Meta’s press release put the settlement at $18 billion, but even the more conservative $17.1 billion figure other states cited is still big enough to eclipse one of Silicon Valley’s biggest AI bets of the past year. To put it in perspective, the $17.1 billion number is a little more than three times the roughly $5 billion personal stake that Alexandr Wang held in Scale AI, a data-labeling company that supplies the human-annotated training data AI models are built on. Last year, Meta paid $14.3 billion for a 49% stake in the company last year and brought in Wang to lead its AI efforts of its new Superintelligence Labs, reporting directly to Mark Zuckerberg.

Put simply: the fine for allegedly hooking kids on Meta’s apps costs about three Alexandr Wangs.

For additional context, the company posted $60.46 billion in net income on $200.97 billion in revenue for full-year 2025, meaning the settlement equals roughly 27% of one year’s profit and about 8% of annual revenue. Meta has continued to spend aggressively even as its legal exposure mounted: the company raised its 2026 capital expenditure guidance to as much as $145 billion, driven largely by its AI buildout.

The company denied wrongdoing and it previously argued the states’ financial demands were “vastly disproportionate,” and in pretrial filings warned that the states’ own damages framework could theoretically produce penalties as high as $1.4 trillion—a figure close to Meta’s entire market capitalization. The states’ lawyers had signaled roughly $200 billion was a more realistic target at trial.

Averting a landmark trial

To put it in perspective, the settlement’s most unusual feature is that roughly $5 billion of the total isn’t guaranteed. Meta’s own announcement puts the total at $18 billion, with states receiving approximately 70%, or $12.7 billion, in annual installments over 10 years regardless of what happens elsewhere in the industry. The remaining 30%, roughly $5.3 billion, is released only if two conditions are met: TikTok and YouTube adopt matching daily time limits, night mode restrictions and age-verification measures, and each of those companies pays a matching sum, split evenly against the contingent pool. Some state attorneys general have cited a slightly lower total, $17.1 billion, built on a similar guaranteed-plus-contingent structure—a roughly $12.1 billion floor plus an additional $5 billion contingent on the same industry-wide adoption. The discrepancy likely reflects differences in how each side scoped the settlement, rather than one figure excluding money the other includes.

The settlement resolves federal Children’s Online Privacy Protection Act claims from all 29 states active in the lawsuit, along with separate state consumer-protection claims that California, Colorado, Kentucky and New Jersey were actively trying before U.S. District Judge Yvonne Gonzalez Rogers in Oakland.

Opening arguments had begun just over a week earlier, with California Deputy Attorney General Megan O’Neill telling the court that “Meta’s business model can be summed up in four simple words: ‘hook’ the users, ‘hold’ them for as long as they can, ‘harvest’ their data, and then ‘hide’ the truth from the public when making public statements.” She added, “It was especially bad for kids.”

As part of the deal, Meta agreed to nationwide safeguards for teen users of Facebook and Instagram, including daily usage limits and nighttime blocks.

Eclipsing Meta’s own record

The settlement more than triples Meta’s previous high-water mark: the $5 billion penalty the Federal Trade Commission imposed in 2019 over Cambridge Analytica-era privacy violations, which regulators at the time called “almost 20 times greater than the largest privacy or data security penalty ever imposed worldwide.”

Across the tech sector more broadly, the new settlement exceeds the EU’s four separate antitrust fines against Google—on search, Android, ad-tech and shopping—which together total roughly $12 billion over nearly a decade. It’s more than 10x Anthropic’s $1.5 billion payout to authors, the largest copyright settlement in U.S. history, and more than 10x Google’s $1.375 billion privacy settlement with Texas last year (Meta had a $1.4 billion settlement of its own with the state).

Wednesday’s deal caps a brutal year for Meta in the courts. A New Mexico jury found in March that the company had willfully violated state consumer-protection law by concealing what it knew about child sexual exploitation on its platforms, awarding $375 million in penalties.

As the case lingered in between phases, New Mexico Attorney General Raul Torrez, who has pursued Meta aggressively, criticized the company in April 2026 for threatening to shut down in the state rather than install safeguards: “Meta is showing the world how little it cares about child safety,” Torrez said. “Meta’s refusal to follow the laws that protect our kids tells you everything you need to know about this company and the character of its leaders. We know Meta has the ability to make these changes. For years the company has rewritten its own rules, redesigned its products, and even bent to the demands of dictators to preserve market access. This is not about technological capability. Meta simply refuses to place the safety of children ahead of engagement, advertising revenue, and profit.”

In March, a Los Angeles jury delivered the first verdict of its kind, finding Meta and Google’s YouTube negligent for designing their platforms to be addictive to children, in a case brought by a then-20-year-old plaintiff identified only as Kaley, or K.G.M., who said she became compulsively hooked on Instagram and YouTube as a child and suffered resulting depression, anxiety and suicidal thoughts. Jurors awarded her $6 million total—$3 million in compensatory damages and $3 million in punitive damages—after concluding both companies knew their products could harm minors and failed to warn users, with Meta shouldering 70% of the liability and Google 30%. The verdict was significant on principle: it marked the first time a jury had treated social media apps as defective products engineered to exploit developing brains, validating a legal strategy that targets platform design rather than content.

Kaley’s case was also the first of nearly 2,500 plaintiffs in a consolidated Southern California proceeding against Meta, Google, TikTok and Snap, meaning the verdict served as an early bellwether for the wave of similar suits still working through the courts. Meta said in a statement it “respectfully disagree[d] with the verdict” and would appeal, arguing that “teen mental health is profoundly complex and cannot be linked to a single app.”

In August, a New Mexico judge added another $567 million, ruling that Meta had created a “public nuisance” similar to air pollution. Two more bellwether cases remain scheduled for trial in October, and Meta faces thousands more similar lawsuits that are currently pending.

Whether Wednesday’s settlement slows that pipeline of litigation, or simply removes the most immediate and costly case from Meta’s docket, remains to be seen when the next bellwether trials begin this fall.

What Meta is actually changing

Pending judicial approval, Meta says the agreement will bring a specific set of default protections to under-18 users of Instagram and Facebook, most of which must remain in place for 10 years:

• A default two-hour daily time limit, cumulative across both apps, that teens can only disable with a parent’s permission.
• A default night mode blocking Facebook and Instagram use between midnight and 6 a.m.
• Muted notifications between 8 a.m. and 3 p.m. on school days, aside from direct messages and safety alerts.
• Usage prompts after every 15 minutes of continuous use, and again at the 60- and 90-minute daily marks.
• An option for a non-algorithmic, non-personalized default feed, which parents can require.
• Hidden like counts, disabled autoplay by default, and a ban on extreme makeup filters in addition to Meta’s existing cosmetic-surgery filter ban.
• Expanded age-verification technology and new parental alerts when a teen links a secondary account or interacts with a flagged adult account.

Notably, Meta structured its own commitment on a sliding scale tied to industry adoption. The Time Limit and Night Mode provisions start on a five-year commitment at the levels above; if TikTok and YouTube sign onto the same framework, Meta will extend those commitments to 10 years and tighten them further, to a one-hour daily limit and a 10 p.m.–7 a.m. night block. Meta also published an open letter Wednesday explicitly calling on TikTok and YouTube to adopt the same standard, arguing that “when teens are restricted on one app, they simply move to another.”

The agreement additionally creates an independent research foundation, to which Meta will contribute consented user data for studies on teen well-being, and calls for an independent auditor to assess Meta’s compliance annually for five years.

C.J. Mahoney, Meta’s chief legal officer, framed the deal as an industry challenge as much as a settlement: “Our new Time Limit commitments, Night Mode features and usage limits during school hours set the right path forward for our whole industry, but this framework will only work if all our peers join us,” he said. “We therefore call on our industry peers, TikTok and YouTube, to implement this new framework, right away.”

For this story, Fortune journalists used generative AI as a research tool. An editor verified the accuracy of the information before publishing.