in this case — and the inquiry it calls for — is not new. At bottom, Texas’s law requires the platforms to carry and promote user speech that they would rather discard or downplay. The platforms object that the law thus forces them to alter the content of their expression — a particular edited compilation of third-party speech. See Brief for NetChoice in No. 22–555, pp. 18–34. That controversy sounds a familiar note. We have repeatedly faced the question whether ordering a party to provide a forum for someone else’s views implicates the First Amendment. And we have repeatedly held that it does so if, though only if, the regulated party is engaged in its own expressive activity, which the mandated access would alter or disrupt. So too we have held, when applying that principle, that expressive activity includes presenting a curated compilation of speech originally created by others. A review of the relevant precedents will help resolve the question here.

The seminal case is Miami Herald Publishing Co. v. Tornillo, 418 U.S. 241 (1974). There, a Florida law required a newspaper to give a political candidate a right to reply when it published “criticism and attacks on his record.” Id., at 243. The Court held the law to violate the First Amendment because it interfered with the newspaper’s “exercise of editorial control and judgment.” Id., at 258. Forcing the paper to print what “it would not otherwise print,” the Court explained, “intru[ded] into the function of editors.” Id., at 256, 258. For that function was, first and foremost, to make decisions about the “content of the paper” and “[t]he choice of material to go into” it. Id., at 258. In protecting that right of editorial control, the Court recognized a possible downside. It noted the access advocates’ view (similar to the States’ view here) that “modern media empires” had gained ever greater capacity to “shape” and even “manipulate popular opinion.” Id., at 249–250. And the Court expressed some sympathy with that diagnosis. See id., at 254. But the cure proposed, it concluded, collided with the First Amendment’s antipathy to state manipulation of the speech market. Florida, the Court explained, could not substitute “governmental regulation” for the “crucial process” of editorial choice. Id., at 258.

Next up was Pacific Gas & Elec. Co. v. Public Util. Comm’n of Cal., 475 U.S. 1 (1986) (PG&E), which the Court thought to follow naturally from Tornillo. See 475 U. S., at 9–12 (plurality opinion); id., at 21 (Burger, C. J., concurring). A private utility in California regularly put a newsletter in its billing envelopes expressing its views of energy policy. The State directed it to include as well material from a consumer-advocacy group giving a different perspective. The utility objected, and the Court held again that the interest in “offer[ing] the public a greater variety of views” could not justify the regulation. Id., at 12. California was compelling the utility (as Florida had compelled a newspaper) “to carry speech with which it disagreed” and thus to “alter its own message.” Id., at 11, n. 7, 16.

In Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622 (1994) (Turner I ), the Court further underscored the constitutional protection given to editorial choice. At issue were federal “must-carry” rules, requiring cable operators to allocate some of their channels to local broadcast stations. The Court had no doubt that the First Amendment was implicated, because the operators were engaging in expressive activity. They were, the Court explained, “exercising editorial discretion over which stations or programs to include in [their] repertoire.” Id., at 636. And the rules “interfere[d]” with that discretion by forcing the operators to carry stations they would not otherwise have chosen. Id., at 643–644. In a later decision, the Court ruled that the regulation survived First Amendment review because it was necessary to prevent the demise of local broadcasting. See Turner Broadcasting System, Inc. v. FCC, 520 U.S. 180, 185, 189–190 (1997) (Turner II ); see infra, at 28, n. 10. But for purposes of today’s cases, the takeaway of Turner is this holding: A private party’s collection of third-party content into a single speech product (the operators’ “repertoire” of programming) is itself expressive, and intrusion into that activity must be specially justified under the First Amendment.

The capstone of those precedents came in Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, Inc., 515 U.S. 557 (1995), when the Court considered (of all things) a parade. The question was whether Massachusetts could require the organizers of a St. Patrick’s Day parade to admit as a participant a gay and lesbian group seeking to convey a message of “pride.” Id., at 561. The Court held unanimously that the First Amendment precluded that compulsion. The “selection of contingents to make a parade,” it explained, is entitled to First Amendment protection, no less than a newspaper’s “presentation of an edited compilation of [other persons’] speech.” Id., at 570 (citing Tornillo, 418 U. S., at 258). And that meant the State could not tell the parade organizers whom to include. Because “every participating unit affects the message,” said the Court, ordering the group’s admittance would “alter the expressive content of the[ ] parade.” Hurley, 515 U. S., at 572–573. The parade’s organizers had “decided to exclude a message [they] did not like from the communication [they] chose to make,” and that was their decision alone. Id., at 574.

On two other occasions, the Court distinguished Tornillo and its progeny for the flip-side reason — because in those cases the compelled access did not affect the complaining party’s own expression. First, in PruneYard Shopping Center v. Robins, 447 U.S. 74 (1980), the Court rejected a shopping mall’s First Amendment challenge to a California law requiring it to allow members of the public to distribute handbills on its property. The mall owner did not claim that he (or the mall) was engaged in any expressive activity. Indeed, as the PG&E Court later noted, he “did not even allege that he objected to the content of the pamphlets” passed out at the mall. 475 U. S., at 12. Similarly, in Rumsfeld v. Forum for Academic and Institutional Rights, Inc., 547 U.S. 47 (2006) (FAIR), the Court reiterated that a First Amendment claim will not succeed when the entity objecting to hosting third-party speech is not itself engaged in expression. The statute at issue required law schools to allow the military to participate in on-campus recruiting. The Court held that the schools had no First Amendment right to exclude the military based on its hiring policies, because the schools “are not speaking when they host interviews.” Id., at 64. Or stated again, with reference to the just-described precedents: Because a “law school’s recruiting services lack the expressive quality of a parade, a newsletter, or the editorial page of a newspaper,” the required “accommodation of a military recruiter[ ]” did not “interfere with any message of the school.” Ibid.

That is a slew of individual cases, so consider three general points to wrap up. Not coincidentally, they will figure in the upcoming discussion of the First Amendment problems the statutes at issue here likely present as to Facebook’s News Feed and similar products.

First, the First Amendment offers protection when an entity engaging in expressive activity, including compiling and curating others’ speech, is directed to accommodate messages it would prefer to exclude. “[T]he editorial function itself is an aspect of speech.” Denver Area Ed. Telecommunications Consortium, Inc. v. FCC, 518 U.S. 727, 737 (1996) (plurality opinion). Or said just a bit differently: An entity “exercis[ing] editorial discretion in the selection and presentation” of content is “engage[d] in speech activity.” Arkansas Ed. Television Comm’n v. Forbes, 523 U.S. 666, 674 (1998). And that is as true when the content comes from third parties as when it does not. (Again, think of a newspaper opinion page or, if you prefer, a parade.) Deciding on the third-party speech that will be included in or excluded from a compilation — and then organizing and presenting the included items — is expressive activity of its own. And that activity results in a distinctive expressive product. When the government interferes with such editorial choices — say, by ordering the excluded to be included — it alters the content of the compilation. (It creates a different opinion page or parade, bearing a different message.) And in so doing — in overriding a private party’s expressive choices — the government confronts the First Amendment.[5]

Second, none of that changes just because a compiler includes most items and excludes just a few. That was the situation in Hurley. The St. Patrick’s Day parade at issue there was “eclectic”: It included a “wide variety of patriotic, commercial, political, moral, artistic, religious, athletic, public service, trade union, and eleemosynary themes, as well as conflicting messages.” 515 U. S., at 562. Or otherwise said, the organizers were “rather lenient in admitting participants.” Id., at 569. No matter. A “narrow, succinctly articulable message is not a condition of constitutional protection.” Ibid. It “is enough” for a compiler to exclude the handful of messages it most “disfavor[s].” Id., at 574. Suppose, for example, that the newspaper in Tornillo had granted a right of reply to all but one candidate. It would have made no difference; the Florida statute still could not have altered the paper’s policy. Indeed, that kind of focused editorial choice packs a peculiarly powerful expressive punch.

Third, the government cannot get its way just by asserting an interest in improving, or better balancing, the marketplace of ideas. Of course, it is critically important to have a well-functioning sphere of expression, in which citizens have access to information from many sources. That is the whole project of the First Amendment. And the government can take varied measures, like enforcing competition laws, to protect that access. Cf., e.g., Turner I, 512 U. S., at 647 (protecting local broadcasting); Hurley, 515 U. S., at 577 (discussing Turner I ). But in case after case, the Court has barred the government from forcing a private speaker to present views it wished to spurn in order to rejigger the expressive realm. The regulations in Tornillo, PG&E, and Hurley all were thought to promote greater diversity of expression. See supra, at 14–16. They also were thought to counteract advantages some private parties possessed in controlling “enviable vehicle[s]” for speech. Hurley, 515 U. S., at 577. Indeed, the Tornillo Court devoted six pages of its opinion to recounting a critique of the then-current media environment — in particular, the disproportionate “influen[ce]” of a few speakers — similar to one heard today (except about different entities). 418 U. S., at 249; see id., at 248–254; supra, at 14–15. It made no difference. However imperfect the private marketplace of ideas, here was a worse proposal — the government itself deciding when speech was imbalanced, and then coercing speakers to provide more of some views or less of others.

B

“[W]hatever the challenges of applying the Constitution to ever-advancing technology, the basic principles” of the First Amendment “do not vary.” Brown v. Entertainment Merchants Assn., 564 U.S. 786, 790 (2011). New communications media differ from old ones in a host of ways: No one thinks Facebook’s News Feed much resembles an insert put in a billing envelope. And similarly, today’s social media pose dangers not seen earlier: No one ever feared the effects of newspaper opinion pages on adolescents’ mental health. But analogies to old media, even if imperfect, can be useful. And better still as guides to decision are settled principles about freedom of expression, including the ones just described. Those principles have served the Nation well over many years, even as one communications method has given way to another. And they have much to say about the laws at issue here. These cases, to be sure, are at an early stage; the record is incomplete even as to the major social-media platforms’ main feeds, much less the other applications that must now be considered. See supra, at 12. But in reviewing the District Court’s preliminary injunction, the Fifth Circuit got its likelihood-of-success finding wrong. Texas is not likely to succeed in enforcing its law against the platforms’ application of their content-moderation policies to the feeds that were the focus of the proceedings below. And that is because of the core teaching elaborated in the above-summarized decisions: The government may not, in supposed pursuit of better expressive balance, alter a private speaker’s own editorial choices about the mix of speech it wants to convey.

Most readers are likely familiar with Facebook’s News Feed or YouTube’s homepage; assuming so, feel free to skip this paragraph (and maybe a couple more). For the uninitiated, though, each of those feeds presents a user with a continually updating stream of other users’ posts. For Facebook’s News Feed, any user may upload a message, whether verbal or visual, with content running the gamut from “vacation pictures from friends” to “articles from local or national news outlets.” App. in No. 22–555, at 139a. And whenever a user signs on, Facebook delivers a personalized collection of those stories. Similarly for YouTube. Its users upload all manner of videos. And any person opening the website or mobile app receives an individualized list of video recommendations.

The key to the scheme is prioritization of content, achieved through the use of algorithms. Of the billions of posts or videos (plus advertisements) that could wind up on a user’s customized feed or recommendations list, only the tiniest fraction do. The selection and ranking is most often based on a user’s expressed interests and past activities. But it may also be based on more general features of the communication or its creator. Facebook’s Community Standards and YouTube’s Community Guidelines detail the messages and videos that the platforms disfavor. The platforms write algorithms to implement those standards — for example, to prefer content deemed particularly trustworthy or to suppress content viewed as deceptive (like videos promoting “conspiracy theor[ies]”). Id., at 113a.

Beyond rankings lie labels. The platforms may attach “warning[s], disclaimers, or general commentary” — for example, informing users that certain content has “not been verified by official sources.” Id., at 75a. Likewise, they may use “information panels” to give users “context on content relating to topics and news prone to misinformation, as well as context about who submitted the content.” Id., at 114a. So, for example, YouTube identifies content submitted by state-supported media channels, including those funded by the Russian Government. See id., at 76a.

But sometimes, the platforms decide, providing more information is not enough; instead, removing a post is the right course. The platforms’ content-moderation policies also say when that is so. Facebook’s Standards, for example, proscribe posts — with exceptions for “newsworth[iness]” and other “public interest value” — in categories and subcategories including: Violence and Criminal Behavior (e.g., violence and incitement, coordinating harm and publicizing crime, fraud and deception); Safety (e.g., suicide and self-injury, sexual exploitation, bullying and harassment); Objectionable Content (e.g., hate speech, violent and graphic content); Integrity and Authenticity (e.g., false news, manipulated media). Id., at 412a–415a, 441a–442a. YouTube’s Guidelines similarly target videos falling within categories like: hate speech, violent or graphic content, child safety, and misinformation (including about elections and vaccines). See id., at 430a–432a. The platforms thus unabashedly control the content that will appear to users, exercising authority to remove, label or demote messages they disfavor.[6]

Except that Texas’s law limits their power to do so. As noted earlier, the law’s central provision prohibits the large social-media platforms (and maybe other entities [7]) from “censor[ing]” a “user’s expression” based on its “viewpoint.” §143A.002(a)(2); see supra, at 7. The law defines “expression” broadly, thus including pretty much anything that might be posted. See §143A.001(2). And it defines “censor” to mean “block, ban, remove, deplatform, demonetize, de-boost, restrict, deny equal access or visibility to, or otherwise discriminate against expression.” §143A.001(1).[8] That is a long list of verbs, but it comes down to this: The platforms cannot do any of the things they typically do (on their main feeds) to posts they disapprove — cannot demote, label, or remove them — whenever the action is based on the post’s viewpoint.[9] And what does that “based on viewpoint” requirement entail? Doubtless some of the platforms’ content-moderation practices are based on characteristics of speech other than viewpoint (e.g., on subject matter). But if Texas’s law is enforced, the platforms could not — as they in fact do now — disfavor posts because they:

• support Nazi ideology;
• advocate for terrorism;
• espouse racism, Islamophobia, or anti-Semitism;
• glorify rape or other gender-based violence;
• encourage teenage suicide and self-injury;
• discourage the use of vaccines;
• advise phony treatments for diseases;
• advance false claims of election fraud.

The list could continue for a while.[10] The point of it is not that the speech environment created by Texas’s law is worse than the ones to which the major platforms aspire on their main feeds. The point is just that Texas’s law profoundly alters the platforms’ choices about the views they will, and will not, convey.

And we have time and again held that type of regulation to interfere with protected speech. Like the editors, cable operators, and parade organizers this Court has previously considered, the major social-media platforms are in the business, when curating their feeds, of combining “multifarious voices” to create a distinctive expressive offering. Hurley, 515 U. S., at 569. The individual messages may originate with third parties, but the larger offering is the platform’s. It is the product of a wealth of choices about whether—and, if so, how—to convey posts having a certain content or viewpoint. Those choices rest on a set of beliefs about which messages are appropriate and which are not (or which are more appropriate and which less so). And in the aggregate they give the feed a particular expressive quality. Consider again an opinion page editor, as in Tornillo, who wants to publish a variety of views, but thinks some things off-limits (or, to change the facts, worth only a couple of column inches). “The choice of material,” the “decisions made [as to] content,” the “treatment of public issues”—“whether fair or unfair”—all these “constitute the exercise of editorial control and judgment.” Tornillo, 418 U. S., at 258. For a paper, and for a platform too. And the Texas law (like Florida’s earlier right-of-reply statute) targets those expressive choices—in particular, by forcing the major platforms to present and promote content on their feeds that they regard as objectionable.

That those platforms happily convey the lion’s share of posts submitted to them makes no significant First Amendment difference. Contra, 49 F. 4th, at 459–461 (arguing otherwise). To begin with, Facebook and YouTube exclude (not to mention, label or demote) lots of content from their News Feed and homepage. The Community Standards and Community Guidelines set out in copious detail the varied kinds of speech the platforms want no truck with. And both platforms appear to put those manuals to work. In a single quarter of 2021, Facebook removed from its News Feed more than 25 million pieces of “hate speech content” and almost 9 million pieces of “bullying and harassment content.” App. in No. 22–555, at 80a. Similarly, YouTube deleted in one quarter more than 6 million videos violating its Guidelines. See id., at 116a. And among those are the removals the Texas law targets. What is more, this Court has already rightly declined to focus on the ratio of rejected to accepted content. Recall that in Hurley, the parade organizers welcomed pretty much everyone, excluding only those who expressed a message of gay pride. See supra, at 18. The Court held that the organizers’ “lenient” admissions policy—and their resulting failure to express a “particularized message”—did “not forfeit” their right to reject the few messages they found harmful or offensive. 515 U. S., at 569, 574. So too here, though the excluded viewpoints differ. That Facebook and YouTube convey a mass of messages does not license Texas to prohibit them from deleting posts with, say, “hate speech” based on “sexual orientation.” App. in No. 22–555, at 126a, 155a; see id., at 431a. It is as much an editorial choice to convey all speech except in select categories as to convey only speech within them.

Similarly, the major social-media platforms do not lose their First Amendment protection just because no one will wrongly attribute to them the views in an individual post. Contra, 49 F. 4th, at 462 (arguing otherwise). For starters, users may well attribute to the platforms the messages that the posts convey in toto. Those messages—communicated by the feeds as a whole—derive largely from the platforms’ editorial decisions about which posts to remove, label, or demote. And because that is so, the platforms may indeed “own” the overall speech environment. In any event, this Court has never hinged a compiler’s First Amendment protection on the risk of misattribution. The Court did not think in Turner—and could not have thought in Tornillo or PG&E—that anyone would view the entity conveying the third-party speech at issue as endorsing its content. See Turner I, 512 U. S., at 655 (“[T]here appears little risk” of such misattribution). Yet all those entities, the Court held, were entitled to First Amendment protection for refusing to carry the speech. See supra, at 14–16. To be sure, the Court noted in PruneYard and FAIR, when denying such protection, that there was little prospect of misattribution. See 447 U. S., at 87; 547 U. S., at 65. But the key fact in those cases, as noted above, was that the host of the third-party speech was not itself engaged in expression. See supra, at 16–17. The current record suggests the opposite as to Facebook’s News Feed and YouTube’s homepage. When the platforms use their Standards and Guidelines to decide which third-party content those feeds will display, or how the display will be ordered and organized, they are making expressive choices. And because that is true, they receive First Amendment protection.

C

And once that much is decided, the interest Texas relies on cannot sustain its law. In the usual First Amendment case, we must decide whether to apply strict or intermediate scrutiny. But here we need not. Even assuming that the less stringent form of First Amendment review applies, Texas’s law does not pass. Under that standard, a law must further a “substantial governmental interest” that is “unrelated to the suppression of free expression.” United States v. ��’B�������, 391 U.S. 367, 377 (1968). Many possible interests relating to social media can meet that test; nothing said here puts regulation of NetChoice’s members off-limits as to a whole array of subjects. But the interest Texas has asserted cannot carry the day: It is very much related to the suppression of free expression, and it is not valid, let alone substantial.

Texas has never been shy, and always been consistent, about its interest: The objective is to correct the mix of speech that the major social-media platforms present. In this Court, Texas described its law as “respond[ing]” to the platforms’ practice of “favoring certain viewpoints.” Brief for Texas 7; see id., at 27 (explaining that the platforms’ “discrimination” among messages “led to [the law’s] enactment”). The large social-media platforms throw out (or encumber) certain messages; Texas wants them kept in (and free from encumbrances), because it thinks that would create a better speech balance. The current amalgam, the State explained in earlier briefing, was “skewed” to one side. 573 F. Supp. 3d, at 1116. And that assessment mirrored the stated views of those who enacted the law, save that the latter had a bit more color. The law’s main sponsor explained that the “West Coast oligarchs” who ran social-media companies were “silenc[ing] conservative viewpoints and ideas.” Ibid. The Governor, in signing the legislation, echoed the point: The companies were fomenting a “dangerous movement” to “silence” conservatives. Id., at 1108; see id., at 1099 (“[S]ilencing conservative views is un- American, it’s un-Texan and it’s about to be illegal in Texas”).

But a State may not interfere with private actors’ speech to advance its own vision of ideological balance. States (and their citizens) are of course right to want an expressive realm in which the public has access to a wide range of views. That is, indeed, a fundamental aim of the First Amendment. But the way the First Amendment achieves that goal is by preventing the government from “tilt[ing] public debate in a preferred direction.” Sorrell v. IMS Health Inc., 564 U.S. 552, 578–579 (2011). It is not by licensing the government to stop private actors from speaking as they wish and preferring some views over others. And that is so even when those actors possess “enviable vehicle[s]” for expression. Hurley, 515 U. S., at 577. In a better world, there would be fewer inequities in speech opportunities; and the government can take many steps to bring that world closer. But it cannot prohibit speech to improve or better balance the speech market. On the spectrum of dangers to free expression, there are few greater than allowing the government to change the speech of private actors in order to achieve its own conception of speech nirvana. That is why we have said in so many contexts that the government may not “restrict the speech of some elements of our society in order to enhance the relative voice of others.” Buckley v. Valeo, 424 U.S. 1, 48–49 (1976) (per curiam). That unadorned interest is not “unrelated to the suppression of free expression,” and the government may not pursue it consistent with the First Amendment.

The Court’s decisions about editorial control, as discussed earlier, make that point repeatedly. See supra, at 18–19. Again, the question those cases had in common was whether the government could force a private speaker, including a compiler and curator of third-party speech, to convey views it disapproved. And in most of those cases, the government defended its regulation as yielding greater balance in the marketplace of ideas. But the Court — in Tornillo, in PG&E, and again in Hurley — held that such an interest could not support the government’s effort to alter the speaker’s own expression. “Our cases establish,” the PG&E Court wrote, “that the State cannot advance some points of view by burdening the expression of others.” 475 U. S., at 20. So the newspaper, the public utility, the parade organizer — whether acting “fair[ly] or unfair[ly]” — could exclude the unwanted message, free from government interference. Tornillo, 418 U. S., at 258; see United States Telecom Assn. v. FCC, 855 F.3d 381, 432 (CADC 2017) (Kavanaugh, J., dissenting from denial of rehearing en banc) (“[E]xcept in rare circumstances, the First Amendment does not allow the Government to regulate the content choices of private editors just so that the Government may enhance certain voices and alter the content available to the citizenry”).[11]

The case here is no different. The interest Texas asserts is in changing the balance of speech on the major platforms’ feeds, so that messages now excluded will be included. To describe that interest, the State borrows language from this Court’s First Amendment cases, maintaining that it is preventing “viewpoint discrimination.” Brief for Texas 19; see supra, at 26–27. But the Court uses that language to say what governments cannot do: They cannot prohibit private actors from expressing certain views. When Texas uses that language, it is to say what private actors cannot do: They cannot decide for themselves what views to convey. The innocent-sounding phrase does not redeem the prohibited goal. The reason Texas is regulating the content- moderation policies that the major platforms use for their feeds is to change the speech that will be displayed there. Texas does not like the way those platforms are selecting and moderating content, and wants them to create a different expressive product, communicating different values and priorities. But under the First Amendment, that is a preference Texas may not impose.

IV

These are facial challenges, and that matters. To succeed on its First Amendment claim, NetChoice must show that the law at issue (whether from Texas or from Florida) “prohibits a substantial amount of protected speech relative to its plainly legitimate sweep.” Hansen, 599 U. S., at 770. None of the parties below focused on that issue; nor did the Fifth or Eleventh Circuits. But that choice, unanimous as it has been, cannot now control. Even in the First Amendment context, facial challenges are disfavored, and neither parties nor courts can disregard the requisite inquiry into how a law works in all of its applications. So on remand, each court must evaluate the full scope of the law’s coverage. It must then decide which of the law’s applications are constitutionally permissible and which are not, and finally weigh the one against the other. The need for NetChoice to carry its burden on those issues is the price of its decision to challenge the laws as a whole.

But there has been enough litigation already to know that the Fifth Circuit, if it stayed the course, would get wrong at least one significant input into the facial analysis. The parties treated Facebook’s News Feed and YouTube’s homepage as the heartland applications of the Texas law. At least on the current record, the editorial judgments influencing the content of those feeds are, contrary to the Fifth Circuit’s view, protected expressive activity. And Texas may not interfere with those judgments simply because it would prefer a different mix of messages. How that matters for the requisite facial analysis is for the Fifth Circuit to decide. But it should conduct that analysis in keeping with two First Amendment precepts. First, presenting a curated and “edited compilation of [third party] speech” is itself protected speech. Hurley, 515 U. S., at 570. And second, a State “cannot advance some points of view by burdening the expression of others.” PG&E, 475 U. S., at 20. To give government that power is to enable it to control the expression of ideas, promoting those it favors and suppressing those it does not. And that is what the First Amendment protects all of us from.

We accordingly vacate the judgments of the Courts of Appeals for the Fifth and Eleventh Circuits and remand the cases for further proceedings consistent with this opinion.

It is so ordered.

Notes

[1] *Justice Jackson joins Parts I, II, and III–A of this opinion.

[2]  The definition of “social-media platforms” covers “any information service, system, Internet search engine, or access software provider” that “[p]rovides or enables computer access by multiple users to a computer server, including an Internet platform or a social media site.” Fla. Stat. §501.2041(1)(g)(1).

[3]  The statute further clarifies that it does not cover internet service providers, email providers, and any online service, website, or app consisting “primarily of news, sports, entertainment, or other information or content that is not user generated but is preselected by the provider.” §120.001(1).

[4]  Although the discussion below focuses on Texas’s content-moderation provisions, it also bears on how the lower courts should address the individualized-explanation provisions in the upcoming facial inquiry. As noted, requirements of that kind violate the First Amendment if they unduly burden expressive activity. See Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U.S. 626, 651 (1985); supra, at 11. So our explanation of why Facebook and YouTube are engaged in expression when they make content-moderation choices in their main feeds should inform the courts’ further consideration of that issue.

[5]  Of course, an entity engaged in expressive activity when performing one function may not be when carrying out another. That is one lesson of FAIR. The Court ruled as it did because the law schools’ recruiting services were not engaged in expression. See 547 U.S. 47, 64 (2006). The case could not have been resolved on that ground if the regulation had affected what happened in law school classes instead.

[6]  We therefore do not deal here with feeds whose algorithms respond solely to how users act online—giving them the content they appear to want, without any regard to independent content standards. See post, at 2 (Barrett, J., concurring). Like them or loathe them, the Community Standards and Community Guidelines make a wealth of user-agnostic judgments about what kinds of speech, including what viewpoints, are not worthy of promotion. And those judgments show up in Facebook’s and YouTube’s main feeds.

[7]  The scope of the Texas law, a matter crucial to the facial inquiry, is unsettled, as previously discussed. See supra, at 10–11. The Texas solicitor general at oral argument stated that he understood the law to cover Facebook and YouTube, but “d[id]n’t know” whether it also covered other platforms and applications. Tr. of Oral Arg. in No. 22–555, pp. 61–62.

[8]  In addition to barring “censor[ship]” of “expression,” the law bars “censor[ship]” of people. More specifically, it prohibits taking the designated “censor[ial]” actions against any “user” based on his “viewpoint,” regardless of whether that “viewpoint is expressed on a social media platform.” §§143A.002(a)(1), (b); see supra, at 7. Because the Fifth Circuit did not focus on that provision, instead confining its analysis to the law’s ban on “censor[ing]” a “user’s expression” on the platform, we do the same.

[9]  The Texas solicitor general explained at oral argument that the Texas law allows the platforms to remove “categories” of speech, so long as they are not based on viewpoint. See Tr. of Oral Arg. in No. 22–555, at 69–70; §120.052 (Acceptable Use Policy). The example he gave was speech about Al-Qaeda. Under the law, a platform could remove all posts about Al-Qaeda, regardless of viewpoint. But it could not stop the “pro-Al-Qaeda” speech alone; it would have to stop the “anti-Al-Qaeda” speech too. Tr. of Oral Arg. in No. 22–555, at 70. So again, the law, as described by the solicitor general, prevents the platforms from disfavoring posts because they express one view of a subject.

[10]  Details on both the enumerated examples and similar ones are found in Facebook’s Community Standards and YouTube’s Community Guidelines. See https://transparency.meta.com/policies/community-standards; https://support.google.com/youtube/answer/9288567.

[11]  Texas claims Turner as a counter-example, but that decision offers no help to speak of. Turner did indeed hold that the FCC’s must-carry provisions, requiring cable operators to give some of their channel space to local broadcast stations, passed First Amendment muster. See supra, at 15. But the interest there advanced was not to balance expressive content; rather, the interest was to save the local-broadcast industry, so that it could continue to serve households without cable. That interest, the Court explained, was “unrelated to the content of expression” disseminated by either cable or broadcast speakers. Turner I, 512 U.S. 622, 647 (1994). And later, the Hurley Court again noted the difference. It understood the Government interest in Turner as one relating to competition policy: The FCC needed to limit the cable operators’ “monopolistic,” gatekeeping position “in order to allow for the survival of broadcasters.” 515 U. S., at 577. Unlike in regulating the parade—or here in regulating Facebook’s News Feed or YouTube’s homepage—the Government’s interest was “not the alteration of speech.” Ibid. And when that is so, the prospects of permissible regulation are entirely different.

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 22–277 and 22–555

_________________

ASHLEY MOODY, ATTORNEY GENERAL OF FLORIDA, et al., PETITIONERS (22–277)

v.

NETCHOICE, LLC, dba NETCHOICE, et al.

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE ELEVENTH CIRCUIT

NETCHOICE, LLC, dba NETCHOICE, et al., PETITIONERS (22–555)

v.

KEN PAXTON, ATTORNEY GENERAL OF TEXAS

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT

[July 1, 2024]

Justice Jackson, concurring in part and concurring in the judgment.

These cases present a complex clash between two novel state laws and the alleged First Amendment rights of several of the largest social media platforms. Some things are already clear. Not every potential action taken by a social media company will qualify as expression protected under the First Amendment. But not every hypothesized regulation of such a company’s operations will necessarily be able to withstand the force of the First Amendment’s protections either. Beyond those broadest of statements, it is difficult to say much more at this time. With these records and lower court decisions, we are not able to adequately evaluate whether the challenged state laws are facially valid.

That is in no small part because, as all Members of the Court acknowledge, plaintiffs bringing a facial challenge must clear a high bar. See ante, at 9–10 (majority opinion); post, at 13–14 (Alito, J., concurring in judgment). The Eleventh Circuit failed to appreciate the nature of this challenge, and the Fifth Circuit did not adequately evaluate it. That said, I agree with Justice Barrett that the Eleventh Circuit at least fairly stated our First Amendment precedent, whereas the Fifth Circuit did not. See ante, at 1 (concurring opinion); see also ante, at 13–19 (majority opinion). On remand, then, both courts will have to undertake their legal analyses anew.

In doing so, the lower courts must address these cases at the right level of specificity. The question is not whether an entire category of corporations (like social media companies) or a particular entity (like Facebook) is generally engaged in expression. Nor is it enough to say that a given activity (say, content moderation) for a particular service (the News Feed, for example) seems roughly analogous to a more familiar example from our precedent. Cf. Red Lion Broadcasting Co. v. FCC, 395 U.S. 367, 386 (1969) (positing that “differences in the characteristics of new media justify differences in the First Amendment standards applied to them”). Even when evaluating a broad facial challenge, courts must make sure they carefully parse not only what entities are regulated, but how the regulated activities actually function before deciding if the activity in question constitutes expression and therefore comes within the First Amendment’s ambit. See Brief for Knight First Amendment Institute at Columbia University as Amicus Curiae 11–12. Thus, further factual development may be necessary before either of today’s challenges can be fully and fairly addressed.

In light of the high bar for facial challenges and the state of these cases as they come to us, I would not go on to treat either like an as-applied challenge and preview our potential ruling on the merits. Faced with difficult constitutional issues arising in new contexts on undeveloped records, this Court should strive to avoid deciding more than is necessary. See Ashwander v. TVA, 297 U.S. 288, 346–347 (1936) (Brandeis, J., concurring). In my view, such restraint is warranted today.

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 22–277 and 22–555

_________________

ASHLEY MOODY, ATTORNEY GENERAL OF FLORIDA, et al., PETITIONERS

22–277v.

NETCHOICE, LLC, dba NETCHOICE, et al.

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE ELEVENTH CIRCUIT

NETCHOICE, LLC, dba NETCHOICE, et al., PETITIONERS

22–555v.

KEN PAXTON, ATTORNEY GENERAL OF TEXAS

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT

[July 1, 2024]

Justice Thomas, concurring in the judgment.

I agree with the Court’s decision to vacate and remand because NetChoice and the Computer and Communications Industry Association (together, the trade associations) have not established that Texas’s H. B. 20 and Florida’s S. B. 7072 are facially unconstitutional.

I cannot agree, however, with the Court’s decision to opine on certain applications of those statutes. The Court’s discussion is unnecessary to its holding. See Jama v. Immigration and Customs Enforcement, 543 U.S. 335, 351, n. 12 (2005) (“Dictum settles nothing, even in the court that utters it”). Moreover, the Court engages in the exact type of analysis that it chastises the Courts of Appeals for performing. It faults the Courts of Appeals for focusing on only one subset of applications, rather than determining whether each statute’s “full range of applications” are constitutional. See ante, at 10, 12. But, the Court repeats that very same error. Out of the sea of “variegated and complex” functions that platforms perform, ante, at 11, the Court plucks out two (Facebook’s News Feed and YouTube’s homepage), and declares that they may be protected by the First Amendment. See ante, at 26 (opining on what the “current record suggests”). The Court does so on a record that it itself describes as “incomplete” and “underdeveloped,” ante, at 12, 20, and by sidestepping several pressing factual and legal questions, see post, at 29–32 (Alito, J., concurring in judgment). As Justice Alito explains, the Court’s approach is both unwarranted and mistaken. See ibid.

I agree with Justice Alito’s analysis and join his opinion in full. I write separately to add two observations on the merits and to highlight a more fundamental jurisdictional problem. The trade associations have brought facial challenges alleging that H. B. 20 and S. B. 7072 are unconstitutional in many or all of their applications. But, Art-icle III of the Constitution permits federal courts to exercise judicial power only over “Cases” and “Controversies.” Accordingly, federal courts can decide whether a statute is constitutional only as applied to the parties before them—they lack authority to deem a statute “facially” unconstitutional.

I

As Justice Alito explains, the trade associations have failed to provide many of the basic facts necessary to evaluate their challenges to H. B. 20 and S. B. 7072. See post, at 22–29. I make two additional observations.

First, with respect to certain provisions of H. B. 20 and S. B. 7072, the Court assumes that the framework outlined in Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio, 471 U.S. 626 (1985), applies. See ante, at 11. In that case, the Court held that laws requiring the disclosure of factual information in commercial advertising may satisfy the First Amendment if the disclosures are “reasonably related” to the Government’s interest in preventing consumer deception. 471 U. S., at 651. Because the trade associations did not contest Zauderer’s applicability before the Eleventh Circuit and both lower courts applied its framework, I agree with the Court’s decision to rely upon Zauderer at this stage. However, I think we should reconsider Zauderer and its progeny. “I am skeptical of the premise on which Zauderer rests—that, in the commercial-speech context, the First Amendment interests implicated by disclosure requirements are substantially weaker than those at stake when speech is actually suppressed.” Milavetz, Gallop & Milavetz, P. A. v. United States, 559 U.S. 229, 255 (2010) (Thomas, J., concurring in part and concurring in judgment) (internal quotation marks omitted).

Second, the common-carrier doctrine should continue to guide the lower courts’ examination of the trade associations’ claims on remand. See post, at 18, and n. 17, 30 (opinion of Alito, J.). “[O]ur legal system and its British predecessor have long subjected certain businesses, known as common carriers, to special regulations, including a general requirement to serve all comers.” Biden v. Knight First Amendment Institute at Columbia Univ., 593 U. S. ___, ___ (2021) (Thomas, J., concurring in grant of certiorari) (slip op., at 3). Moreover, “there is clear historical precedent for regulating transportation and communications networks in a similar manner as traditional common carriers” given their many similarities. Id., at ___ (slip op., at 5). Though they reached different conclusions, both the Fifth Circuit and the Eleventh Circuit appropriately strove to apply the common-carrier doctrine in assessing the constitutionality of H. B. 20 and S. B. 7072 respectively. See 49 F. 4th 439, 469–480 (CA5 2022); NetChoice v. Attorney Gen., Fla., 34 F. 4th 1196, 1219–1222 (CA11 2022).

The common-carrier doctrine may have weighty implications for the trade associations’ claims. But, the same factual barriers that preclude the Court from assessing the trade associations’ claims under our First Amendment precedents also prevent us from applying the common-carrier doctrine in this posture. At a minimum, we would need to pinpoint the regulated parties and specific conduct being regulated. On remand, however, both lower courts should continue to consider the common-carrier doctrine.

II

The opinions in these cases detail many of the considerable hurdles that currently preclude resolution of the trade associations’ claims. See ante, at 9–10; ante, at 1–4 (Barrett, J., concurring); post, at 22–32 (opinion of Alito, J.). The most significant problem of all, however, has yet to be addressed: Federal courts lack authority to adjudicate the trade associations’ facial challenges.

Rather than allege that the statutes impermissibly regulate them, the trade associations assert that H. B. 20 and S. B. 7072 are actually unconstitutional in most or all of their applications. This type of challenge, called a facial challenge, is “an attack on a statute itself as opposed to a particular application.” Los Angeles v. Patel, 576 U.S. 409, 415 (2015).

Facial challenges are fundamentally at odds with Article III. Because Article III limits federal courts’ judicial power to cases or controversies, federal courts “lac[k] the power to pronounce that [a] statute is unconstitutional” as applied to nonparties. Americans for Prosperity Foundation v. Bonta, 594 U.S. 595, 621 (2021) (Thomas, J., concurring in part and concurring in judgment) (internal quotation marks omitted). Entertaining facial challenges in spite of that limitation arrogates powers reserved to the political branches and disturbs the relationship between the Federal Government and the States. The practice of adjudicating facial challenges creates practical concerns as well. Facial challenges’ dubious historical roots further confirm that the doctrine should have no place in our jurisprudence.

A

1

Article III empowers federal courts to exercise “judicial Power” only over “Cases” and “Controversies.” This Court has long recognized that those terms impose substantive constraints on the authority of federal courts. See Muskrat v. United States, 219 U.S. 346, 356–358 (1911); see also Steel Co. v. Citizens for Better Environment, 523 U.S. 83, 102 (1998). One corollary of the case-or-controversy requirement is that while federal courts can judge the constitutionality of statutes, they may do so only to the extent necessary to resolve the case at hand. “It is emphatically the province and duty of the judicial department to say what the law is,” but only because “[t]hose who apply the rule to particular cases, must of necessity expound and interpret that rule.” Marbury v. Madison, 1 Cranch 137, 177 (1803); see Liverpool, New York & Philadelphia S. S. Co. v. Commissioners of Emigration, 113 U.S. 33, 39 (1885) (“[The Court] has no jurisdiction to pronounce any statute . . . irreconcilable with the Constitution, except as it is called upon to adjudge the legal rights of litigants in actual controversies”). Accordingly, “[e]xcept when necessary” to resolve a case or controversy, “courts have no charter to review and revise legislative and executive action.” Summers v. Earth Island Institute, 555 U.S. 488, 492 (2009); see United States v. Raines, 362 U.S. 17, 20–21 (1960).

These limitations on the power of judicial review play an essential role in preserving our constitutional structure. Our Constitution sets forth a “tripartite allocation of power,” separating different types of powers across three co-equal branches. DaimlerChrysler Corp. v. Cuno, 547 U.S. 332, 341 (2006) (internal quotation marks omitted). “[E]ach branch [is vested] with an exclusive form of power,” and “no branch can encroach upon the powers confided to the others.” Patchak v. Zinke, 583 U.S. 244, 250 (2018) (plurality opinion) (internal quotation marks omitted). In the Judicial Branch’s case, it is vested with the “ultimate and supreme” power of judicial review. Chicago & Grand Trunk R. Co. v. Wellman, 143 U.S. 339, 345 (1892). That power includes the authority to refuse to apply a statute enacted and approved by the other two branches of the Federal Government. But, the power of judicial review can be wielded only in specific circumstances and to limited ends—to resolve cases and controversies. Without that limitation, the Judiciary would have an unchecked ability to enjoin duly enacted statutes. Respecting the case-or-controversy requirement is therefore necessary to “preven[t] the Federal Judiciary from intruding upon the powers given to the other branches, and confin[e] the federal courts to a properly judicial role.” Town of Chester v. Laroe Estates, Inc., 581 U.S. 433, 438 (2017) (internal quotation marks and alteration omitted).

2

Facial challenges conflict with Article III’s case-or-controversy requirement because they ask a federal court to decide whether a statute might conflict with the Constitution in cases that are not before the court.

To bring a facial challenge under our precedents, a plaintiff must ordinarily “establish that no set of circumstances exists under which the Act would be valid.” United States v. Salerno, 481 U.S. 739, 745 (1987). In the First Amendment context, we have sometimes applied an even looser standard, called the overbreadth doctrine. The overbreadth doctrine requires a plaintiff to establish only that a statute “prohibits a substantial amount of protected speech,” “relative to [its] plainly legitimate sweep.” United States v. Williams, 553 U.S. 285, 292 (2008).

Facial challenges ask courts to issue holdings that are rarely, if ever, required to resolve a single case or controversy. The only way a plaintiff gets into a federal court is by showing that he “personally has suffered some actual or threatened injury as a result of the putatively illegal conduct of the defendant.” Blum v. Yaretsky, 457 U.S. 991, 999 (1982) (internal quotation marks omitted). And, the only remedy a plaintiff should leave a federal court with is one “limited to the inadequacy that produced the injury in fact that the plaintiff has established.” Lewis v. Casey, 518 U.S. 343, 357 (1996). Accordingly, once a court decides whether a statute can be validly enforced against the plaintiff who challenges it, that case or controversy is resolved. Either the court remedies the plaintiff ’s injury, or it determines that the statute may be constitutionally applied to the plaintiff.

Proceeding to decide the merits of possible constitutional challenges that could be brought by other plaintiffs is not necessary to resolve that case. Instead, any holding with respect to potential future plaintiffs would be “no more than an advisory opinion—which a federal court should never issue at all, and especially should not issue with regard to a constitutional question, as to which we seek to avoid even nonadvisory opinions.” Chicago v. Morales, 527 U.S. 41, 77 (1999) (Scalia, J., dissenting) (citation omitted).

Unsurprisingly, facial challenges are at odds with doctrines enforcing the case-or-controversy requirement. Pursuant to standing doctrine, for example, a plaintiff can maintain a suit in a federal court—and thus invoke judicial power—only if he has suffered an “injury” with a “traceable connection” to the “complained-of conduct of the defendant.” Steel Co., 523 U. S., at 103. Facial challenges significantly relax those rules. Start with the injury requirement. Facial challenges allow a plaintiff to challenge applications of a statute that have not injured him. But see Acheson Hotels, LLC v. Laufer, 601 U.S. 1, 10 (2023) (Thomas, J., concurring in judgment) (“To have standing, a plaintiff must assert a violation of his [own] rights”). In fact, under our First Amendment overbreadth doctrine, a plaintiff need not be injured at all; he can challenge a statute that lawfully applies to him so long as it would be unlawful to enforce it against others. See United States v. Hansen, 599 U.S. 762, 769 (2023).

Facial challenges also distort standing doctrine’s redressability requirement. The Court has held that a plaintiff has standing to sue only when his “requested relief will redress the alleged injury.” Steel Co., 523 U. S., at 103. With a facial challenge, however, a plaintiff seeks to enjoin every application of a statute—including ones that have nothing to do with his injury. A plaintiff can ask, “Do [I] just want [the court] to say that this statute cannot constitutionally be applied to [me] in this case, or do [I] want to go for broke and try to get the statute pronounced void in all its applications?” Morales, 527 U. S., at 77 (opinion of Scalia, J.). In this sense, the remedy sought by a facial challenge is akin to a universal injunction—a practice that is itself “inconsistent with longstanding limits on equitable relief and the power of Article III courts.” Trump v. Hawaii, 585 U.S. 667, 713 (2018) (Thomas, J., concurring); see Department of Homeland Security v. New York, 589 U. S. ___, ___–___ (2020) (Gorsuch, J., concurring in grant of stay) (slip op., at 2–3); FDA v. Alliance for Hippocratic Medicine, 602 U.S. 367, 402 (2024) (Thomas, J., concurring).

Because deciding the constitutionality of a statute as applied to nonparties is not necessary to resolve a case or controversy, it is beyond a federal court’s constitutional authority. Federal courts have “no power per se to review and annul acts of Congress on the ground that they are unconstitutional. That question may be considered only when the justification for some direct injury suffered or threatened, presenting a justiciable issue, is made to rest upon such an act.” Massachusetts v. Mellon, 262 U.S. 447, 488 (1923). Resolving facial challenges thus violates Article III.[1]

3

Adjudicating facial challenges also intrudes upon powers reserved to the Legislative and Executive Branches and the States. When a federal court decides an issue unnecessary for resolving a case or controversy, the Judiciary assumes authority beyond what the Constitution granted. Supra, at 5–6. That necessarily alters the balance of powers: When one branch exceeds its vested power, it becomes stronger relative to the other branches. See Free Enterprise Fund v. Public Company Accounting Oversight Bd., 561 U.S. 477, 500 (2010).

Moreover, by exceeding their Article III powers, federal courts risk interfering with the executive and legislative functions. Facial challenges enable federal courts to review the constitutionality of a statute in many or all of its applications—often before the statute has even been enforced. In practice, this provides federal courts a “general veto power . . . upon the legislation of Congress.” Muskrat, 219 U. S., at 357. But, the Judicial Branch has no such constitutional role in lawmaking. When courts take on the supervisory role of judging statutes in the abstract, they thus “assume a position of authority over the governmental acts of another and co-equal department, an authority which plainly [they] do not possess.” Mellon, 262 U. S., at 489.

Comparing the effects of as-applied challenges and facial challenges makes this point clear. With an as-applied challenge, the Judiciary intrudes only as much as necessary on the will “ ‘of the elected representatives of the people.’ ” Washington State Grange v. Washington State Republican Party, 552 U.S. 442, 451 (2008). Assuming a court adheres to traditional remedial limits, a successful as-applied challenge only prevents application of the statute against that plaintiff. The Executive Branch remains free to enforce the statute in all of its other applications. And, the court’s decision provides some notice to the political branches, enabling the Executive Branch to tailor future enforcement of the statute to avoid violating the Constitution or Congress to amend the statute.

Facial challenges, however, force the Judiciary to take a maximalist approach. A single plaintiff can immediately call upon a federal court to declare an entire statute unconstitutional, even before it has been applied to him. The political branches have no opportunity to correct course, making legislation an all-or-nothing proposition. The end result is that “the democratic process” is “short circuit[ed]” and “laws embodying the will of the people [are prevented] from being implemented in a manner consistent with the Constitution.” Ibid.

In a similar vein, facial challenges distort the relationship between the Federal Government and the States. The Constitution “establishes a system of dual sovereignty between the States and the Federal Government.” Gregory v. Ashcroft, 501 U.S. 452, 457 (1991). The States retain all powers “not delegated” to the Federal Government and not “prohibited by [the Constitution] to the States.” Amdt. 10. Facial challenges can upset this division by shifting power from the States to the Federal Judiciary. Most obviously, when a state law is challenged, a facial challenge prevents that State from applying its own statute in a constitutional manner. But, facial challenges can also force federal courts to appropriate the role of state courts. To analyze whether a statute is valid on its face, a court must determine the statute’s scope. If a state court has yet to determine the scope of its statute (a common occurrence with facial challenges), the federal court must do so in the first instance. Facial challenges thus increase the likelihood that federal courts must interpret novel state-law questions—a role typically and appropriately reserved for state courts.

B

In addition to their constitutional infirmities, facial challenges also create practical problems. The case-or-controversy requirement serves as the foundation of our adversarial system. Rather than “ ‘sit[ting] as self-directed boards of legal inquiry and research,’ ” federal courts serve as “ ‘arbiters of legal questions presented and argued by the parties before them.’ ” NASA v. Nelson, 562 U.S. 134, 147, n. 10 (2011) (quoting Carducci v. Regan, 714 F.2d 171, 177 (CADC 1983) (opinion for the court by Scalia, J.)). This system “assure[s] that the legal questions presented to the court will be resolved . . . in a concrete factual context conducive to a realistic appreciation of the consequences of judicial action.” Valley Forge Christian College v. Americans United for Separation of Church and State, Inc., 454 U.S. 464, 472 (1982).

Facial challenges disrupt the adversarial system and increase the risk of judicial error as a result. A plaintiff raising a facial challenge need not have any direct knowledge of how the statute applies to others. In fact, since a facial challenge may be brought before a statute has been enforced against anyone, a plaintiff often can only guess how the statute operates—even in his own case. For this reason, “[c]laims of facial invalidity often rest on speculation,” Washington State Grange, 552 U. S., at 450, and “factually barebones records,” Sabri v. United States, 541 U.S. 600, 609 (2004). Federal courts are often called to give “premature interpretations of statutes in areas where their constitutional application might be cloudy.” Raines, 362 U. S., at 22. In short, facial challenges ask courts to resolve potentially thorny constitutional questions with little factual background and briefing by a party who may not be affected by the outcome.

C

The problems with facial challenges are particularly evident in the two cases before us. Even though the trade associations challenge two state laws, the state actors have been left out of the picture. State officials had no opportunity to tailor the laws’ enforcement. Nor could the legislatures amend the statutes before they were preliminarily enjoined. In addition, neither set of state courts had a chance to interpret their own State’s law or “accord [that] law a limiting construction to avoid constitutional questions.” Washington State Grange, 552 U. S., at 450. Instead, federal courts construed these novel state laws in the first instance. And, they did so with little factual record to assist them. The trade associations’ reliance on our questionable associational-standing doctrine is partially to blame.[2] But, the fact that the trade associations raise facial challenges has undeniably played a significant role. With even simple fact patterns, a court has little chance of determining whether a novel, never-before-enforced state law can be constitutionally enforced against nonparties without resorting to mere speculation. For cases such as these, where the constitutional analysis depends on complex, fact-specific questions, the task becomes impossible.

D

Facial challenges are particularly suspect given their origins. They appear to be the product of two doctrines that are themselves constitutionally questionable, vagueness and overbreadth.

At the time of the founding, it was well understood that federal courts could hold a statute unconstitutional only insofar as necessary to resolve a particular case or controversy. See supra, at 5–6. The Founders were certainly familiar with alternative systems that provided for the free-floating review of duly enacted statutes. For example, the New York Constitution of 1777 created a Council of Revision, composed of the Governor, Chancellor, and New York Supreme Court. See Hansen, 599 U. S., at 786 (Thomas, J., concurring). The Council of Revision could object to “any measure of a [prospective] bill” based on “not only [its] constitutionality . . . but also [its] policy.” Id., at 787. If the Council lodged an objection, the Legislature’s only options were to “conform to [the Council’s] objections, override them by a two-thirds vote of both Houses, or simply let the bill die.” Ibid. (internal quotation marks omitted).

In our Constitution, the Founders refused to create a council of revision or involve the Federal Judiciary in the business of reviewing statutes in the abstract. “Despite the support of respected delegates . . . the Convention voted against creating a federal council of revision on four different occasions. No other proposal was considered and rejected so many times.” Id., at 789 (citation omitted). Instead, the Founders created a Judiciary with “only the authority to resolve private disputes between particular parties, rather than matters affecting the general public.” Ibid. (internal quotation marks omitted). They considered judges “of all men the most unfit to have a veto on laws before their enactment.” Ibid. (internal quotation marks omitted). Therefore, they refused to enlist judges in the business of reviewing statutes other than “as an issue for decision in a concrete case or controversy.”[3] Ibid.

For more than a century following the founding, the Court generally adhered to the original understanding of the narrow scope of judicial review. When the Court first discussed the concept of judicial review in Marbury v. Madison, it made clear that such review is limited to what is necessary for resolving “a particular cas[e]” before a court. 1 Cranch, at 177; see also supra, at 5–6. And, in case after case that followed Marbury, the Court reiterated that federal courts have no authority to reach beyond the parties before them to facially invalidate a statute.[4]

As best I can tell, the Court’s first departure from those principles was the development of the vagueness doctrine. See Johnson v. United States, 576 U.S. 591, 616–620 (2015) (Thomas, J., concurring in judgment) (describing history of vagueness doctrine). Before and at the time of the founding, American and English courts dealt with vague laws by “simply refus[ing] to apply them in individual cases.” Id., at 615. After the unfortunate rise of “substantive” due process, however, American courts began striking down statutes wholesale as “unconstitutionally indefinite.” Id., at 617. This Court first adopted that approach in 1914, see International Harvester Co. of America v. Kentucky, 234 U.S. 216, and has since repeatedly used the vagueness doctrine “to strike down democratically enacted laws” in the name of substantive due process, Sessions v. Dimaya, 584 U.S. 148, 210 (2018) (Thomas, J., dissenting); see Johnson, 576 U. S., at 618–621 (opinion of Thomas, J.). As I have explained, I doubt that “our practice of striking down statutes as unconstitutionally vague is consistent with the original meaning of the Due Process Clause.” Dimaya, 584 U. S., at 206 (opinion of Thomas, J.); see Johnson, 576 U. S., at 622 (opinion of Thomas, J.).

The vagueness doctrine was the direct ancestor of one subset of modern facial challenges, the overbreadth doctrine. See United States v. Sineneng-Smith, 590 U.S. 371, 385 (2020) (Thomas, J., concurring) (noting that the overbreadth doctrine “developed as a result of the vagueness doctrine’s application in the First Amendment context”). In Thornhill v. Alabama, 310 U.S. 88 (1940), the Court deemed an antipicketing statute “invalid on its face” due to its “sweeping proscription of freedom of discussion.” Id., at 101–106. The Thornhill Court did so “[w]ithout considering whether the defendant’s actual conduct was entitled to First Amendment protection,” instead invalidating the law because it “ ‘swept within its ambit . . . activities that in ordinary circumstances constitute an exercise of freedom of speech or of the press.’ ” Sineneng-Smith, 590 U. S., at 383 (opinion of Thomas, J.) (quoting Thornhill, 310 U. S., at 97; alteration omitted).

Thornhill’s approach quickly gained traction in the First Amendment context. In the years to follow, the Court “invoked [its] rationale to facially invalidate a wide range of laws” concerning First Amendment rights—a practice that became known as the overbreadth doctrine. Sineneng-Smith, 590 U. S., at 383. Under that doctrine, a court can invalidate a statute if it “prohibits a substantial amount of protected speech,” “relative to the statute’s plainly legitimate sweep.”[5] Williams, 553 U. S., at 292. The Court has never attempted to ground the overbreadth doctrine “in the text or history of the First Amendment.” Sineneng-Smith, 590 U. S., at 384 (opinion of Thomas, J.). Instead, the Court has supplied only “policy considerations and value judgments.” Ibid.

The overbreadth and vagueness doctrines’ method of facial invalidation eventually spread to other areas of law, setting in motion our modern facial challenge doctrine. For several decades after Thornhill, the Court continued to resist the broad use of facial challenges. For example, in Broadrick v. Oklahoma, 413 U.S. 601 (1973), the Court emphasized that “[c]onstitutional judgments, as Mr. Chief Justice Marshall recognized, are justified only out of the necessity of adjudicating rights in particular cases between the litigants brought before the Court.” Id., at 611. In that vein, the Court characterized “facial overbreadth adjudication [as] an exception to our traditional rules of practice.” Id., at 615. But, the Court eventually entertained facial challenges more broadly where a plaintiff established that “no set of circumstances exists under which the Act would be valid.” [6] Salerno, 481 U. S., at 745. Just as with the overbreadth doctrine, the Court has yet to explain how facial challenges are consistent with the Constitution’s text or history.

Given how our facial challenge doctrine seems to have developed—with one doctrinal mistake leading to another—it is no wonder that facial challenges create a host of constitutional and practical issues. See supra, at 6–13. Rather than perpetuate our mistakes, the Court should end them. “No principle is more fundamental to the judiciary’s proper role in our system of government than the constitutional limitation of federal-court jurisdiction to actual cases or controversies.” Simon v. Eastern Ky. Welfare Rights Organization, 426 U.S. 26, 37 (1976). Because that requirement precludes courts from judging and enjoining statutes as applied to nonparties, the Court should discontinue the practice of facial challenges.

* �ă* �ă*

The Court has recognized the problems that facial challenges pose, emphasizing that they are “disfavored,” Washington State Grange, 552 U. S., at 450, and “best when infrequent,” Sabri, 541 U. S., at 608. The Court reiterates those sentiments today. Ante, at 9, 30. But, while sidelining facial challenges provides some measure of relief, it ignores the real problem. Because federal courts are bound by Article III’s case-or-controversy requirement, holding a statute unconstitutional as applied to nonparties is not simply disfavored—it exceeds the authority granted to federal courts. It is high time the Court reconsiders its facial challenge doctrine.

Notes

[1]  This is not to say that federal courts can never adjudicate a constitutional claim if a plaintiff styles it as a facial challenge. Whenever a plaintiff alleges a statute is unconstitutional in many or all of its applications, that argument nearly always includes an allegation that the statute is unconstitutional as applied to the plaintiff. Federal courts are free to consider challenged statutes as applied to the plaintiff before them and limit any relief accordingly. See generally Americans for Prosperity Foundation v. Bonta, 594 U.S. 595, 618–619 (2021); id., at 621 (Thomas, J., concurring in part and concurring in judgment).

[2]  The trade associations do not allege that they are subject to H. B. 20 and S. B. 7072, but have brought suit to vindicate the rights of their members. There is thus not a single party in these suits that is actually regulated by the challenged statutes and can explain how specific provisions will infringe on their First Amendment rights. Instead, the trade associations assert their understanding of how the challenged statutes will regulate nonparties. As I have recently explained, “[a]ssociational standing raises constitutional concerns.” See FDA v. Alliance for Hippocratic Medicine, 602 U.S. 367, 399 (2024) (concurring opinion). Associational standing appears to conflict with Article III’s injury and redressability requirements in many of the same ways as facial challenges. I have serious doubts that either trade association has standing to vicariously assert a member’s injury. See id., at 400.

[3]  “The later history of the New York Council of Revision demonstrates the wisdom of the Framers’ decision.” United States v. Hansen, 599 U.S. 762, 790 (2023) (Thomas, J., concurring). The Council’s ability to lodge objections proved significant: “Over the course of its existence, [the Council] returned 169 bills to the legislature; the legislature, in turn, overrode only 51 of those vetoes and reenacted at least 26 bills with modifications.” Ibid. The Council did not shy away from controversial or weighty matters either. It vetoed, among other things, “a bill barring those convicted of adultery from remarrying” and a bill “declar[ing] Loyalists aliens.” Ibid. In fact, the bill authorizing the Erie Canal’s construction—“one of the most important measures in the Nation’s history—survived the Council’s review only because Chancellor James Kent changed his deciding vote at the last minute, seemingly on a whim.” Ibid. Concerns over the Council’s “intrusive involvement in the legislative process” eventually led to its abolition in 1820. Ibid.

[4]  See, e.g., Austin v. Aldermen, 7 Wall. 694, 699 (1869) (holding that the Court could “only consider the statute in connection with the case before” it and thus “our jurisdiction [wa]s at an end” once it “ascertained that [the case] wrought no effect which the act forbids”); Liverpool, New York & Philadelphia S. S. Co. v. Commissioners of Emigration, 113 U.S. 33, 39 (1885) (the Court “has no jurisdiction to pronounce any statute . . . irreconcilable with the Constitution, except as it is called upon to adjudge the legal rights of litigants in actual controversies”); Chicago & Grand Trunk R. Co. v. Wellman, 143 U.S. 339, 345 (1892) (explaining that judicial review of a statute’s constitutionality “is legitimate only in the last resort, and as a necessity in the determination of real, earnest, and vital controversy between individuals”); Muskrat v. United States, 219 U.S. 346, 357 (1911) (“[T]here [i]s no general veto power in the court upon the legislation of Congress”); Yazoo & Mississippi Valley R. Co. v. Jackson Vinegar Co., 226 U.S. 217, 219 (1912) (rejecting argument that statute was “void in toto,” because the Court “must deal with the case in hand and not with imaginary ones”); Dahnke-Walker Milling Co. v. Bondurant, 257 U.S. 282, 289 (1921) (“[A] litigant can be heard to question a statute’s validity only when and so far as it is being or is about to be applied to his disadvantage”); Massachusetts v. Mellon, 262 U.S. 447, 488 (1923) (Federal courts “have no power per se to review and annul acts of Congress on the ground that they are unconstitutional. That question may be considered only when the justification for some direct injury suffered or threatened, presenting a justiciable issue, is made to rest upon such an act”).

[5]  Although the Court’s precedents describe an unconstitutionally overbroad statute as facially “invalid,” “federal courts have no authority to erase a duly enacted law from the statute books.” J. Mitchell, The Writ-of-Erasure Fallacy, 104 Va. L. Rev. 933, 936 (2018); see Sineneng-Smith, 590 U. S., at 387 (opinion of Thomas, J.).

[6]  Some Members of the Court subsequently sought to apply a more lenient standard to all facial challenges. See Washington State Grange v. Washington State Republican Party, 552 U.S. 442, 449 (2008) (noting that “some Members of the Court have criticized the Salerno �ڴǰ����ܱ����پ��Dz�”); United States v. Stevens, 559 U.S. 460, 472 (2010) (reserving the question of which standard applies to “a typical facial attack”).