Meta agreed last Wednesday to pay up to $17.1bn to 47 states, the District of Columbia and several US territories, and to rebuild the way teenagers use Instagram and Facebook: a two-hour daily cap, a blackout from midnight to 6am, notifications silenced through the school day, “like” counts switched off, cosmetic-surgery filters withdrawn, and an outside auditor marking the homework every year for a decade.
It is the largest state consumer-protection settlement since Big Tobacco. It is also, on the evidence of the document itself, one of the shrewdest deals a defendant has struck in the history of American technology litigation.
Begin with proportion. The four states that took Meta to trial in Oakland this month had asked for something near $200bn in civil penalties. They settled for a base of about $12bn spread over 10 years, with a further $5.1bn withheld against a condition Meta itself devised.
The company’s market capitalisation is around $1.45tn, which makes the maximum penalty roughly 1% of the firm. Meta earned $15.85bn in profit in a single quarter this year. “$17bn sounds like accountability, until you look at the numbers,” said Lisa Strohman, the clinical psychologist who founded Digital Citizen Academy, a US body working on technology addiction. Annualised, this is a line item.
Money, though, was never the point. Tennessee’s attorney general Jonathan Skrmetti, one of the two lead negotiators for the states, put it plainly: the states wanted “to protect kids and we wanted to reform the platforms.” The reform is what cost Meta, and the reform is what makes the deal historic.
How it came about is a study in leverage. In January, Meta hired CJ Mahoney, a former Microsoft general counsel and a lead negotiator of the Trump administration’s North American trade deal, as chief legal officer reporting directly to Mark Zuckerberg, Meta’s founder and chief executive. Within days he was on the phone to Skrmetti and his Colorado counterpart Phil Weiser, hunting a grand bargain.
The talks stalled for months over one thing: Meta would pay, but it would not redesign. Then it began losing. A Los Angeles judge ruled that a teenager’s addiction claim was a product-liability case, not a speech case, and refused to dismiss it; Zuckerberg spent a day on the stand insisting he was being mischaracterised, while internal documents showed his own staff likening themselves to drug dealers. New Mexico’s consumer-protection judgment followed.
On August 6, Mahoney flew to Nashville with a $19bn offer and went home empty-handed. Four days later the Ninth Circuit, the federal appeals court covering California, refused to throw the addiction cases out on Section 230 grounds, and the floor gave way. Daily negotiations began. By the time Adam Mosseri, the head of Instagram, had finished testifying, Zuckerberg’s own second appearance was scheduled and the company had run out of reasons to fight.
That sequence matters more than the number, because it marks the collapse of a defence that held for two decades. Section 230 and the First Amendment protect what users say and how platforms carry it. They say nothing about how a product is engineered to keep a fourteen-year-old awake. By recasting infinite scroll and variable-reward notifications as defective design rather than protected expression, the plaintiffs walked around the shield instead of at it. California’s attorney general Rob Bonta was entitled to his victory lap: the two laws, he said, “are not impenetrable shields to holding Meta accountable.”
And yet the most consequential clause in the settlement is not a concession. It is a weapon. Meta pays the withheld $5.1bn only if TikTok and YouTube accept comparable limits and comparable penalties; if they do, Meta’s own teen cap tightens from two hours to one. The company published an open letter the same morning and bought full pages in three American newspapers to press the case. “This framework will only work if all our peers join us,” Mahoney wrote.
Read that again. A regulator has just converted the target of its enforcement action into a lobbyist for extending that enforcement to the target’s competitors. The logic is not mysterious. Restrictions borne by Meta alone are a handicap, and teenagers locked out of Instagram at midnight will simply migrate. Restrictions borne by everyone are a floor beneath the entire market, and floors are what incumbents build. Compliance behaves like a fixed cost: age-assurance systems, auditors, parental-control architecture, six-hour response guarantees on abuse reports in two languages. Meta will absorb it without blinking. A three-person start-up will not. Consumer-protection law has quietly delivered the barrier to entry that antitrust law exists to prevent.
There is a second paradox, smaller and sharper. Part of the original grievance was that Meta harvested too much data on children. The remedy obliges Meta to identify children far more accurately than it does today. We are protecting minors’ privacy by requiring a company to know precisely who the minors are.
Notice, too, what survived. The recommendation engine is untouched. Minors keep personalised feeds; they merely gain the option of a chronological one, which few will choose. Haley Hinkle, policy counsel at Fairplay, a US advocacy group for children’s online privacy and safety, welcomed that option while noting what it dodges: personalised recommendation to minors remains “a huge driver of excessive time online.” The states capped consumption and left the machine that drives it running. They legislated the packet, not the leaf.
The conservative objection deserves an airing. No legislature debated whether two hours was the right number or midnight the right curfew; prosecutors and a defendant settled it over bagels in Nashville, and an auditor will now police a national technology policy that Congress never wrote. Parents buy the phones and pay for the wi-fi; teenagers have evaded house rules since long before Zuckerberg was born; Australia’s under-16 ban has not abolished adolescence. All true, and all beside the point. Congress held more than a dozen hearings across a decade and produced nothing. Regulation by litigation is what happens when legislation by legislature does not. The tobacco master settlement was drafted the same way, for the same reason, and it changed an industry.
So what does the verdict portend? Three things. Product-liability pleading is now the standard key to the American platform economy, and it will be turned next on AI companions, on algorithmic pricing, on whatever is built to hold attention rather than merely to host it. Governance by consent decree is now the default mode, with fifty attorneys general functioning as the regulator the United States declined to create. And Meta has demonstrated that a defendant with enough cash can convert a judgment into an industrial standard of its own drafting.
The settlement runs 10 years. The teenagers it protects were toddlers when Instagram appeared. It is entirely possible that by 2036 the app they cannot open after midnight will be a curiosity, and that the rulebook Meta wrote for its rivals will bind an industry Meta no longer leads. That would be the tidiest outcome of all: a company that paid one per cent of itself to make sure everyone who comes after has to pay the same.
- The writer is Deputy Managing Editor, Gulf Times.