The court copied Meta's own bedtime , then added school

What the court actually ordered, and what it refused to touch#

On 7 August 2026 the First Judicial District Court in Santa Fe entered final judgment in State of New Mexico v Meta Platforms, D-101-CV-2023-02838. A jury had already found 75,000 violations of the state's Unfair Practices Act at the maximum $5,000 each. Judge Bryan Biedscheid then added a $567,000,000 abatement fund on top, and ordered five years of supervised product change.

Read the New Mexico v Meta remedy and a pattern appears. The court declined to order hard age verification, reasoning that a choice of that scale belongs with legislators. It declined to touch autoplay, infinite scroll or algorithmic recommendations, finding in the judgment itself that those features are too close to how content is presented and therefore tangled with the First Amendment and Section 230. It went nowhere near the ad auction, and nothing in it touches paid social advertising directly.

What it did touch was the clock. Paragraph 170 of the order is unambiguous: Meta "shall restrict the usage of all such users to not more than 90 hours of use per month cumulatively across Facebook and Instagram". Paragraph 169 hides like counts by default for under-18s unless a parent permits otherwise. And paragraph 168 does the thing everyone has half-reported.

New Mexico v Meta copied Meta's own bedtime almost exactly#

Here is the part worth the coffee. Paragraph 168 orders Meta to eliminate push notifications to under-18s "from 10:00 PM to 7:00 AM (i.e. hours when the vast majority of children are sleeping or should be asleep) on all days".

Now open Meta's own announcement of Instagram Teen Accounts from September 2024. Its words: "Sleep mode will be turned on between 10 PM and 7 AM, which will mute notifications overnight and send auto-replies to DMs". The court did not invent an overnight window. It took the one Meta had already built, and removed the option to switch it off. Anyone planning paid social advertising against this audience has been prowling the same trail for two years without noticing the fence.

That last row deserves a moment. Meta stripped teen targeting itself, three years ago. In its January 2023 announcement: "Age and location will be the only information about a teen that we'll use to show them ads", with gender removed and on-platform engagement no longer feeding the ads a teen sees. Search Engine Land told advertisers at the time they would need to rework their approach.

So the targeting fight for this audience was over before it started, and facebook ads targeting for under-18s has meant age and location for three years. What the court has now taken is the other half: not who you may reach, but when the platform may go and fetch them. That is a different lever, and almost nobody in paid social advertising has a plan for it.

Meta is appealing, and that matters#

Andy Stone, Meta's communications chief, told Forbes: "We disagree with the ruling and will appeal." Treat every figure in New Mexico v Meta as a live order under challenge, not a settled fact. The reason to read it now is that it is the most detailed public description of what a court thinks a defensible teen product looks like, and other states have said they intend to follow.

The second window nobody reported#

Every account of New Mexico v Meta we could find describes one notification ban. TechCrunch, on the day: "Push notifications to underaged users in the state are also to be paused between 10 p.m. and 7 a.m., and their usage should be limited to 90 hours a month". Accurate, and half of New Mexico v Meta. The missing half is the one that moves paid social advertising.

The New Mexico Attorney General's own press release lists the reform as "Eliminating push notifications for users under 18 during overnight hours". Also accurate. Also half. Two careful sources, one quiet omission, and a scent nobody followed.

New Mexico v Meta itself, at paragraph 168, has two limbs. The overnight one, and then this: "(b) from 8:00 AM to 3:00 PM (i.e., typical school hours) during the academic year, excluding weekends". Meta may still send messages from connected users and urgent, targeted alerts such as security or hazard warnings. Everything else stops for seven hours of every school day.

The evidence the court leaned on is not thin. The US Surgeon General's 2023 advisory records that up to 95% of 13 to 17 year olds use a social platform, that a survey of 8th and 10th graders put average use at 3.5 hours a day, and that a systematic review of 42 studies found "a consistent relationship between social media use and poor sleep quality, reduced sleep duration". Set the court's 90 hours a month against that 3.5 hours a day and the cap is a trim, not a demolition.

The night-time limb has its own literature. A cross-sectional analysis of 11,872 UK adolescents in the Millennium Cohort Study, published in BMJ Open, found very high social media users more likely than average users to report late sleep onset on school days (OR 2.14, 95% CI 1.83 to 2.50) and trouble falling back asleep after waking in the night (OR 1.36, 95% CI 1.10 to 1.66). A notification at 23:40 is the mechanism in that second number, and it is the moment a nocturnal audience is manufactured rather than found.

Run the arithmetic and the shape is stark. Strip 22:00 to 07:00 and fifteen hours remain. Strip 08:00 to 15:00 as well, on a term-time weekday, and the platform's permitted window collapses to roughly eight: one hour before school and seven in the evening. The teenage attention a media plan quietly assumed was spread across the day is now legally concentrated into after-school and evening.

For paid social advertising this is a compression, not a prohibition. Ads are not notifications, and the order does not touch ad delivery. But notifications are what pull a teenager back into the app, and sessions are where impressions live. Squeeze the re-entry points into eight hours and the inventory follows them. Anyone buying reach against a young audience in New Mexico should expect the daily curve to sharpen, not flatten.

What this does to a media plan, and to paid social advertising#

Three consequences of New Mexico v Meta for paid social advertising, in the order you will meet them.

The first is arithmetic. If your audience skews young and your delivery is even across the day, your frequency curve is about to bunch. Dayparting stops being a nicety in paid social advertising and becomes a description of reality, and the quarry is simply somewhere else at noon. The second is measurement: New Mexico v Meta, a five-year, state-specific product change, makes New Mexico a natural control against neighbouring states, which is the closest thing to a free experiment paid social advertising has been handed in years, a clean trail through otherwise thick undergrowth.

The third is contagion. The Attorney General has said plainly that he wants age verification put into statute rather than left in a court order, and that other states can follow the route. Deseret News reported the fund's allocation and the finding that the platforms' purpose and effect is to drive engagement in ways detrimental to teenagers. Plan for a patchwork, not a one-off.

Regulated categories, read this twice#

If you sell anything age-gated, this cuts the other way and in your favour. In New Mexico v Meta, a court has just written down, in numbered paragraphs, what it considers a defensible under-18 experience. That is a free specification for any operator, lender or clinic being asked how it keeps minors away from its marketing. Our paid social work uses exactly this kind of document, and our healthcare practice has been arguing the same point about age signals for months.

The wider ruling on personalised feeds published by folkfox this week runs alongside this one: see the judge who called the feed a mirror. The audience-side consequences are worked through in age assurance and your advertising audience and in first-party data and audience decay. If the question on your mind is are meta ads still worth it, the cost side is in social media ad costs in 2026.

Five moves for paid social advertising in the next planning cycle#

Cheap, specific, and none of them requires waiting for the appeal. Treat this as a patient prowl rather than a panic.

None of this is a reason to abandon the channel or to abandon paid social advertising. Instagram ads still reach an enormous audience, and the order applies to one state, to under-18s, and to notifications rather than advertising. The reason to move early is that the cost of learning this on a live campaign in a state you did not break out is far higher than the cost of adding a reporting line this week. Fresh tracks are cheap to read; cold ones are not.

What has genuinely changed is the assumption underneath the plan. Availability used to be a platform decision, tuned by engagement and auction pressure. In one state, for five years, it is a court's decision, published in numbered paragraphs, with compliance reports filed twice a year. Read the paragraphs. They are more useful than the headline, they are the part nobody else has read, and they are where the next year of paid social advertising planning quietly begins. The den has been rearranged while everyone stared at the number on the door.

Frequently asked questions#