IN RE: SOCIAL MEDIA ADOLESCENT ADDICTION/PERSONAL INJURY PRODUCTS LIABILITY LITIGATION (2026)

ORDER DENYING DEFENDANT'S MOTION FOR SUMMARY JUDGMENT AND DENYING IN PART AG PLAINTIFFS' CROSS-MOTION FOR PARTIAL SUMMARY JUDGMENT

Before the Court are (i) defendant Meta Platforms, Inc.'s (“Meta”) motion for summary judgment on the state Attorneys General (“AG”) plaintiffs' claims for deception, unfairness, and violations of the Children's Online Privacy and Protection Act (“COPPA”) (Dkt. No. 2704 [“MSJ”]) and (ii) the AG plaintiffs' cross-motion for partial summary judgment regarding certain elements of the AGs' deception and COPPA claims (Dkt. No. 2695 [“Partial MSJ”]).

The Court has carefully considered the parties' briefs, evidence, and argument at the April 15, 2026 hearing and, as explained below, DENIES defendant's motion for summary judgment and DENIES IN PART plaintiffs' cross-motion for summary judgment.¹ In this regard, the Court finds that material disputes of fact exist as to whether Meta is required to comply with COPPA. That said, no dispute exists that Meta has not, in fact, complied with COPPA's notice and parental consent requirements which is consistent with its position. As to this element, summary judgment is GRANTED.

I. INTRODUCTION

A. MOTIONS TO SEAL

As a preliminary matter, both sides have submitted numerous motions to seal documents or portions of documents offered in support of their summary judgment briefing. At summary judgment, materials may be sealed if there are “compelling reasons” to do so. Ctr. for Auto Safety v. Chrysler Grp., LLC, 809 F.3d 1092, 1096-97 (9th Cir. 2016). The Court finds that many of the sealing requests here based upon “confidential” or “proprietary” information are overbroad and compelling reasons have not been established to seal certain documents to the extent requested. At the June 26, 2026 hearing, the Court warned that it would not entertain requests to seal documents without a specifically identified compelling reason. Therefore, regarding all materials referenced in this Order sought to be sealed on the basis of confidential or proprietary information, the Court DENIES the motions to seal. The Court will address the remaining sealing disputes at a later date.²

B. PROCEDURAL BACKGROUND

As discussed in the Court's prior order on defendant's motion to dismiss, the AG plaintiffs assert that Meta violated their consumer protection laws by (i) designing and deploying platform features it knew were harmful to young users, and (ii) misleading and concealing from the public its knowledge of this harm. (Dkt. No. 1214 [“MTD Order”]). The AGs also contend that Meta collects and uses personal information from children under 13 years old (“U13s”) without complying with COPPA's notice and consent requirements. There, the Court held that Meta's design, development, and deployment of certain product features plausibly constituted unfair or unconscionable practices, but that Section 230 posed a limitation on the AGs' claims. Given the scope of the briefing, the Court limited claims such that they could only be based on the following features: appearance-altering features, features related to restricting time spent on the platform, and Instagram's “multiple accounts” function. (MTD Order at 29.)

C. FACTUAL OVERVIEW

The Court discussed the background of this case in detail in its October 15, 2024 Motion to Dismiss Order. Plaintiffs are the AGs of twenty-nine states (Arizona, California, Colorado, Connecticut, Delaware, Hawai‘i, Idaho, Illinois, Indiana, Kansas, Kentucky, Louisiana, Maine, Maryland, Minnesota, Nebraska, New Jersey, New York, North Carolina, Ohio, Oregon, Pennsylvania, Rhode Island, South Carolina, South Dakota, Virginia, Washington, West Virginia, and Wisconsin.) Through a consolidated complaint, they have sued Meta based on conduct relating to its Facebook and Instagram social media platforms.

As discussed below, the evidence presented reveals that the record contains numerous disputes of fact. The Court provides here an overview to the extent it is relevant for the purposes of this motion.

1. Alleged Design and Deployment of Harmful Platform Targeting Young Users

This Court held at the motion to dismiss stage that the “alleged exchange of users' use of Meta's platforms for their personal data is Meta's ‘primary trade or commerce,’ at least insofar as its Facebook and Instagram platforms are concerned.” Meta's business model focuses on monetizing user information and attention on its platforms. Meta then converts this user engagement into revenue obtained from advertisers, who can deploy targeted advertising based on the personal data Meta collects from each user. In 2024, “substantially all” of Meta's revenue was generated from marketers advertising to users on Facebook and Instagram. (Dkt. No. 2785, Exs. 63 & 64, Deposition of Susan Li ([“Li Dep.”]), 57:4–58:6; 96:10–97:2; 88-96, 100.) Meta generates revenue “by showing ads on [certain] parts of the family of apps [Facebook, Instagram, WhatsApp, Messenger, and Threads].” (Id. at 41:23–45:12.) To maintain revenue generation, Meta must maintain or increase users and engagement on Facebook and Instagram. (Id. at 42:23–45:12, 138:7–142:12; Dkt. No. 2785-17 (email in which CFO asks for updates on goals related to teens, including goal of millions of teen users).)

Meta's platforms contain a significant teen user base. Indeed, millions of teens use the platforms between midnight and four a.m. on a weekly basis. (Dkt. No. 2787-6.) Now that discovery is over, the parties dispute whether Meta's platforms are, or are designed to be, addictive. Evidence exists that Meta employees have authored internal documents indicating that the platforms are addictive and that teens interact with them in an addictive manner. (See, e.g., Dkt. No. 2783-17 (“It seems clear from what's presented here that some of our users are addicted to our products. And I worry that driving sessions [of use] incentivizes us to make our product more addictive.”); Dkt. No. 2783-13 (“app addiction is common on IG”); Dkt. No. 2783-18 (“It seems relatively clear that younger people are less equipped to be able to handle social media addictions.”))

Internal Meta documents also show that at least some employees believed the platforms cultivated compulsive use to maximize engagement. (See, e.g. Dkt. No. 2784-10) (“our product exploits weaknesses in the human psychology to promote product engagement and time spent [on the platforms]”); Dkt. No. 2784-11 at 67:5–17 (discussing “the mechanisms by which harm was reproduced,” including “clicking, liking, [and] sharing,” “given that the whole News Feed system is aimed at engagement”); Dkt. No. 2784-12 at 152–53 (“people are binging on IG so much they can't feel reward anymore ․ Its [sic] biological and psychological. the top down directives drive it all towards making sure people keep coming back for more.”.)

2. Alleged Deceptive Statements

Evidence further indicates that Meta executives and other high-ranked employees were included in email and other conversations discussing that the platforms may induce compulsive use by teens. (See, e.g., Dkt. No. 2783-2, (2019 email to Zuckerberg, with Mosseri and others copied, regarding statistics about problematic use and “increasing scientific evidence” of negative effect of Facebook on well-being); id., Dkt. No. 2785-6, (2019 email sent to Zuckerberg with research regarding problematic use of Meta's platforms); Dkt. No. 2785-1, Deposition of Mark Zuckerberg [“Zuckerberg Dep.”] 124:21–24 (Q: “Meta knew that problematic use is a real issue for a meaningful portion of its users, correct?” A: “Yes.”); id., Dkt. No. 2784-12 (“IG is a drug․ Adam [Mosseri] doesn't want to hear it –he freaked out when I talked about dopamine in my teen fundamentals leads review but it's undeniable!”); id., Dkt. No. 2785-13 (email from Davis describing “issues we are seeing crop up, such as around addiction,” in context of “youth issues” and an “integrated youth plan”).)

The AGs assert that, while developing its platform to be addictive, Meta affirmatively misrepresented to the public the contrary. That is, it messaged the safety of its platforms despite possessing knowledge of harms to young users.

As noted above, the AGs failed to solidify the statements on which it was proceeding prior to briefing on this motion. The statements referenced in the parties' briefing and discussed herein are:

Meta's statement to the BBC in 2018 that “at no stage does wanting something to be addictive factor into” Meta's platform design process. (Dkt. No. 2705-73 at 4.)

In 2021, Adam Mosseri, Head of Instagram, stated to Congress: “I don't believe the research suggests that our products are addictive.” (Dkt. No. 2705-6 at 13.)

Also in 2021, Facebook Global Head of Safety Antigone Davis' told Congress: “I disagree with calling our products addictive.” (Dkt. No. 2705-4 at 38.)

In the same committee hearing, a senator subsequently asked Davis “to drill down on this addictive element ․ isn't part of your business model to have more eyeballs for a longer amount of time engaged using your services?” to which Davis responded: “Respectfully, Senator that is not actually how we build our products.” (Id. at 39.)

3. Alleged COPPA-Related Practices

The parties have stipulated that Instagram and Facebook are “websites” and “online services” as those terms are used in COPPA. Meta collects a wide array of data from its users, including email addresses, phone numbers, images of a user's face, user's voice, cookie IDs, IP addresses, profile information, and information reflecting a user's interactions on the platform. (See Dkt. No. 2696-4, Deposition of Allison L. Hartnett ([“Hartnett Dep.”]).) Meta has collected at least some of this type of data from Instagram and Facebook users since 2012. (Id. at 30:1-4.) Meta collects various types of information belonging to logged-off users, including IP addresses, device IDs, cookie data, and user clicks. (Id. at 59:13-60:18.)

Meta has engaged in numerous practices to detect and remove U13 users, including: identifying children who self-report a date of birth under the age of 13 by changing the date of birth on their account after signup (id. at 103:5-7); automatically suspending (or “checkpointing”) the accounts of Facebook users who attempted to change their date of birth to reflect an age under 13 (See Dkt. No. 2696-7); reviewing reports from users and human review to detect U13 users. (See Hartnett Dep. at 101:10-14); and through “hard-linking” or “soft-matching” other accounts already identified as belonging to the same child (id. at 110:8-17).

II. LEGAL STANDARD

Summary judgment is appropriate when “there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P. 56(a). In reviewing summary judgment motions, courts must view all evidence in the light most favorable to the non-moving party and draw all justified inferences on its behalf. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986); see also Walls v. Cent. Contra Costa Transit Auth., 653 F.3d 963, 966 (9th Cir. 2011) (courts must “draw all inferences supported by the evidence in favor of the non-moving party.”). A court may grant summary judgment on a “part of each claim or defense” to eliminate questions for trial where there is no genuine dispute of fact. Fed. R. Civ. P. 56(a). Additionally, in an MDL, the transferee court applies the law of its circuit to issues of federal law, but on issues of state law it applies the state law that would have been applied to the underlying case as if it had never been transferred into the MDL. In re Anthem, Inc. Data Breach Litig., 2015 WL 5286992, at *2 (N.D. Cal. Sept. 9, 2015) (collecting Ninth Circuit cases).

III. ANALYSIS

Meta moves for summary judgment on all of the AGs' claims, namely those based on (A) deception, (B) unfairness, (C) disgorgement as a remedy, and (D) COPPA. The Court takes each in turn.

A. DECEPTION CLAIMS

The AGs have proffered evidence of more than one hundred purportedly deceptive statements. As a preliminary matter, the Court notes that as long as one actionable misrepresentation presents a triable issue—that is, a genuine dispute of material fact—summary judgment on the AGs' deception claims as a whole is denied. Also, as noted in the MTD Order, “each statement must be considered in the context of the alleged deceptive scheme as a whole, and so the Court is unwilling to parse and isolate every alleged misrepresentation as Meta requests.” (MTD Order at 37.)³

Meta asserts seven independent grounds ⁴ for summary judgment on the deception claims, namely:

One, that statements regarding the addictiveness of its platforms are not false, because social media addiction does not exist and plaintiffs cannot show that Meta's platforms cause it.

Two, that Meta's statements are not knowingly false, because the science of social media addiction is at best uncertain and plaintiffs lack evidence that the speakers knew otherwise.

Three, that plaintiffs lack evidence to show that the statements were misleading.

Four, that certain statements are immunized by the Noerr-Pennington doctrine because they were made to government actors in the context of petitioning activity.

Five, certain statements were not made in the course of commerce or pursuant to a commercial transaction, as required by state laws.

Six, certain “opinion” statements are protected by the First Amendment.

Seven, certain statements were inactionable “puffery” or statements of aspiration. The Court addresses each argument.

1. Falsity

Meta first contends that statements regarding Meta's platforms' addictive nature were not actually false.⁵ The parties contest whether falsity is in fact required under the pertinent state laws. Meta contends that plaintiffs' evidence must show that the statements were, in “full context,” false or untrue when made. See Catholdi-Jankowski v. CVS Health Corp., 656 F.Supp.3d 412, 428 (W.D.N.Y. 2023). The AGs contend that at least certain states do not require falsity and that deceptive statements can be merely “misleading” or have “a capacity, likelihood or tendency to deceive or confuse the public”. Moore v. Mars Petcare US, Inc., 966 F.3d 1007, 1017 (9th Cir. 2020) (citations omitted). Regardless, for the reasons discussed below, the Court finds a genuine dispute of material fact as to the falsity of Meta's statements reflecting the proposition that its platforms are not designed to be addictive.

Meta urges the Court to conclude that “social media addiction” is not an accepted diagnosis, and thus statements that its platforms are not addictive cannot be false. Specifically, Meta hinges the argument on the ground that the Diagnostic and Statistical Manual of Mental Disorders (“DSM-5,” a manual published by the American Psychiatric Association) and the International Classification of Diseases 11th Edition (“ICD-11”) do not list “social media addiction” as a disorder. This Court has already held that the absence of the term “social media addiction” from those publications is not dispositive in this context, in part because more than a decade of research on social media has passed since the DSM was published in 2013, and more recent statements from the American Psychiatric Association have in fact defined “social media addiction,” suggesting that forthcoming publications may include it. (See Dkt. No. 2857 [“GC Experts Order”].)

Additionally, the Court is not persuaded that Meta's statements could be false only if “social media addiction” is recognized as a diagnosable mental health disorder. Statements discussed in the briefing ⁶ deny in general terms that the platforms are addictive ⁷ without reference to a particular diagnosis. The AGs present a reasonable interpretation of those statements that Facebook and Instagram are not designed in ways that cause teens to compulsively use the platforms to their detriment. To the extent plaintiffs' evidence shows that the platforms are in fact designed to do just that, a jury could reasonably find the statements were untrue to a reasonable person.

Further, the AG plaintiffs cite Meta's own internal documents to support their theory that Meta's platforms cause addictive use by teens. (See, e.g., Dkt. No. 2783-13 at 37 (2019 presentation titled “Instagram Teen Well-Being Study: US Topline Findings” discussing that “addiction is common on IG” and that “[o]lder teen girls are most likely to experience this issue”); Dkt. No. 2783-17 (“It seems clear from what's presented here that some of our users are addicted to our products. And I worry that driving sessions [of platform use] incentivizes us to make our product more addictive”); Dkt. No. 2783-24 (2023 Instagram presentation finding that “Research has shown that problematic social media use is both prevalent among teens and is a driver of churn: 56% of IG teens surveyed say it's difficult to manage time on social media, and teens reporting life interference were 2.1x more likely to delete Instagram.”); Dkt. No. 2783-20 (report noting that a “majority of clinicians believe that social media can be addictive”); Dkt. No. 2783-21 (presentation titled “Teen Mental Health Deep Dive” concluding that “Teens talk about the amount of time they spend on Instagram as one of the ‘worst’ aspects of their relationship to the app,” and noting that those teens “have an addicts' narrative about their use –it can make them feel good, feel bad. They wish they could spend less time caring about it, but they can't help themselves.”).)

Plaintiffs' experts opine that Meta's platforms are designed in ways that cause addiction. For example, Dr. Mitch Prinstein, a clinical psychologist, concluded that “Meta platform features encourage problematic social media use, consistent with clinical dependency,” and interfered with teens' important daily activities. (Dkt. No. 2405-15 at ¶¶ 2, 67.) Likewise, Dr. Bradley Zicherman opines that “[s]ocial media addiction is a serious concern acknowledged by mental health professionals that resembles other addiction disorders and can be measured and diagnosed by adapting addiction criteria published by the American Psychiatric Association,” and that “[s]ocial media use is particularly problematic for teens and youth due to their underdeveloped brains.” (Dkt. No. 2405-11 at ¶¶ 1, 2.)⁸

The Court concludes there is a material dispute of fact as to the existence, qualities, and/or scope of social media addiction as well as whether Meta's statements denying that their platforms are designed to be addictive were false.

Thus, the motion on this ground is DENIED.

2. Knowingly False

Meta contends that because social media addiction is not an established condition, the speakers could not knowingly make a false statement that the platforms are not addictive.⁹ Additionally, Meta argues that to the extent some individuals within Meta believed their products were addictive, that knowledge cannot be imputed to the individual speakers. JTS Choice Enters., Inc. v. E.I. DuPont De Nemours & Co., 2014 WL 793525, at *7 (D. Colo. 2014) (in fraud context, “it is not enough to show that the statement was false based upon all of the knowledge imputed to a corporation.”).

The Court addressed Meta's first argument above. The AGs have offered sufficient evidence to create a material dispute of fact regarding whether the products were designed to be addictive. Supra Section III(A)(1). Regarding the second argument, the Court finds that plaintiffs adduced sufficient evidence that Meta's speakers, including CEO Mark Zuckerberg, Adam Mosseri (Head of Instagram), and Antigone Davis (Meta global head of safety), may have had individualized knowledge regarding Instagram and Facebook's addictiveness. (See, e.g., Dkt. No. 2783-2 (2019 email to Zuckerberg, with Mosseri and others copied, regarding statistics about problematic use and “increasing scientific evidence” of negative effect of Facebook on well-being); Dkt. No. 2785-6, (2019 email sent to Zuckerberg with research regarding problematic use of Meta's platforms); Zuckerberg Dep. at 124:21–24 (Q: “Meta knew that problematic use is a real issue for a meaningful portion of its users, correct?” A: “Yes.”); Dkt. No. 2784-12 (“IG is a drug․ Adam [Mosseri] doesn't want to hear it – he freaked out when I talked about dopamine in my teen fundamentals leads review but it's undeniable!”); Dkt. No. 2785-13 (email from Davis describing “issues we are seeing crop up, such as around addiction,” in context of “youth issues” and an “integrated youth plan”).)

The motion on this ground is DENIED.

3. Misleading

Meta next contends that the AGs lack evidence to prove that the statements had a tendency, or were likely, to mislead consumers, as required by many of the AGs' state laws.¹⁰ A representation that is “obviously misleading” requires no additional evidence, but extrinsic “evidence is necessary where the advertising is not clearly misleading on its face and materiality is in doubt.” Beardsall v. CVS Pharmacy, Inc., 953 F.3d 969, 976 (7th Cir. 2020).

Relying on Weaver v. Champion Petfoods USA Inc., 3 F.4th 927, 936 (7th Cir. 2021), Meta contends that the AGs were required, but have failed, to identify “consumer surveys or market research showing that a reasonable consumer would interpret” the statements in such a way that they would be misleading. Weaver, however, concerned whether a consumer would misunderstand the phrase “biologically appropriate” to mean BPA-free, an inference not clearly misleading on its face. Id. at 935. Here, however, Meta has made statements including (i) “I don't believe the research suggests that our products are addictive,” (ii) “I disagree with calling our products addictive,” and (iii) having “more eyeballs for a longer amount of time engaged” with the platforms is “not actually how we build our products.” See supra at 8, n. 5. These statements unambiguously suggest that the products are not addictive. To the extent Meta's platforms are in fact designed to be addictive, the statements are misleading on their face. As such, the Court is not persuaded that evidence such as market research is required to create a triable issue. Nor is the issue beyond the reasoning of a reasonable juror.

The motion on this ground is DENIED.

4. Noerr-Pennington

Meta argues that Noerr-Pennington immunizes any statements made to government bodies, including to Congress and the Barnet Coroner's Court, a foreign court. “The Noerr-Pennington doctrine protects ‘those who petition any department of the government for redress ․ from statutory liability for their petitioning conduct.’ ” In re Apple Inc. Securities Litigation, 2020 WL 2857397, at *18 (N.D. Cal., 2020) (quoting SOSA v. DIRECTV, Inc., 437 F.3d 923, 929 (9th Cir. 2006)).

To determine whether a defendant's conduct is immunized under Noerr-Pennington, courts consider “(1) whether the lawsuit imposes a burden on petitioning rights, (2) whether the alleged activities constitute protected petitioning activity, and (3) whether the statute at issue may be construed to avoid that burden.” B&G Foods N. Am., Inc. v. Embry, 29 F.4th 527, 535 (9th Cir.) (cleaned up), cert. denied, ––– U.S. ––––, 143 S. Ct. 212, 214 L.Ed.2d 83 (2022). As noted in the MTD Order, Noerr-Pennington immunizes only “protected petitioning activity,” B&G Foods, 29 F.4th at 535, and Meta must have evidence that it “was seeking any redress from Congress that implicates its First Amendment right to petition.” In re Apple Inc. Sec. Litig., 2020 WL 2857397, at *18.

Meta contends that certain statements were protected petitioning activity because, for example, the Congressional hearings were aimed at shaping future legislation or legislative proposals. The AGs argue that Meta offers no evidence that “any particular legislation or government effort was pending” that these challenged statements were intended to influence. Instead, they argue that the hearings had generalized purposes. See id. They also argue that the statements constitute commercial speech and may thus be subjected to “more restrictions.” Ariix, LLC v. NutriSearch Corp., 985 F.3d 1107, 1115 (9th Cir. 2021).

A commercial speech analysis is not needed. The AGs provide evidence that Meta's practice was to amplify the statements made to governmental bodies after the fact. (See, e.g. Dkt. No. 2786-5, (discussing methods to shape public narrative through both consumers and “policy elites”); Dkt. No. 2786-6 at 2 (describing strategies for changing “parent sentiment,” including via “policy elites”); Dkt. No. 2783-16 (stating that Meta expected the story regarding statements to the Barnet Coroner's Court to travel broadly); Dkt. No. 2785-20 (discussing media coverage of Meta executive “giving evidence” in the inquest).) The Court finds that plaintiffs have provided sufficient evidence that these statements were communicated separately to the public at large as part of a series of messaging campaigns. Such statements cannot be immunized by simultaneously repeating them to Congress or other governmental bodies.

The motion on this ground is DENIED.

5. Commerce

Next, Meta seeks summary judgment of the AG plaintiffs' deception claims on the basis that ninety-five ¹¹ of the statements were not made in connection with business, commerce, trade, or consumer transactions, as required by sixteen of eighteen states. Meta asserts that five states' statutes (DE, IN, KS, NJ, and VA) require misrepresentations to be made in connection with a “consumer transaction”; that six states (CO, DE, IL, MN, NE, and NY) require the statements be made “in the course of business”; and that nine states (CT, IL, KY, LA, NC, NE, NY, PA, and SC) require statements to be made in the conduct of “trade or commerce.” Meta contends that certain types of statements, including (i) internal talking points; (ii) statements on earnings calls; (iii) statements made to a foreign court; (iv) statements made to Congress; (v) statements made at technology conferences; and (vi) statements on Meta's website, were not directed at consumers to convince them to use its platforms. Silver v. Pep Boys-Manny, Moe & Jack of Del., Inc., 2018 WL 1535285, at *6 (D.N.J. 2018) (misrepresentation must be “material to the transaction” and “made to induce the buyer”).

This Court has previously analyzed the states' law on this question, holding that while the states each “present various formulations of how to determine whether an activity takes place in ‘trade or commerce’ ” (MTD Order at 69), the various “trade or commerce” definitions are also broadly defined and liberally construed in each state (id. at 68). The Court also held that these variations “do not impact the final analysis” and the Court declines to parse each state's standards again here. (Id. at 70.)

As this Court held at the motion to dismiss stage, the “alleged exchange of users' use of Meta's platforms for their personal data is Meta's ‘primary trade or commerce,’ at least insofar as its Facebook and Instagram platforms are concerned,” and thus that “the alleged unfair and deceptive acts or practices occur “in the conduct of” Meta's “trade or commerce,” regardless of whether Meta's users pay for use of Facebook or Instagram.” (Id. at 71). Nothing has changed to find otherwise.

Meta also contends that plaintiffs have failed to show that Meta's statements about safety and addictiveness were “material to” their consumers' decision to use the platforms, as required by certain states. See, e.g., Silver v. Pep Boys-Manny, Moe & Jack of Del., Inc., 2018 WL 1535285, at *6 (D.N.J. 2018). The AGs counter that they have provided sufficient evidence to show that safety is a significant concern to consumers. Moreover, plaintiffs' expert Alter opines that because of the platforms' highly technical nature, and resulting information asymmetries, consumers are particularly likely to rely on company messaging concerning health and safety.¹² (Dkt. No. 2781-3 [“Alter Rep.”] at 1-2.) As a result, the Court concludes there is a material dispute of fact as to whether the statements were material.

Meta also argues that many of the statements, such as those only made in internal documents, were never received by consumers. Statements made on public earnings calls—such as executive Sheryl Sandberg's claim that Meta has “to keep people safe and give them control over their experience on our apps. And we are[,]” were facially directed not at consumers but only at investors. (Dkt. No. 2705-19 at 5). Meta will have the opportunity at trial to make their case that statements were not sufficiently directed at consumers or otherwise not made in the course of commerce. Sufficient for purposes of this motion, Meta knows that information provided on public earnings are widely reported to consumers. With respect to “internal documents,” without specificity, the motion cannot be granted.

Plaintiffs' cross-motion for summary judgment asks the Court to rule in their favor on the commerce element. Again, as discussed supra at n. 1, the Court will not make legal or factual conclusions regarding the alleged statements because the AGs failed to properly identify the statements in the first instance.

Both motions on this ground are DENIED.

6. First Amendment Opinion Statements

Meta next contends that four 2021 statements of opinion are protected by the First Amendment. Three of those statements are by Meta CEO Mark Zuckerberg:

(1) describing Meta's reporting on the prevalence of harmful content on Facebook as a “model.” (Dkt. No. 2705-5 at 49.)

(2) “Overall, the research that we have seen is that using social apps to connect with other people can have positive mental health benefits and well-being benefits by helping people feel more connected and less lonely.” (Id. at 56.)

(3) that Meta's platforms “help people stay connected to people they care about, which I think is one of the most fundamental and important human things that we do.” (Id. at 100.)

They also cite one statement by Adam Mosseri, Head of Instagram: “I don't believe the research suggests that our products are addictive.” (Dkt. No. 2705-6 at 13.)

As a threshold issue, the First Amendment does not protect misleading commercial speech. Ariix, LLC v. NutriSearch Corporation, 985 F.3d 1107, 1119 (9th Cir. 2021). Commercial speech is “not limited simply to the expectation of a direct commercial transaction with consumers,” but can relate to “improvements to a brand's image” and “general exposure of a product.” Ariix, 985 F.3d at 1117.

With respect to Meta's claims that the statements are related to a matter of public concern—the alleged harm to teens and children caused by social media use—rather than commercial speech, that is an argument, not an undisputed fact. It is equally plausible that these statements, which suggest that Meta's platforms have positive impacts on wellbeing and are good for human connection, relate to allaying consumer concerns about safety, improving brand image, and exposing consumers to positive messages about the platforms. Given plaintiffs' evidence that Meta's business benefits from exposure to those positive public messages, there is a triable issue of fact regarding whether the statements are unprotected commercial speech.

The motion on this ground is DENIED.

7. Puffery

Finally, Meta argues that sixty-two ¹³ of the statements identified by plaintiffs are inactionable puffery. At the motion to dismiss stage, the Court analyzed this argument in some depth (MTD Order at 37-40) and held that while some of the statements may be sufficiently non-specific to constitute puffery, others “may be considered deceptive in context.” (Id. at 39.) Ultimately, the Court concluded that “these isolated representations are part of a cohesive whole which, as alleged, form a deceptive scheme by Meta to obfuscate the risks of serious harm stemming from platform use.” (Id. at 40.) Upon summary judgment, defendant proffers nothing new to warrant a change in the prior analysis. Nor, as already discussed, will the Court determine which of the sixty-two statements are puffery in deciding Meta's motion for summary judgment.

The motion on this ground is DENIED.

B. UNFAIRNESS CLAIMS

Meta seeks summary judgment on the AGs' unfair practices claims premised on the design features of Meta's platforms.

1. Section 230 & First Amendment

Meta first argues that the AGs' claims are barred by Section 230 and the First Amendment because they have failed to adduce evidence disentangling harm from social media generally from those of the actionable features.

For the reasons already discussed in the Court's prior orders in this MDL, the motion on this ground is DENIED. (See, e.g., Dkt. No. 2728, [“Breathitt MSJ Order”]; GC Experts Order.)

2. Harm

Meta next contends that the AGs adduce insufficient evidence of harm. Meta makes these arguments as to (i) the “Daily Limit” and “Take a Break” time restriction features; (ii) the multiple accounts function on Instagram; and (iii) appearance-altering filters.

As already discussed supra in section III(A)(1), and as noted in the Court's prior orders in this MDL, the AGs adduce sufficient document and expert opinion evidence ¹⁴ to create a triable issue that the design of Meta's platforms harm teens by causing compulsive use. (See, e.g., GC Experts Order.)

The motion on this ground is DENIED.

3. Unique Evidentiary Burdens under State Law

Meta argues that certain states have not met the evidentiary burdens required by each state's law for their unfairness claim. However, the only reference to specific deficiencies were made in footnotes. (See, e.g. Mtn. at 38 n. 90; id. at 39 n. 93.) The Court need not consider substantive arguments merely referenced in footnotes. United States v. Strong, 489 F.3d 1055, 1060 n.4 (9th Cir. 2007).

4. Evidence of Intent

State law varies on the standard for “intent” when required for civil penalties. The parties agree that plaintiffs must show that Meta acted “knowingly or recklessly” under Colorado law but differ on whether the same standard applies in Kentucky.

Regardless, as discussed supra Section III(A)(2), the AGs have presented evidence of intent. For example, the AGs cite to an internal Meta document discussing how to “look inward to biological factors that are pretty consistent across adolescent development and gain valuable relatively stable insights to inform product strategy” and includes that “Teens are sensitive to the feel good effects of dopamine and that they always want to feel good ․ This presents opportunities for us.” (Dkt. No. 3012-3.) Upon review of the record, the Court finds plaintiffs have adduced sufficient evidence of scienter by Meta to meet the highest standard set by state law.

The motion on this ground is DENIED.

C. DISGORGEMENT

Meta next asserts that certain states are barred from recovering disgorgement for five reasons, namely that: (1) certain states' laws do not explicitly authorize disgorgement, so they are barred from pursuing it; (2) the AGs have an adequate remedy at law; (3) the AGs cannot show that the challenged features were the source of profit nor put forth, via the expert report of Dr. Carl Saba, a reasonable approximation of the profits flowing from the challenged conduct; and (4) disgorgement must be related to harms within state boundaries and states cannot assert parens patriae standing over harms to out of state individuals.

1. State Law Analysis

Meta argues that the law of five states (CA, IN, KY, NC, and NY) do not authorize disgorgement or catch-all equitable relief. The AGs counter that in all five states, the relevant statutes do not expressly prohibit disgorgement. Thus, they argue the Court should recognize equitable jurisdiction without such a restriction. Meta has not identified caselaw in any of the five states that supports its proposition that disgorgement is not an available remedy. As the Court explained more fully on the record at the June 26, 2026 hearing, the Court is not required to entertain arguments not briefed in full. The Court denies the argument on this record.¹⁵

2. Equitable Jurisdiction

Next, defendant argues that even for states that allow disgorgement, this Court does not have equitable jurisdiction because plaintiffs (1) did not plead that they lack an adequate remedy at law and (2) in any event, plaintiffs do have an adequate remedy at law.¹⁶

Under federal common law, a plaintiff may not bring an equitable action for a claim that has an adequate remedy at law because, “a suit in equity [that would] accomplish a result which could be attained by an action at law ․ would deprive the litigant of his constitutional right to a trial by jury in the law action.” 14 WRIGHT AND MILLER'S FEDERAL PRACTICE AND PROCEDURE § 4513 (3d ed. 2016) (quoting The Effect of State Statutes on Equity Jurisdiction in the Federal Courts, 33 Yale L.J. 193, 195 (1923)). In Sonner v. Premier Nutrition Corp., the Ninth Circuit found that federal common law applies to state law claims for equitable restitution, and therefore plaintiffs must establish that they lack “an adequate remedy at law before securing equitable restitution for past harm under the UCL.” 971 F.3d 834, 844 (9th Cir. 2020).

The procedural implications of the Ninth Circuit's ruling in Sonner are subject to some debate. For example, district courts are split on how to apply Sonner at the pleading stage.¹⁷ However, there is consensus that plaintiffs must plead they have no adequate legal remedy available.

In March 2025, the Ninth Circuit explained “equitable jurisdiction is waivable.” Ruiz v. Bradford Exch., Ltd., 153 F.4th 907, 915 (9th Cir. 2025). Citing to Supreme Court precedent, the court added that such an objection is waived “where the defendant has expressly consented to action by the court, or has failed to object seasonably.” Id. (citing Am. Mills Co. v. Am. Sur. Co. of N.Y., 260 U.S. 360, 363, 43 S.Ct. 149, 67 L.Ed. 306 (1922)) (emphasis supplied). Here, defendant did not object until its summary judgment motion, despite plaintiffs' multiple references to disgorgement in both the original and amended complaints. (See Dkt. Nos. 1 & 207.) Therefore, the Court finds that this issue was likely waived. Further, and out of an abundance of caution, the Court briefly addresses the parties' arguments on this issue.¹⁸

It is not clear to the Court that any legal remedy is available in all eighteen states.¹⁹ In any event, there is certainly no adequate legal remedy.

Disgorgement allows a plaintiff to recover ill-gotten gains from defendant. Restatement (Third) of Restitution and Unjust Enrichment § 51. Unlike in Sonner, where the plaintiff sought the same relief from equitable restitution and from available legal remedies, here the AGs seek the disgorgement of any proceeds or profit that defendant received because of its allegedly wrongful conduct. Sonner v. Premier Nutrition Corp., 971 F.3d 834, 844 (9th Cir. 2020). Should the AGs prove liability, relief through compensatory damages alone would not be adequate or complete, as Sonner requires, because defendant would be permitted to retain the proceeds of their alleged misconduct.²⁰ Id.

Therefore, the Court finds that the AGs lack an adequate remedy at law and thus may pursue equitable relief under Sonner. As a result, the Court also finds that it has equitable jurisdiction. See also Guzman v. Polaris Indus. Inc., 49 F.4th 1308, 1313 (9th Cir. 2022) (“[T]o entertain a request for equitable relief, a district court must have equitable jurisdiction, which can only exist under federal common law if the plaintiff has no adequate legal remedy.”).

The motion is DENIED.

3. Evidentiary Basis

Meta argues the AGs have failed to proffer any evidence that the challenged conduct caused Meta's alleged unjust enrichment and that the AGs cannot meet their burden to identify a reasonable approximation of profits causally connected ²¹ to Meta's alleged conduct, citing SEC v. First Jersey, 101 F.3d 1450, 1475 (2d Cir. 1996) and SEC v. Liu, 2022 WL 3645063, at *2 (9th Cir. Aug. 24, 2022). Specifically, Meta critiques the expert report of Carl Saba for failing to (1) show that the youth users would not have used the platform even if the alleged conduct had not occurred or (2) show that the time spent on the platform equates to increased profits.

The AGs contend that Saba's report suffices to show that Meta acted to maximize young users' time spent on its platforms, and that additional evidence demonstrates that time spent on the platform equates to profits. Saba purports to have calculated the advertising profits “associated with users who, while teenagers, averaged more than .5, 1, 2, 3, 4, or 5 hours per day in a given month on Meta's platforms.” (Dkt. No. 2787-20 [“Saba Rep.”] at ¶¶ 175-213.)

Additionally, plaintiffs point to evidence of Meta's interest in increasing teen usage for the purpose of maximizing long-term revenue growth by retaining these users into adulthood, when ad revenue typically increases. (See Dkt. No. 2786-4 (company goals included increasing time spent by 12% and described U.S teens as “key”); Dkt. No. 2787-21, at 641-42 (noting that younger joiners to Facebook, particularly tweens, exhibited higher retention rates); Dkt. No. 2787-22 at 605 (discussing strategy of reducing ad load for teens on Instagram Reels to increase their engagement and “maximize long-term revenue growth”); Dkt. No. 2787-23 at 425 (same for Facebook); Li Dep. at 283:8–288:3.)

Therefore, the Court finds there is a dispute of material fact as to the evidentiary basis for disgorgement. Meta's arguments are factual and go to weight; they are not dispositive.

4. Extraterritoriality and Statutes of Limitations

Next, Meta argues the claims for disgorgement and civil penalties are limited by extraterritoriality and statues of limitations. The AGs respond that their civil penalties and disgorgement calculations will be presented on a state-by-state basis. The AGs report, and Meta does not dispute, that Saba's report splits out the calculations by state, albeit the lack of clarity is noticeable. Therefore, the Court does not find any extraterritoriality issue. The motion on this ground is DENIED.

Meta also notes that disgorgement must be limited by any applicable statutes of limitations. The limitations arguments are addressed supra at n. 4 and infra in Section III(D)(1).

D. CHILDREN'S ONLINE PRIVACY PROTECTION ACT (COPPA)

Meta moves for summary judgment based on (1) overarching defenses and (2) on the grounds that COPPA does not apply to Meta's platforms. For COPPA to apply, a website must either be “directed to children” or the operator must have actual knowledge of under-13 users. Meta contends that its platforms are not directed to children and that it does not have actual knowledge of child users.

The AGs cross-move for partial summary judgment on several elements of its COPPA claims, including that (i) Meta Platforms, Inc. and Instagram, LLC are “operators” of Instagram; (ii) Meta has “actual knowledge” of U13 users on its platforms; (iii) Meta has not complied with COPPA's notice and parental consent requirements; and (iv) Meta violates COPPA by using personal information of U13 users to train its machine learning and generative AI models.

1. Overarching Defenses

a) Standing

As a threshold issue, Meta asserts that the AGs lack standing to bring the COPPA action. COPPA authorizes an attorney general to bring a civil action to enforce the statute “[i]n any case in which the attorney general of a State has reason to believe that an interest of the residents of that State has been or is threatened or adversely affected.” 15 U.S.C. § 6504(1)(a). In response, the AGs point to the expert report of Ryan Sheatsley, who analyzes data produced by Meta to conclude that there were accounts disabled for being underage in each of the AGs' states. (Dkt. No. 2899-2.)

The Court concludes that the AGs have sufficient reason to believe that the interest of their residents “is threatened or adversely affected” by Meta's alleged violations of COPPA.

The motion on this ground is DENIED.

b) Laches

Meta additionally contends that the doctrine of laches bars the AGs' COPPA claims. A defendant asserting laches as a defense must show both “an unreasonable delay by plaintiff in bringing suit” and prejudice thereby to the defendant, with conduct outside the limitations period creating a presumption that laches applies. Miller v. Glenn Miller Prods. Inc., 454 F.3d 975, 997 (9th Cir. 2006).

Here, the AGs explain that they were not aware of Meta's internal practices until after the company's alleged COPPA violations began, and certainly not as early as 2011. Additionally, the AGs claim that once they were reasonably aware of the violations, they began to investigate and coordinate multistate efforts that resulted in this lawsuit. Therefore, the Court finds no unreasonable delay to support a laches defense.

The motion on this ground is DENIED.

c) Limitations Period

The parties disagree on whether there is a statute of limitations applicable to the AGs' COPPA claims. Meta argues that 28 U.S.C. Section 1658, which provides for a catchall four-year limitations period for any civil action arising under an act of Congress, applies to COPPA.²² Meta thus concludes that the statute of limitations bars the AGs from relying on facts that pre-date the limitations period, including conduct alleged that dates back to 2012.

The AGs seek to apply the nullum tempus occurrit regi 23 doctrine, which provides that “actions by government officials are not subject to state, or indeed any other, limitations periods” when suit is brought to vindicate public rights or interests, unless there is clear congressional intent for those periods to apply. S.E.C. v. Rind, 991 F.2d 1486, 1491 (9th Cir. 1993). Meta cites no authority or legislative history to suggest that Congress intended for Section 1658 to apply to government actions enforcing public rights or protecting public interests. Therefore, the Court declines to so find.²⁴

The motion on this ground is DENIED.

d) Disgorgement

Finally, Meta argues that plaintiffs are not entitled to disgorgement because (1) as a matter of law, the statute does not permit the “punitive relief” of disgorgement, (2) even if disgorgement were available under COPPA, a federal court would lack equitable jurisdiction to award such relief; and (3) as a matter of fact, they failed to meet their burden to provide a “reasonable approximation of profits causally connected to the [alleged] violation.” S.E.C. v. Platforms Wireless Intern. Corp., 617 F.3d 1072, 1096 (9th Cir. 2010).

i. COPPA Remedies

“Unless a statute in so many words, or by a necessary and inescapable inference, restricts the court's jurisdiction in equity, the full scope of that jurisdiction is to be recognized and applied.” Porter v. Warner Holding Co., 328 U.S. 395, 398, 66 S.Ct. 1086, 90 L.Ed. 1332 (1946).

Meta does not cite any authority to support its proposition that COPPA forecloses disgorgement as a remedy. Instead, it argues that disgorgement is punitive, and therefore not permitted, if the funds will not be returned to the alleged victims.

There is no basis to support defendant's assertion that COPPA does not permit disgorgement, whether it is punitive or not. The statute permits plaintiffs to seek “damage, restitution, or other compensation on behalf of residents of the State; or ․ other relief as the court may consider to be appropriate.” 15 U.S.C. § 6504(a)(1) (emphasis supplied).

The motion is DENIED on this ground.

ii. Equitable Jurisdiction

Meta cites Sonner to suggest that even if disgorgement were available under COPPA, a federal court lacks equitable jurisdiction to permit disgorgement where an adequate legal remedy is available. See supra Section III(C)(ii). However, the Ninth Circuit in Sonner was focused on whether a plaintiff could pursue equitable remedies under state law where legal remedies were available. 971 F.3d at 841 (“state law cannot expand or limit a federal court's equitable authority”) (emphasis supplied). Here, plaintiff's COPPA claim is under federal law. Federal statutes, including COPPA, may allow for the recovery of both legal and equitable remedies because Congress can displace federal common law as it sees fit. Am. Elec. Power Co. v. Connecticut, 564 U.S. 410, 423, 131 S.Ct. 2527, 180 L.Ed.2d 435 (2011). “The test for whether congressional legislation excludes the declaration of federal common law is simply whether the statute speaks directly to the question at issue.” Id. at 424, 131 S.Ct. 2527.

COPPA expressly provides for a range of remedies in actions brought by states. 15 U.S.C. § 6504. This provision gives district courts authority to order legal (“damage, restitution, or other compensation on behalf of residents of the State”) and equitable (“enjoin that practice ․ enforce compliance with the regulation ․ obtain such other relief as the court may consider to be appropriate”) remedies. Id. Therefore, the Court finds that COPPA speaks directly to the issue of whether equitable relief is available to the AGs.²⁵

iii. Reasonable Approximation

Meta also argues that plaintiffs have failed to provide a reasonable approximation of profits causally connected to the alleged violation. The AGs counter that there are disputed facts regarding the value Meta derived from the U13 data allegedly collected. Despite Meta's denial that such U13 users exist, plaintiffs cite internal documents that suggest Meta may have used U13 data to train machine learning and generative AI models. (See Dkt. No. 2789-20 at 119 (“increasing model training data drives value”); Dkt. No. 2789-21 at 738-740 (algorithms and machine learning models “help [Meta] personalize what people see,” including “the ads they are shown,” and personalization is “a central component of the value [Meta's] services provide”); Li Dep. at 88:18-21.)

The Court agrees there are material disputes of fact as to the reasonable approximation of profits. Therefore, the motion on this ground is DENIED.

2. Elements of COPPA

Title 15 Section 6502, otherwise known as COPPA, provides: “It is unlawful for [1a] an operator of a website or online service [1b] directed to children, or [2] any operator that has actual knowledge that it is collecting personal information from a child, [3] to collect personal information from a child in a manner that violates the regulations prescribed under subsection (b).” (annotations provided for clarity).

a) Operator

Meta acknowledges that Meta Platforms, Inc. currently “operates” Facebook and Instagram, and operated Facebook at all relevant times. This issue is undisputed. The remaining dispute briefed was whether (i) Instagram, LLC is an operator of Instagram, and (ii) Meta Platforms, Inc. operated Instagram in the past, as opposed to the present.

After briefing concluded, the parties filed a stipulation dismissing Instagram. (Dkt. No. 3108.) Pursuant to the stipulation, Meta agrees not to assert as a defense to liability in this matter that it is not responsible for any actions or inactions alleged by the State AGs as wrongful on the ground that such actions or inactions were taken in whole or in part by Instagram, LLC. The Court granted this joint stipulation in Pretrial Order No. 3. Therefore, the Court deems this dispute resolved.

b) Directed to Children

COPPA applies to operators of websites or portions thereof that are “directed to” children. 15 U.S.C. § 6502(a)(1).

Meta asserts that its platforms are directed at a general audience, not to U13s, including via explicit age restrictions and statistics showing that most users are adults. The AGs contend that at least portions of the platforms are directed to children, including certain child-friendly design choices and arrangements with influencers popular with children.

Whether a website is entirely, or in part, targeted to children depends upon the totality of the circumstances. 16 C.F.R. § 312.2; 90 Fed. Reg. 16918, 16937 (Apr. 22, 2025). This kind of inquiry is a classic question of fact. Here, plaintiffs have identified sufficient evidence to support their theory the platform was at least in part directed at children. For example, plaintiffs point to Meta's practice of tracking 12 to 15-year olds in a group (Dkt. No. 2788-25 ¶¶ 424-425); an internal document noting that “Facebook already has ~5M[illion] monthly web visitors under 13 in the US” and that “Google+ is way behind in this segment” (Dkt. No. 2789-1 at 389); and an internal document noting that “11 and 12 year olds aren't so different from 13 or 14 year olds.” (Dkt. No. 2789-2).

The Court finds the evidence plaintiffs cite is sufficient to show there are material disputes of fact as to whether the platforms are directed, albeit partially, at children. Therefore, the motion on this ground is DENIED.

c) Collection of Personal Information

The AGs ask the Court to enter partial summary judgment finding that Meta collects personal information of U13 users. Meta argues (1) that this is an improper use of Rule 56 because whether they collect the personal information is not an element of COPPA; and (2) that it does not collect “personal information” as defined by the statute.

If a court finds that there is no genuine dispute of material fact as to only a single claim or defense or as to part of a claim or defense, it may enter partial summary judgment. Fed. R. Civ. Pro. 56(a). The Court disagrees with Meta that this finding would not streamline the trial and thus rejects the Rule 56 argument.

COPPA defines “personal information” as “individually identifiable information about an individual collected online, including--(A) a first and last name; (B) a home or other physical address including street name and name of a city or town; (C) an e-mail address; (D) a telephone number; (E) a Social Security number; (F) any other identifier that the Commission determines permits the physical or online contacting of a specific individual; or (G) information concerning the child or the parents of that child that the website collects online from the child and combines with an identifier described in this paragraph.” 15 U.S.C. § 6501(8). The AGs have proffered evidence that Meta collects many types of data from its users, including: email addresses; phone numbers; images of a person's face; the sound of a person's voice; cookie IDs; IP addresses; information that the user provides on their profile, such as content created by the user; and information that reflects the user's interactions on the platform, such as likes, the videos they watch, and who they interact with. (See, e.g. Hartnett Dep. at 28:17-29:19.) Some of this data is explicitly recognized by the statute as personal information, such as email addresses and telephone numbers. (Id. at 29:17-19 (30(b)(6) testimony confirming Meta collects either an email address or phone number from users with a Facebook or Instagram account)).

Therefore, the Court finds that Meta does collect personal information from all its users, including any users found to be U13. That Meta may disagree with respect to whether other kinds of information fall within the scope of COPPA, the Court need not resolve that dispute here.

d) Actual Knowledge

Meta and the AGs each move for summary judgement on the issue of whether Meta has “actual knowledge that it is collecting information” from U13s. 15 U.S.C. § 6502(a)(1).

“Actual knowledge” is not defined in the statute, but the Supreme Court has concluded it requires more than “constructive” or “hypothetical” knowledge, and that it may be proved through “inference from circumstantial evidence.” Intel Corporation Investment Policy Committee v. Sulyma, 589 U.S. 178, 185, 189, 140 S.Ct. 768, 206 L.Ed.2d 103 (2020). Additionally, “willful blindness,” which requires taking deliberate action to avoid learning a fact, is also considered tantamount to actual knowledge. Global-Tech Appliances, Inc. v. SEB S.A., 563 U.S. 754, 769, 131 S.Ct. 2060, 179 L.Ed.2d 1167 (2011).

The parties vigorously dispute the facts surrounding this element. The AGs seek summary judgment based on evidence in four situations: (i) users who, after sign-up, change their listed age on the platform to under 13; (ii) users who Meta checkpoints based on human review of profiles; (iii) users who are “hard-linked” or “soft-matched” to checkpointed accounts; and (iv) users whose accounts Meta ultimately disables after checkpointing. Meta responds that its efforts at checkpointing and deleting accounts in those circumstances are based on “suspicions” of users being under 13, not actual knowledge, and thus there is no violation. The Court takes each claim in turn.

i. Users who change their age to under-13

The parties do not dispute that when a user changes their date of birth to a date that would make their age under 13, Meta “checkpoints” the user's account. “Checkpointing” an account is when Meta disables an account based on evidence that the user may be U13. (See Hartnett Dep. at 105:11-21.) The account owner may regain access to their checkpointed account by providing identification verifying they are 13 or over within 30 days. (Id. at 105:23-106:17.) If they fail to do so within 30 days, Meta schedules the account for deletion. (Id. at 106:3-7.) Meta retains the data for an additional 45 days, which it contends is standard industry practice and is done to ensure legal compliance. (Id. at 299:14-301:13.) Meta acknowledges that this data retention period has, in the past, at times been as long as six months or one year. (Id. at 299:22-24, 300:21-25.) After the data retention period, Meta permanently deletes the data in a process which takes another 90 days. (Id. at 301:15-19.)

The AGs focus on evidence which it interprets as admissions: for instance, Meta told the Federal Trade Commission (“FTC”) that “[b]ecause these users have themselves indicated they are under 13, Meta will presumptively determine that these users are too young to use the service, absent additional proof of age.” (Id.) Also, a Meta employee described a platform functionality involving users changing their date of birth to an age under 13 as constituting “explicit knowledge of U13.” (Dkt. No. 2696-41.)

Meta provides a different interpretation. It responds that accounts of users who change their date of birth to reflect an age under 13 often are in fact not under 13. Meta proffers evidence that bad actors use date of birth changes to purposefully get their accounts checkpointed and deleted to conceal a history of illicit online activities. (Dkt. 2781-8 (Meta employee reporting that “[w]e have recently found that 70% of accounts who edit DOB [date of birth] to U13 seem to be involved in fraud․ They are purposefully reporting their accounts as U13 so that their account gets deleted”).) As a result, Meta maintains, the fact that a user's account shows a date of birth reflecting an age under 13 does not provide actual knowledge that the user is a child. Even if true, Meta does not claim that 100% of the accounts fall into this category.

Nonetheless, based on the foregoing disputes, the Court concludes there is a material dispute of fact as to the claim that a date of birth changed to reflect an age under 13 provides Meta with actual knowledge of accounts belonging to U13s. The AGs' and Meta's cross motions for summary judgment on this ground are both DENIED.

ii. Meta checkpoints after human review

Meta checkpoints the accounts of users following human review if the user has listed their age as under 13 or posted photographs suggesting the account holder is under 13. (Dkt. No. 2696-4 at 184:18–186:8.) The AGs describe that Meta has implemented this policy including by, at times, requiring human reviewers to checkpoint an account when and a quarter to one half of photos on the account feature a particular child, and fewer than three of the photos are of a given teen or adult. Id. The AGs contend that, as a matter of common sense, when human reviewers review the aforementioned information and put the account into an age checkpoint, Meta has obtained actual knowledge that the user is a child.

Meta responds that checkpointing based on human review merely shows that the human reviewer concluded the account might belong to a user under 13. Meta adduces evidence that human reviewers are instructed to “err on the side of” checkpointing a user if they could be under 13. (Hartnett Dep. at 105:16-21). Additionally, Meta proffers evidence suggesting that human review determinations are “subjective,” particularly with regard to assessing age based on photos, and so reviewers are instructed to “checkpoint if they are unsure.” (Dkt. No. 2781-14.) Meta contends because human reviewers checkpoint many accounts belonging to non-children, checkpointing based on human review does not give Meta actual knowledge that a user is a child.

Given the evidence proffered by each side, the Court concludes there is a material dispute of fact as to whether accounts that are checkpointed based on human review provide Meta with actual knowledge under COPPA. The AGs' and Meta's motions for summary judgment on this ground are both DENIED.

iii. Users with “hard-linked” or “soft-matched” accounts

The parties agree that user accounts can be “hard-linked” through Meta's Account Center, meaning that they are connected to one another. Meta also acknowledges that it has data that permits it to “soft-match” accounts, meaning that it predicts that multiple accounts likely are the same person.

In this context, the AGs contend that Meta failed to use hard-link and soft-match data to enforce underage policies, and in fact, chose not to checkpoint and disable user accounts linked to accounts which had previously been checkpointed or disabled for being under 13. The AGs argue that Meta's inaction against accounts hard-linked or soft-matched accounts constitute willful blindness, in that they reflect a deliberate choice to avoid knowledge that the owner of those matched accounts is under 13.

In support, the AGs proffer documents suggesting employees were concerned about the failure to enforce Meta's policies against underage accounts through hard-linking and soft-matching. (Dkt. No. 2696-45 (internal 2020 memo stating “We have deprioritized u13 enforcement because COPPA compliance is at risk when we sit on knowledge of u13, [it] incentivizes not knowing”; recommending “cross-enforcement of u13s on IG who have hard-linked FB accounts”; and suggesting that “basic COPPA compliance is at risk” when failing to enforce via hard-linking); id. (“It is difficult to defend the use of [soft-matching] for advertising purposes without also using these links for age management.”); id. Dkt. No. 2696-46 (describing that, in certain contexts, accounts soft-matched with “>0.99 confidence” leads to account disabling).)

Meta offers a different interpretation of the evidence. As already discussed, Meta contends that checkpointed accounts do not provide it with actual knowledge. Meta next argues that hard-linked accounts only confirm that the users “know each other or have the logins,” and that it is “fairly common for the accounts to be linked together” but operated by different people. (Dkt. No. 2781-6 at 212:17-213:10.) For example, Meta suggests that parents may hard-link accounts with their teenager to monitor the latter's activity on the platforms, meaning that a hard-linked user is not a child. Third, Meta proffers expert opinion that soft-matching does not identify with certainty if an account belongs to a common user, but merely produces a probabilistic linkage. (See Dkt. No. 2781-17.) Finally, Meta disputes the AGs' suggestion that soft-matching reflects a 99% probability that the accounts belong to the same users. Meta asserts that that level of accuracy is only achieved in limited situations and, more generally, soft-matching is not accurate at the level of a single user. (Dkt. No. 2781-8 at 11, 13.)

Based on the record before it, the Court concludes there is a material dispute of fact as to the claim that hard-linking and soft-matching provides Meta with actual knowledge, including willful blindness, of accounts belonging to U13s. This inquiry will be better informed with a complete record. The dueling motions for summary judgment on this ground are both DENIED.

iv. Users whose accounts are deleted after checkpointing

Finally, the AGs argue the Court should find that Meta violated this element with respect to the users whose accounts it deleted after checkpointing. Meta disagrees, asserting that it only disables an account after the time period expires for a user to appeal Meta's decision. Thus, Meta reasons that it does not have actual knowledge that an account is under 13—rather, plaintiffs can only prove knowledge that a user failed to appeal successfully. Meta adduces evidence that large portions of checkpointed accounts demonstrably do not belong to children, including in connection with cyber attacks against users designed to trigger Meta's account disabling.

This category, too, raises numerous factual issues.

In summary, Meta's operational processes regarding how accounts are reviewed, checkpointed, disabled, and whether data is retained are fact intensive and technical. Ultimately, each side adduces evidence regarding those processes that must be weighed by the finder of fact at trial. Because a material dispute of fact exists as to whether Meta has actual knowledge, including willful blindness, of u-13 users on its platform, the motions on this ground are DENIED.

e) Retention of Under-13 Data for AI Model

The AGs argue that the Court should grant summary judgment on its claim that Meta retains U13 data for its large language models (“LLMs,” or “AI models”), despite checkpointing those accounts as belonging to U13s. Meta counters that this claim is untimely; that COPPA does not prohibit maintenance of personal data belonging to individuals that were identified as children after that data was collected; and that Meta does not actually use or maintain that data for its AI models. The Court addresses each.

Timeliness: Meta first asserts that the AGs are barred from raising this issue as it was not specifically alleged in their complaint as a basis for COPPA liability.

“A plaintiff must provide the defendant with adequate notice of his or her claims.” In re TFT-LCD (Flat Panel) Antitrust Litigation, 2012 WL 6708866, at *3 (N.D.Cal. 2012), aff'd, 637 F. App'x 981 (9th Cir. 2016). The “federal rules contemplate that the process of defining and narrowing the issues raised in the pleadings will be accomplished through discovery and other pretrial procedures” and that “a party may even be permitted to alter the legal theory of its case” during discovery. William Inglis & Sons Baking Co. v. ITT Continental Baking Co., Inc., 668 F.2d 1014, 1053 (9th Cir. 1981) (answers to interrogatories sufficient to put party on notice of claims).

The AGs disagree with Meta and reference a portion of the complaint that alleges (i) Meta collects “personal data to train its Recommendation Algorithms” (Compl. ¶ 175) and (ii) separately alleges that Meta “has repeatedly collected, used, or shared personal information about children under the age of 13 and continues to systematically do so” without complying with COPPA. (Id. ¶ 852). Complaints do not need to assert theories.

While the foregoing might have been enough to put Meta on notice, the AGs also point to evidence that Meta was put on notice through the discovery process. In January 2025, the AGs sought deposition testimony from Meta on “data collected from Child or Teen Users in training or developing algorithms or models,” and submitted a request for admission (“RFA”) in February 2025 regarding whether Meta, once it identifies a user as potentially being under 13, “determine[s] whether the user's data was used to develop, train, or validate any of your algorithms or models.” (Dkt. No 2696-5; Dkt. No. 2696-6.) After obtaining additional data during discovery that the AGs contend validated this theory, the AGs provided interrogatory responses in October 2025 that laid out their claim that “Meta continues to use the personal information of children under 13 after determining that they are likely under 13 to train its machine learning and generative AI models, without obtaining parental consent.” (See Dkt. No 2899-16 at 6–10.)

Thus, this argument is rejected. Meta has received adequate notice.

Maintenance of Data Subsequent to Discovery of Under-13 Users: Meta also contends that even if it does eventually obtain knowledge that users are U13, it cannot be liable for using their data for model training. COPPA's statutory text provides that parental notice and consent requirements attach when an operator “has actual knowledge that it is collecting personal information from a child.” 15 U.S.C. § 6502(a)(1) (emphasis supplied). Meta contends the statute is silent on data previously collected without such knowledge. Thus, Meta argues that if it collects data from individuals whom it does not know to be under 13 but later discovers that they were, COPPA cannot give rise to liability for the subsequent use of their data.

Meta's interpretation is contrary to federal guidance. The FTC has explained that “if such companies collect personal information from visitors who click on their ads at general audience sites, and that information reveals that the visitor is a child, then they will be subject to the Act.” Children's Online Privacy Protection Rule, 64 Fed. Reg. 59888, 59892 (Nov. 3, 1999) (emphasis added). Thus, it indicates that knowledge obtained after the data is collected is sufficient to form a basis for liability.

The FTC further explained that “operators may acquire actual knowledge that they are dealing with children under 13” where they ask “age identifying” questions that “do not directly ask age or grade.” Id. Meta repeatedly ignores that “actual knowledge” includes willful blindness. The FTC promulgated a rule that attaches COPPA liability when a website operator has “actual knowledge that it is collecting or maintaining personal information from a child.” 16 C.F.R. § 312.3 (emphasis supplied). Under the FTC rule, once Meta discovers a user is a child, it cannot keep or use their personal information. The FTC has further clarified that if, after collecting a user's personal information, “the operator later determines that a particular user is a child under age 13, COPPA's notice and parental consent requirements will be triggered.” (Dkt. No. 2899-20, Gooler Decl., Ex. J.) This reading is entirely consistent with the statute.

Meta responds that the FTC's inclusion of the phrase “or maintaining” impermissibly expands COPPA's reach without basis in the statutory text. They thus argue the FTC violated the APA (1) procedurally because the FTC failed to provide notice or a reasoned explanation; and (2) by exceeding the rulemaking authority delegated by Congress.

The Court not only declines to engage with Meta's procedural argument ²⁶ , but finds the argument ironic given its own expert's opinions regarding the primacy of the FTC's guidance where it suits Meta's purposes. The Court also finds it unnecessary to address Meta's substantive challenge that the FTC exceeded its authority. Based on the language of the statute alone, the Court finds Meta's position untenable. The statute requires an operator “to obtain verifiable parental consent for the collection, use, or disclosure of personal information from children.” 15 U.S.C. § 6502(b)(1)(A)(ii). Nowhere does the statute suggest that collection of U13 data can only serve as a basis for liability if the operator verifies that the user is U13 before the data is collected. Common sense instructs that user data must necessarily be collected for an operator to determine whether that user is U13 or not. Therefore, Meta's interpretation would render the provision meaningless. “It is our duty to give effect, if possible, to every clause and word of a statute.” City of Los Angeles v. Barr, 941 F.3d 931, 939 (9th Cir. 2019) (citing United States v. Menasche, 348 U.S. 528, 538–39, 75 S.Ct. 513, 99 L.Ed. 615 (1955)).

Next, 15 U.S.C. § 6501 defines “disclosure” in part as “making personal information collected from a child by a website or online service directed to children or with actual knowledge that such information was collected from a child, publicly available in identifiable form, by any means.” (emphasis supplied). As noted above, parental consent is required for the disclosure of U13 personal information. Thus, the statute expressly contemplates that disclosure before actual knowledge can form the basis of liability if the service is directed to children. The Court sees no reason to read into the statute a requirement that is not there.

Whether Data from Checkpointed Accounts is used in LLMs: The AGs argue it is undisputed that “Meta continues to use [U13] personal information for machine learning and generative AI model training after it develops actual knowledge that they are [U13].” They point to evidence that Meta generally uses all user data to train models (see, e.g. Dkt. No. 2696-19 at 25:4-21) (30(b)(6) deponent confirming Meta uses all user data to train machine learning models, and in fact specifically uses teen data to do so)); that Meta has increased its use of teen data to improve its models (Dkt. No. 2696-24 (an internal document discussing using teen data to enable better model training to “deliver model improvements for Teens and Gen pop.”)); and that Meta continues training its machine learning and generative AI models on user data from accounts that are checkpointed, disabled, and/or scheduled to be deleted (see, e.g. Dkt. No. 2696-22 (30(b)(6) testimony suggesting that where data is retained prior to deletion, it is likely used during that period)). While Meta deletes data from certain accounts flagged as potentially U13, Meta acknowledges that the data retention period has, in the past, at times been as long as six months or one year. (Hartnett Dep. at 299:22-24, 300:21-25.) However, Meta also claims that any such data is marked “not intended for business purposes” and thus not used for model training.

The Court finds that the AGs have proffered sufficient evidence to establish a material dispute of fact, but not enough to warrant partial summary judgment on this point. Both motions are DENIED on this ground.

f) Compliance with COPPA Notice Requirements

The AGs seek a legal ruling that Meta has met none of COPPAs requirements: (i) providing notice to parents; (ii) seeking and obtaining parental consent; and (iii) providing parents with a means of reviewing personal data collected, requesting deletion, and refusing further data collection. The AGs base their motion on Meta's own responses to RFAs.

Meta's denial strains credulity. Meta has repeatedly urged that it need not comply. Based on this position, it has not. The issue is whether liability attaches in the first instance. The lack of compliance is obvious.²⁷

While Meta's RFA responses are each styled as denials, each response makes clear that Meta has not provided parental notice, obtained parental consent, or provided a way for parents to review data. The denials are based solely on its contention that Meta lacks “actual knowledge,” discussed above. (See Dkt. No. 2696-26 at 3) (“Because during the Relevant Time Period, Meta has not knowingly collected personal information from individuals under 13 on Facebook or Instagram ․ Meta has not provided notice as described in 16 CFR § 312.4”); id. at 9 (“Meta has not provided parents means to review personal information of their children under 13 whom Meta does not know exist”); id. at 10 (“Meta has not provided parents a means to refuse to permit the further use of personal information from their children whom Meta does not know exist”); id. at 5 (“Meta has not sought verifiable consent from parents of individuals under 13 whom Meta does not know exist”) (emphasis supplied); Dkt. No. 2696-27 (similar for Instagram).)

Meta's responses “denying” admission speak only to why it has not provided the relevant functions, not whether it has done so. As such, no genuine dispute of fact exists as to this element. Therefore, plaintiff's motion as to this element is GRANTED.

IV. CONCLUSION

Defendant's motion for summary judgment is DENIED. Plaintiffs' motion for summary judgment is DENIED IN PART and GRANTED IN PART.

IT IS SO ORDERED.

This Order terminates Dkt. Nos. 2695, 2704, 2739 in Case No. 22-md-03047.

FOOTNOTES

1.   The Court notes that, as is relevant to much of the analysis in this Order, and as defendants conceded on the record at the April 15, 2026 hearing, neither party identified for the Court the actual statements on which plaintiffs proceed for their deception claims. Therefore, the Court declines to make any legal or factual determinations as to those specific statements. The Court ordered plaintiffs to identify these statements in advance of trial. (Dkt. No. 3139, Pretrial Order No. 2.)

2.   Where the parties have agreed to seal employee names, emails, and other personal information, including as to exhibits cited in this Order, the motions are GRANTED.

3.   If the goal of this motion had been to narrow the scope of the trial, this was not the appropriate motion to achieve that goal. This Court has had repeated conferences with counsel. The topic could have been easily raised so as to avoid wasting limited judicial resources.

4.   Meta also asserts that some of the deceptive statements, including statements made in 2018, 2020, and 2021, are barred by the statutes of limitations of certain states.Meta identifies the following states' relevant limitations periods: California (FAL), Colorado, New York, South Carolina, and Wisconsin (three years); California (UCL) and Nebraska (four years); and Delaware and Indiana (five years). The parties also entered into two agreements, one in 2021 and one in 2023, tolling “the running of any applicable statute of limitations or by application of any defense of laches.” (Dkt. No. 2705-89 at 2.) As discussed supra, defendants have not identified the full list of statements they are moving on. Nor do they explain how the statements fall outside of the limitations periods. Indeed, they concede that statutes of limitations bar only a few of the purportedly deceptive statements for only some states. Given the lack of briefing of this issue, the Court declines to grant summary judgment on this basis.

5.   The Court notes the following statements regarding addiction that plaintiffs contend are deceptive. Meta's statement to the BBC in 2018 that “at no stage does wanting something to be addictive factor into” Meta's platform design process. Dkt. No. 2705-73 at 4. In 2021, Adam Mosseri, Head of Instagram, stated to Congress: “I don't believe the research suggests that our products are addictive.” Dkt. No. 2705-6 at 13. Also in 2021, Facebook Global Head of Safety Antigone Davis told Congress: “I disagree with calling our products addictive.” Dkt. No. 2705-4 at 38. In the same committee hearing, a senator subsequently asked Davis “to drill down on this addictive element ․ isn't part of your business model to have more eyeballs for a longer amount of time engaged using your services?” to which Davis responded: “Respectfully, Senator that is not actually how we build our products.” Id. at 39.

6.   Because neither party identified the full list of statements, the Court cannot make wholesale conclusions about the statements plaintiff is proceeding on.

7.   As the Court has already noted, plaintiffs' experts, the scientific literature, and defendants' own documents use a variety of terms interchangeably, referring to “addictive,” “excessive,” “problematic,” or “compulsive” use. See GC Experts Order at 11-12. The Court declines to draw lines between these terms. For example, an internal Meta document describes that “[t]he concept of problematic use of social media has been defined by researchers as a reported lack of control over social media use that is perceived to lead to negative life impacts.” Dkt. No. 2783-2, O'Neill Decl., Ex. 22.

8.   The Court previously declined to exclude Prinstein's and Zicherman's opinions. See GC Experts Order.

9.   Meta contends that the laws of thirteen states (CA, CO, DE, IL, IN, KS, KY, NJ, NC, LA, PA, SC, and VA) contain some variation of a knowledge requirement.

10.   As discussed above, the Court cannot make wholesale conclusions about the statements plaintiffs are proceeding on for these claims, because neither party identified the full list of statements in briefing these motions.

11.   Defendant was unable to identify these statements for the Court when asked at the April 15, 2026 hearing. Therefore, the Court is unable to make any conclusions regarding statements that were not specifically identified in the motion briefing.

12.   The Court denied defendant's 702 objection to this opinion in its May 20, 2026 Order. (Dkt. No. 3068 at 5.)

13.   To reiterate the point made throughout this Order, defendants were unable to identify which statements plaintiffs are proceeding on when asked at the April 15, 2026 hearing. The parties confirmed that these were never identified for the Court.

14.   For example, Meta's own documents support plaintiff's theory that time restriction tools were merely a “public relations stunt” employed because defendant knew that more time spent on social media was linked with negative outcomes for teens. See Dkt. No. 3029-7 at 9–10, 36–41 (“Internal document reporting that millions of teens use Meta's platforms between midnight and four a.m. on a weekly basis.”); Dkt. No. 3029-8 at 700 (“suggesting that [Meta's] products [like Quiet Mode or Take a Break] served as proof points but weren't effective at reducing behaviors.”); Dkt. No. 3029-9 at 940 (internal document noting that time restriction tools like Teen Accounts were part of a “public affairs strategy to improve parental alignment when it comes to the safety of teens,” and “contributes to growing perception among parents that Instagram is on the same side as parents.”)

15.   In the AGs' briefing, they assert that California seeks disgorgement only for the period beginning January 1, 2024, which is the effective date of California Government Code § 12527.6. This law provides a disgorgement remedy under the UCL and FAL. Meta argues such relief is not available for conduct that occurred after the complaint was filed. The Court has ordered the AGs to submit additional disclosures regarding their request for disgorgement and will address any related issues in the context of further trial proceedings.

16.   The parties agree that the disgorgement requested is an equitable remedy.

17.   Some courts allow equitable claims to survive dismissal when pled in the alternative. See, e.g., Carroll v. Myriad Genetics, Inc., 2022 WL 16860013, at *6 (N.D. Cal. 2022); Yeomans v. World Fin. Grp. Ins. Agency, Inc., 2022 WL 844152, at *7-8 (N.D. Cal. 2022). Other courts do not permit plaintiffs to “choose between two available inconsistent remedies” because, in their view, the question is “whether equitable remedies are available to [the plaintiff] at all.” Nacarino v. KSF Acquisition Corp., 642 F.Supp.3d 1074, 1083 (N.D. Cal. 2022).

18.   Regarding the alleged pleading deficiency, pursuant to the Court's authority to grant leave for amendments pursuant to Federal Rule of Civil Procedure 15(a)(2), the Court ORDERS the States to file a one-page supplement to the complaint explicitly identifying that there is no adequate legal remedy no later than July 10, 2026. Meta is reminded that Rule 15 permits the Court, where appropriate, to give leave to amend the complaint, even during trial.

19.   In fact, at the May 27, 2026 conference, Meta's counsel confirmed on the record that they agree everything is equitable.

20.   The same is true for the States' request for injunctive relief. Id.; see also Linton v. Axcess Fin. Servs., Inc, 2023 WL 4297568, at *3 (N.D. Cal. June 30, 2023) (holding that Sonner does not apply to claims for injunctive relief where defendants cannot avoid future harm, like continued elevated interest rates).

21.   Plaintiffs point out that Meta incorrectly asserts State v. Minn. Sch. of Bus., 935 N.W.2d 124 (Minn. 2019) requires the Minnesota AG to prove a “causal nexus” between harm suffered by consumers and Meta's unlawful conduct. (MSJ at n. 98.) That is a requirement for restitution, not disgorgement.

22.   The parties disagree on whether the parties' tolling agreements—which date back to 2021 for Instagram and 2022 for Facebook—should apply to the COPPA claims. Meta argues that they do not, and thus any claims relating to conduct before October 2019 fall outside of the four-year limitations period. Because the Court determines, infra, that the statute of limitations argument fails on other grounds, the Court does not reach this issue.

23.   Latin for “no time runs against the king.”

24.   Neither party cites Ninth Circuit authority that addresses whether the federal statute of limitations under Section 1658 applies to parens patriae (Latin for “parent of the country”) actions brought by state attorneys general pursuant to federal statutes. Still, the Ninth Circuit has recognized that parens patriae suits brought by state attorneys general are fundamentally different from ordinary civil actions. See, e.g., Washington v. Chimei Innolux Corp., 659 F.3d 842 (9th Cir. 2011) (determining that such actions are not civil actions subject to the Class Action Fairness Act).

25.   See Porter v. Warner Holding Co., 328 U.S. 395, n. 5, 66 S.Ct. 1086, 90 L.Ed. 1332 (1946) (finding that an amendment adding a legal remedy to a statute did not provide any “indication that [it] was intended to whittle down the equitable jurisdiction recognized by [the statute] so as to preclude a suit for restitution”).

26.   Procedural challenges to agency rules under the Administrative Procedure Act are subject to the general six-year limitations period in the U.S. Code. Under Wind River, challenges to a “mere procedural violation in the adoption of a regulation or other agency action” must be brought within six years of the agency rulemaking, whereas challenges to “the substance of an agency's decision as exceeding constitutional or statutory authority” may be brought any time “within six years of the agency's application of the disputed decision to the challenger.” See Wind River Mining Corp. v. United States, 946 F.2d 710 (9th Cir. 1991) (citing 28 U.S.C. § 2401(a)).

27.   The Court declines Meta's invitation to not reach this question given its arguments on the issue of “actual knowledge.” Contrary to Meta's suggestion, evidence of knowledge exists. Whether liability exists remains disputed.

YVONNE GONZALEZ ROGERS, UNITED STATES DISTRICT JUDGE