In late August 2026, US District Judge John Cronan dismissed X Corp.'s First Amendment challenge to New York's Stop Hiding Hate Act, holding that "the First Amendment poses no obstacle" to requiring large platforms to disclose how they moderate hate speech, extremism and disinformation (Insurance Journal/Reuters). The case is X Corp. v. James, No. 25-cv-05068 (S.D.N.Y.). Press reports describe the dismissal as with prejudice. The ruling matters beyond Musk's company, because it marks a boundary between regulation that speech advocates can live with and regulation they should keep fighting.
The strongest case for the law
Supporters have a serious argument. Platforms make consequential, largely invisible decisions about what hundreds of millions of people see. Users, researchers and advertisers cannot compare services if each describes its rules in vague terms. A disclosure mandate does not remove a single post. It asks companies to say truthfully what they already claim to do. Judge Cronan adopted this reasoning. He said the law requires companies only "to speak truthfully about their offerings," which serves the free-speech goal of "the discovery of truth" (Insurance Journal/Reuters).
According to press accounts, the law applies to platforms with more than $100 million in annual revenue. They must file reports on how their terms define and enforce categories such as hate speech, extremism, disinformation, harassment and foreign political interference. Penalties reach $15,000 per violation per day for missing, incomplete or misleading reports, with a 30-day cure period (Reclaim The Net).
Why this differs from California's failed law
The result is not a blank check for states. In X Corp. v. Bonta (September 4, 2024), the Ninth Circuit held that California's AB 587 was likely unconstitutional. The statute required platforms to report how, if at all, they moderate contested categories of speech. The court treated those reports as compelled non-commercial speech that was content-based and not narrowly tailored (Ninth Circuit opinion). The Southern District of New York reached the opposite result on a similar subject matter. Cronan reasoned that the reporting obligation concerns "purely factual and uncontroversial information" about a company's own terms of service, which is compatible with the First Amendment.
This is a real split in reasoning, not a quirk. Two courts in two circuits are treating similar disclosure statutes differently. The outcome turns on whether the compelled content is framed as facts about the product or as a state-prompted statement about contested categories. The Second Circuit will likely get to say which framing is right, if X appeals. Companies and legislators should treat the question as unsettled.
The companion ruling that shows how courts police drafting
The disclosure ruling follows a June 23, 2026 decision of the New York Court of Appeals in Volokh v. James, which concerned a different statute, the Hateful Conduct Law (General Business Law § 394-ccc). That law requires platforms to maintain a mechanism for users to report hateful conduct. Answering three questions certified by the Second Circuit, the court held that a platform can comply without explicitly referencing the statutory definition of hateful conduct, and that the statute does not require a platform to respond to a user's report (NY Court of Appeals, Volokh v. James). Commentators read this narrow construction as likely to help the law survive First Amendment review (Simpson Thacher).
The pattern is consistent. New York's statutes are surviving where they are read, or drafted, to compel mechanisms and factual disclosures. They are not surviving, or have not been tested, where they would pressure platforms to adjudicate what counts as hateful.
What this means for a pro-speech, pro-innovation policy
Three implications follow.
- Transparency is the least speech-restrictive tool on the menu. If lawmakers want to respond to online hate, requiring truthful descriptions of enforcement is better than takedown mandates, which push platforms to over-remove lawful speech to avoid liability. A disclosure regime preserves the open internet, because users and rivals can see the rules and choose.
- Disclosure only works if it is proportionate. The $100 million revenue threshold spares startups. That matters for innovation, since a regime with daily penalties and detailed category-by-category reports can become a compliance moat that only incumbents can afford. Regulators should keep thresholds high and report formats standardized.
- The risk is mission creep. Once a state collects how platforms define "disinformation," the temptation will be to judge those definitions. Cronan's reasoning rests on the absence of such judgment. Attorneys general who use reports to pressure platforms toward particular moderation outcomes would invite the strict scrutiny that doomed AB 587. Civil-liberties groups such as EFF have long argued that state pressure on moderation choices raises First Amendment problems, and enforcement practice will be the real test.
A fair reading of the ruling is therefore neither a defeat for free speech nor a victory for it. It confirms that courts will tolerate factual transparency about private moderation, and it signals that the next fight will be over enforcement, not text. Platforms that dislike the law have an obvious, speech-protective response: publish clear rules, follow them, and report accurately. Legislators who want to build on it should resist adding definitions of hate speech or duties to act on reports. Those additions would convert a transparency law into a speech regulation, and courts such as the Ninth Circuit have shown they will notice.