Meta’s $17 Billion Child Safety Settlement Puts a Meter on the Machine — but Leaves the Machine Running

Meta has agreed to pay up to $17 billion and impose new restrictions on how teenagers use Facebook and Instagram as part of a historic settlement. California calls it transformative. One litigator calls it a toll booth. Somewhere in Menlo Park, the infinite scroll keeps scrolling.

The tech giant has promised billions of dollars and a substantial new collection of guardrails for teenagers on Facebook and Instagram.

The settlement ends the states’ landmark federal case against the social-media giant just as the trial was gathering momentum in Oakland. California Attorney General Rob Bonta declared that the agreement would “make social media less dangerous for our kids,” while posting that the changes would arrive “in the next few months.”

The numbers certainly look suitably enormous. Meta will pay the participating states up to roughly $17 billion over 10 years, while California says it stands to receive between $1.5 billion and $2.1 billion.

There is, however, an important word in that sentence: up to.

The agreement divides Meta’s financial obligations into guaranteed and contingent payments. The guaranteed money is paid in 10 installments, while additional contingent payments depend on an industry-wide trigger involving rival platforms. The agreement defines the relevant “Core Industry Members” as Snap, TikTok and YouTube, with the monetary trigger tied to industry-wide adoption and financial obligations imposed on the larger rivals.

In other words, this is a multibillion-dollar punishment with something resembling a group-discount clause.

We put the settlement at up to $16.68 billion, while California rounds it to $17 billion; other reporting has described the broader package at approximately $18 billion. Reuters also reported that Meta shares rose 2.3% following the announcement — not usually Wall Street’s traditional way of signaling that a company has just been mortally wounded.

And that gets to the argument now surrounding the deal: is this the long-awaited regulatory reckoning for social media, or a very large bill that Meta can comfortably pay to keep most of its underlying business intact?

Meta’s New Teen Rules Add Friction, Not a Full Redesign

There appears to be no question that the settlement requires meaningful changes.

Teen users will face a default limit of 60 minutes on an individual Meta platform and 120 minutes across Meta platforms per day. Once the cap is reached, much of the platform becomes inaccessible until midnight, although messaging, settings and long-form video are treated separately. A parent must approve any move to a less restrictive limit.

Teen accounts will also be blocked by default from using the platforms between midnight and 6 a.m., while push notifications will be switched off between 10 p.m. and 7 a.m. unless a supervising parent changes the setting.

During school hours, push notifications will generally be disabled, and parents will be able to shut down functionality other than messaging altogether. Meta must also introduce “productive pauses” at 60 and 90 minutes of cumulative daily use, plus notices after 15 minutes of continuous use.

There are other changes. Cosmetic-procedure filters will be disabled for teenagers, and Meta must maintain and improve safeguards intended to reduce exposure to inappropriate material. California says numerical like and reaction counts will be hidden from under-18s, teens will be offered a non-personalized feed, and Meta must respond to 90% of enhanced teen reports of potentially harmful content within six hours.

“Meta paid a toll, and what that toll buys is the right to keep running the machine it built for exploiting children.” – James Rubinowitz, Civil Litigator

Age assurance gets considerably more serious, too. Meta must develop systems for identifying users under 13, put mechanisms in place to detect attempts to circumvent age checks, and subject parts of that process to an independent auditor. Data collected for age assurance is also subject to minimization requirements, and data concerning under-13 users retained for the age model cannot be used for advertising, marketing or algorithmic optimization.

Bonta’s office calls those changes “real enforceable protections,” saying Meta must make “massive transformations” within months.

The court was similarly approving. Judge Yvonne Gonzalez Rogers concluded that the settlement represents a “fair, reasonable, comprehensive, and good faith approach” that provides monetary relief while attempting to change conduct and address the negative effects of the platforms.

The parties’ own joint motion was naturally no less enthusiastic, describing the agreement as the result of “difficult, good-faith, and arms-length negotiations” after years of litigation and the beginning of trial.

That is one interpretation.

James Rubinowitz, a civil litigator focusing on AI regulatory regimes and a regular commentator on the subject, sees something closer to Meta paying for an unusually expensive parking permit.

“Meta will pay up to $17 billion over the harm 47 states say its platforms did to children, and its shares rose in early trading,” Rubinowitz told Techopedia. “I read the reported terms, and this feels like a major win for Meta, and a major loss for everyone else. Meta paid a toll, and what that toll buys is the right to keep running the machine it built for exploiting children.”

The Meta Features at the Heart of the Case Largely Survive in Settlement

Rubinowitz’s criticism is less about what the settlement contains than what it does not remove.

“The states went to trial demanding removal of the features their own case called harmful to kids: infinite scroll, autoplay, disappearing Stories, beauty filters, and the algorithms trained on children’s data,” he said. “None of those removals made the deal.”

The agreement itself describes restrictions on autoplay as an “Optional Protective Setting.” The personalized feed survives too, although teenagers are to be offered the option of switching to a chronological, non-personalized alternative.

“If a court was ever going to order a real redesign of these products, it was this one, and the states took the check instead.” – James Rubinowitz, Civil Litigator

Former Meta engineer and whistleblower Arturo Béjar had already testified that autoplay, like counters and infinite scroll were “inherently unsafe for teenagers.” He also told the court that some of Meta’s existing safety features were “designed to fail,” criticizing protections that users had to activate themselves rather than receiving by default.

Rubinowitz therefore describes the settlement as “a band-aid on a bullet wound.”

“Autoplay stays on unless a parent finds the setting to turn it off, and infinite scroll, Stories, and the personalized feed survive untouched, running through every capped hour,” he said.

His metaphor for the settlement is even less flattering: “The states put a meter on the machine. The machine itself is exactly what it was.”

Meta, unsurprisingly, sees the agreement rather differently. C.J. Mahoney, the company’s chief legal officer, said the framework would give parents greater control over their children’s use of its platforms while establishing what Meta believes should become an industry-wide standard for teen safety.

“Our new Time Limit commitments, Night Mode features and usage limits during school hours set the right path forward for our whole industry,” Mahoney said. But he argued the framework would “only work if all our peers join us,” pointing to the ease with which teenagers move between competing apps. He specifically called on TikTok and YouTube to adopt the framework “right away,” adding that its success depends on other social-media platforms “following Meta’s lead.”

That appeal to Meta’s rivals is more than corporate throat-clearing. The settlement itself makes part of Meta’s potential multibillion-dollar payout contingent on broader industry adoption, with Snap, TikTok and YouTube explicitly named as the agreement’s “Core Industry Members.” In other words, Meta would quite like everyone else to follow its lead — and, under the terms of the deal, there are several billion reasons why.

The States Settled Just as the Trial Was Turning Up the Heat

The lawsuit was filed in 2023, accusing Meta of using Facebook and Instagram to “entice, engage, and ultimately ensnare youth and teens,” misleading the public about the platforms’ dangers and violating COPPA through its handling of children’s data. Meta denied those allegations and liability.

Trial began on August 18. Before it even started, Meta had failed in an attempt to end the case on summary judgment. Judge Gonzalez Rogers also agreed with California on an important COPPA issue, finding that Meta had not obtained parental consent in a manner sufficient to satisfy the federal children’s privacy law.

Then came Béjar.

The former Meta engineer testified that Zuckerberg’s approach made addressing safety problems extremely difficult and that safety features were “designed to fail.”

Rubinowitz argues that this was precisely the moment for the states to keep going.

“The states stopped at the worst possible moment to stop,” he told Techopedia. “They were in week two of a six week trial. The judge had already ruled against Meta on part of the case before opening statements. Arturo Bejar, a former Meta engineer, had just told the jury that one of the company’s own safety features was ‘designed to fail.’ Mark Zuckerberg came off the witness list without answering a single question.”

For Meta, ending the trial also eliminated the possibility that a court might impose more fundamental product changes. The company’s own SEC disclosures acknowledged that plaintiffs in its youth-related litigation were seeking remedies ranging in some cases to more than $1 trillion, as well as “extensive changes” to Meta’s business practices and third-party oversight.

A $17 Billion Headline Looks Smaller Against Meta’s Balance Sheet

Then there is the fine — or, more accurately, the monetary settlement.

Meta reported $200.97 billion in revenue in 2025. Against that backdrop, even a $17 billion headline figure spread over a decade looks rather different from $17 billion leaving the corporate checking account on Friday afternoon.

The settlement provides for guaranteed payments over 10 installments, plus contingent payments that are forfeited if the required industry trigger never occurs.

Rubinowitz argues the structure gives Meta another significant advantage.

“No admission of wrongdoing,” he said. “Roughly $5 billion of the money, as the AP reports the terms, pays out only if YouTube and TikTok adopt similar safeguards, which means Meta wrote its own competitors into its penalty and surrenders no competitive ground if the whole industry signs on.”

The agreement does indeed expressly state that it constitutes no admission by Meta of “liability, wrongdoing, or violation” of law, while its obligations generally expire after 10 years.

“The scale tells you who blinked,” Rubinowitz said. “Meta reported $201 billion in revenue for 2025, so ten annual payments of about $1.7 billion works out to three days of revenue a year.”

The exact annual amount is more complicated because a substantial portion of the headline maximum is contingent rather than guaranteed. But his larger comparison is hard to miss: Meta’s annual revenue is about 12 times the entire headline value of the decade-long settlement.

The Deal Changes How Teens Use Meta — Not How Meta Works

The settlement is not nothing. Far from it.

Teenagers will encounter harder time limits, overnight restrictions, quieter school days, tougher age checks, fewer visible popularity metrics and restrictions on cosmetic-surgery filters. Parents receive greater control. An independent auditor receives access. And the agreement gives state attorneys general a mechanism for enforcing those obligations.

The final consent judgment also leaves enforcement with the parties and the federal court, which retains jurisdiction to enforce or modify it.

Bonta celebrated accordingly. In one social-media post accompanying NPR coverage, he wrote that the settlement would “make social media less dangerous for our kids and make a world of difference for children and their families.” In another, he said the agreement would “change its platforms, enhance safety measures, and increase transparency,” adding: “Protecting our children can’t wait.”

But the settlement simultaneously resolves the states’ claims without a final adjudication of the allegations, Meta admits no wrongdoing, and some of the design mechanisms that critics regard as central to compulsive use remain alive and well.

That is why Rubinowitz sees the agreement less as the demolition of Meta’s engagement machine than the installation of operating hours.

“What the states collected here is a toll rather than a reckoning,” he said. “Meta pays the fee, admits nothing, and the road it built for exploiting children stays open.”

And while the attorneys general are heading home with a multibillion-dollar agreement, the general legal war is not over. Related litigation from families, individuals and school districts continues; Meta’s SEC filings describe extensive litigation and mass arbitration demands concerning alleged social-media addiction.

Rubinowitz believes those remaining plaintiffs may now become the test of whether a court will go further than the states were prepared to.

“The families and the more than 1,300 school districts still headed to trial are now carrying the fight the states walked away from,” he said. “I hope they press it to the end.”

For now, both sides can claim something.

The states can point to billions of dollars and protections that did not exist before. Meta can point out that it admitted nothing, avoided the remainder of a dangerous trial and kept much of the machinery underneath Facebook and Instagram intact.

And teenagers?

They can still scroll.

Just, eventually, with a meter running.