Well, that was quick. Not even a quarter of the way into an expected 19-day federal trial over Meta’s alleged failure to protect young users, the social media giant agreed to settle the case. The company has agreed to pay up to $16.7 billion to 51 US states and territories and make substantive product changes for how teens engage with its platforms.

Attorneys general representing some of the 29 US states that had taken Meta to trial announced the deal on Wednesday, shortly before Instagram head Adam Mosseri was set to return to the witness stand. In a separate announcement detailing the settlement, Meta urged YouTube and TikTok to adopt similar changes to their platforms.

The settlement is a stunning development; while other social media services, including YouTube, Snapchat, and TikTok, resolved similar lawsuits out of court, Meta has repeatedly chosen to go to trial. Earlier this year, it lost state trials in California and New Mexico, notching nearly $1 billion in penalties. The company’s first federal trial on the youth issues began last week and threatened to leave the company on the hook for what it estimated to be over $1 trillion in damages.

“Ensuring teens have a safe and productive experience on our platforms is an absolute imperative for Meta,” the company wrote in its announcement. “We want to get this right for parents and teens, and that’s why we partnered with state attorneys general to set a new industry standard.”

“Today, we have secured a settlement with Meta that will make social media less dangerous for our kids and make a world of a difference for children and their families,” said California attorney general Rob Bonta in a statement.

The product changes that Meta has agreed to will apply to Instagram and Facebook and will focus on both features and time limits. Pending judicial approval, teens on Instagram and Facebook will have a two-hour time limit by default across both platforms; they’ll need parental approval to turn it off. Those limits don’t apply, however, to “longform content”—defined as audio or video at least 22 minutes long—or messaging, meaning teens can still DM for as long as they please.

The apps will also be blocked between midnight and 6 am for teens by default, and notifications will be muted between 8 am and 3 pm—from August 15 to June 15—to minimize distractions during school hours. Teens will also get prompts every 15 minutes of continuous screen time, with additional notifications when they hit 60 and 90 minutes of use.

Parents will also have more control over how teens interact with Instagram and Facebook. They can set the default experience to a chronological feed rather than algorithmic and can similarly turn off autoplay by default. Similarly, Meta won’t display the number of likes and reactions on posts by default for teen users. (Bonta described these steps as a “ban” for teens rather than a default setting; it appears that parents will need to approve any changes from the default settings either way.) The company will also disable “cosmetic surgery and extreme makeup filters” for teens.

The $16.7 billion payout comes with an interesting asterisk. While around $12.7 billion of that is guaranteed and will be paid out over 10 years in annual installments, the remainder is contingent on other “core industry members”—defined as Snap, TikTok, and YouTube—implementing similar safeguards. Meta says those changes include a one-hour daily limit, a comparable night mode, and age assurance measures. (If adopted, Meta’s own daily limit will also go down from two hours to one.) TikTok and YouTube would also have to make a combined payment that’s comparable to the amount that’s outstanding from Meta.

All 29 states argued that Meta violated federal child privacy laws by collecting data on users under 13 years old. Four of the states—California, Colorado, Kentucky, and New Jersey—alleged Meta deceived parents with misleading statements about protections for adolescent users and that it engaged in unconscionable practices in designing some features for Facebook and Instagram. Meta denied all of the accusations.

The states began presenting their case by calling on current and former Meta employees to describe an alleged culture of maintaining or growing engagement above all else, to the detriment of young users. An academic psychologist testified that adolescents, and especially girls, who spent more hours on social media were more likely to be depressed or unhappy, which contributed to increases in teen loneliness and self-harm over the past 15 years.

This week, the attorneys general coalition zeroed in on Instagram’s well-being team, which has been developing features to address some of the concerns about teen social media use since 2018. Testimony from a former Meta data scientist, current product design director, and veteran executive Mosseri started to show how the states hoped to convince Judge Yvonne Gonzalez Rogers, and the eight jurors advising her, that Meta engaged in unfair and deceptive business.

On Tuesday, the focus was largely on a feature called Take a Break, which reminds teen users to limit extended use of Instagram. According to testimony, less than 1 percent of Instagram’s weekly teen users had actually activated the feature months after its launch. Two years in, adoption still sat at only 1.8 percent. Uptake for another usage-limiting feature, called Quiet Mode, was barely any better, documents presented during the trial on Tuesday showed.

As part of the settlement, Meta has also agreed to enforcement measures, including an independent auditor who can relay any concerns to the attorneys general and the development and deployment of more robust tools for parental supervision. The company will also be prohibited from “making further false, misleading, or deceptive statements around its safety features.”