A Judge Called Meta a Polluting Factory. Then Made It Pay for Therapy.

A New Mexico judge has declared that Meta's social media platforms are a "public nuisance" and ordered the company to pay $567 million to abate the damage. That is the legal category usually reserved for illegal landfills, blocked waterways, and actual environmental contamination. The kind of thing you call when a factory has been dumping chemicals into a river.

That was weird. The basic point is that a court just decided Meta is basically a polluting factory whose product creates toxic runoff — except the runoff is teen depression and sexual exploitation, and the abatement remedy is mental health treatment.

The ruling came down late Thursday, August 6, from Chief District Judge Bryan Biedscheid in Santa Fe. It's the second phase of a trial that began in 2023. In March, a jury found Meta willfully violated the state's Unfair Practices Act and assessed $375 million in civil penalties — $5,000 per violation across what the jury counted as 75,000 separate infractions. Now, in the public nuisance phase heard without a jury, the judge added $567 million more, pushing the New Mexico tab to nearly $1 billion.

The $567 million is broken down like a state health budget: $420 million for treatment services for young people, $90 million for screening and assessment, $33 million for awareness and prevention, $15 million for referral and coordination, and $9 million for implementation and evaluation. The money goes into a state-managed fund, to be spent over five years. The state had originally asked for $953 million. The judge gave them a bit less.

But here is the part that matters structurally. The attorney general wanted the judge to actually change how Meta's products work — strip out infinite scroll, disable autoplay, remove algorithmic feeds. The judge said no. He refused to order changes to core engagement features, ruling that doing so would risk violating the First Amendment and Section 230 (the 1996 law that shields platforms from liability for user-generated content), and would competitively damage Meta to an "unreasonable extent." He also couldn't force real age verification, because federal privacy law — COPPA — bars courts from mandating that companies collect personal data from children under 13.

So the judge found Meta guilty of creating a public nuisance comparable to air pollution, then made the company write a check for treatment services while leaving the factory's production line untouched. The state gets a fund. Meta keeps its algorithms. The thing that supposedly created the harm continues to create it, just with a line item on the balance sheet attached.

The operational changes the judge did order are narrower but still concrete: Meta must eliminate push notifications for users under 18 between 10 p.m. and 7 a.m., and during school hours on weekdays. It must impose a mandatory 90-hour monthly usage cap for teen accounts on Facebook and Instagram. Like counts are hidden by default for under-18 users. Teen accounts are private by default. Unconnected adults can't message teens. Meta must develop an AI-based "under-13 prediction model" within two years, request age proof from suspected underage users, and delete the accounts of those who can't verify within 30 days.

WhatsApp was excluded from all of this. The judge noted it doesn't contribute to the nuisance. Which makes sense if you think about what the nuisance actually is — not encryption, not messaging, but algorithmic engagement features designed to keep people scrolling. WhatsApp has none of that. It's a phone in your pocket that does one thing. The court's own remedy implicitly confirmed what the product theory of the case was all along.

Meta vowed to appeal. The stock dropped less than 0.5% on the news, trading around $590 on Friday. That is not because investors thought $942 million in combined New Mexico liability was immaterial — it's not, but it's also not $942 million worth of new information. The first-phase jury verdict came in March, when the stock fell 6% in a single day. Since then, the market has been pricing a pipeline, not a single judgment.

The pipeline is the real number to stare at. There are roughly 2,893 cases consolidated in a federal multidistrict litigation in the Northern District of California. About 800 of those are school district claims. The rest are individual personal injury lawsuits filed by parents and adolescents. More than 40 state attorneys general have filed their own suits. Snap and TikTok settled the first federal bellwether case (Breathitt County, Kentucky) before it went to trial in June, along with YouTube and Meta — Meta's share was reportedly around $9 million. The next bellwethers, Tucson and Charleston school districts, are scheduled for February 2027.

The legal mechanism making all of this possible is narrower than most headlines suggest, and more consequential. Section 230 still protects platforms from liability for what users post. But courts are increasingly allowing claims to proceed when the harm is attributed not to third-party content but to the platform's own design choices — infinite scroll, autoplay, variable-reward notifications, defective age gating. In April, the Massachusetts Supreme Judicial Court held explicitly that Section 230 doesn't immunize platforms from unfair business practice claims arising from addictive design features. The harm, the court said, resulted from the platform's own prolonging of user time, not from third-party statements.

This is an old legal move dressed in new clothing. It's the same distinction courts have drawn for decades in products liability: you're not responsible for what the customer does with your product, but you might be responsible if the product itself is designed in a way that causes harm. The tobacco cases in the 1990s worked on a similar theory — not that nicotine was illegal, but that the industry misrepresented its safety and engineered addiction.

The question for an investor isn't whether Meta will ultimately pay hundreds of millions or even low billions across these cases. It's whether the design-defect theory becomes binding appellate precedent that survives appeal, and whether it forces changes to the engagement features that are literally the product. If Meta loses on appeal and a court orders structural changes to the feed algorithm, the notification system, or the recommendation engine, that is a margin problem. If Meta just writes checks, it's a cost of doing business — large, ugly, and annoying, but absorbed into a company that reported roughly $60 billion in annual profit last year.

The $9 million Breathitt County settlement suggests the companies know they can buy off individual bellwethers before they set bad precedent. The $567 million New Mexico judgment suggests states can extract meaningful payments without actually changing the product. Both outcomes keep the machine running.

Meta's business model is attention extraction, monetized through advertising. The engagement features plaintiffs call "addictive by design" are the same engagement features that drive ad revenue. A court that orders Meta to pay for the consequences of those features while leaving the features in place is essentially treating them as a tax. The product stays. The price gets adjusted.

That is the structural answer to the question that matters. The verdicts so far are penalties, not redesigns. The factory keeps operating. The abatement fund is a surcharge, not a shutdown. Whether that changes depends on whether an appellate court takes the next step — not just ruling that Meta's design caused harm, but ruling that the design itself must change. That would be the moment the case stops being a tax and starts being a product problem. Until then, the liability is real, the pipeline is long, and the machine keeps running.