Westlaw's Case Law Is Free. The Cage Around It Is What You Pay For.

Pull up any reported court decision - a Supreme Court opinion, an obscure appellate ruling from 1974 - and you can copy it, reprint it, sell it, train a robot on it. It belongs to no one. It is, by law, the property of the public. So here is the puzzle at the center of one of the most durable monopolies in professional software: if the raw material of Westlaw is free for anyone to take, why can no one build a cheaper Westlaw? Thomson Reuters has an answer, and in February 2025 a federal court wrote it down.

The official story is that Westlaw dominates because its case-law database is uniquely complete. It isn't. The opinions inside are public domain and freely copyable. What is proprietary - and what the law protects - is the thin editorial layer Thomson Reuters wrapped around them: the headnotes that summarize each ruling and the West Key Number System that files them. The product isn't the law. It's the index to the law. And you can't buy the index anywhere else.

The cage is the headnote, not the case

Westlaw's real asset is a filing system. For more than a century, West editors have read each opinion and written a short summary of every legal point it decides - a headnote - then slotted that point into the Key Number System, a hierarchical index of legal topics built since the 19th century.⁶⁷ A lawyer who learned research on Westlaw doesn't think in opinions; they think in Key Numbers. That is the lock-in's deepest layer: not a contract, but a habit of mind. The competitor's problem is that you cannot offer the same shortcut without copying the same headnotes - and the headnotes are not free. When Ross Intelligence, an AI legal-research startup, asked Thomson Reuters for a license to train on Westlaw content, the answer was no, precisely because Ross was a competitor.⁶ So Ross built its training set from Westlaw material anyway, and Thomson Reuters sued.

In February 2025, a U.S. District Court in Delaware granted Thomson Reuters partial summary judgment. Westlaw's headnotes, the court held, are copyrightable. Ross's fair use defense failed because its use was not transformative and harmed Thomson Reuters' potential derivative markets - the AI-training market it might itself enter.⁵ The ruling did Thomson Reuters one better than a private contract: it made the moat a matter of federal copyright law, enforceable against anyone, forever.

That parenthetical is the whole game in four words. 'Not that creative' - and yet protected, and yet un-licensable to a rival. The genius of the moat is that it doesn't require brilliance. It requires only that the editorial layer be original enough to copyright and entrenched enough that lawyers can't research without it. Thomson Reuters didn't build a better mousetrap. It built the only mousetrap a court will let you own, around a mouse everyone is free to chase.

Why 96% of the money comes back every year

A legally protected switching cost shows up in exactly one place: the renewal rate. In its Legal Professionals segment, Thomson Reuters' recurring revenues climbed to 96% of total segment revenue in the fourth quarter of 2023, with 7% organic growth.¹ Read that slowly. Ninety-six cents of every dollar that segment earns is a subscription that renewed - the revenue equivalent of a customer who never leaves. Earlier in the same year the figure ran at 94%, with an adjusted EBITDA margin of 44.6%.² Software-company economics, on a base of public-domain documents.

The pricing machinery is built to keep it that way. Westlaw's published 'Advantage' plans start at $256.75 a month for single-circuit coverage and $399.75 for all states plus federal - but those prices are for new buyers only. Existing customers cannot see published online pricing at all; they must call Thomson Reuters to learn what they pay.⁸ That opacity is not an accident of a clunky website. A customer who cannot easily compare their own renewal price to the market is a customer who cannot easily decide to switch. The cage has a fog machine.

Buying the escape hatch before anyone else could

Generative AI looked, briefly, like the thing that could break the cage. If a chatbot can read every public-domain opinion and answer your question directly, who needs a century-old index? Thomson Reuters' response was not to defend the old moat but to buy the new one. In August 2023 it completed the acquisition of Casetext for $650 million in cash. Casetext already served more than 10,000 law firms and legal departments, and its flagship product, CoCounsel, was an AI legal assistant running on GPT-4.³ This is not a content company slowly waking up to AI. The company committed to spending $100 million-plus a year on AI integration and, by outside reporting, earmarked up to $10 billion for AI-focused acquisitions through 2025.⁴

Isn't this just a monopoly the law happened to bless?

The fair objection is that this is rent dressed as innovation - a toll on documents the public already owns, protected by a copyright the court itself called 'not that creative.' There's truth in it. But two things complicate the easy outrage. First, no U.S. government antitrust action has successfully challenged Thomson Reuters' dominance as unlawful monopoly behavior, and the 2025 ruling was a copyright case, not an antitrust loss for the company - the contested antitrust counterclaim is not what produced the headline.⁷⁵ What looks like a monopoly verdict was actually a property-rights verdict. Second, the stickiness is not as uniform as the 96% figure suggests: that peak belongs to the Legal Professionals segment specifically, not to the whole company, where other segments renew at materially lower rates. The honest read is that the deepest lock-in lives where the editorial layer is densest - law - and thins out where Thomson Reuters sells more commoditized data.

And yet the structure holds, because the thing that protects it is rare: a piece of intellectual property that is simultaneously essential to a profession's daily work, original enough to copyright, and impossible to license from anyone else. Thomson Reuters never needed to own the law. It needed only to own the path everyone takes through it - and then to get a federal court to agree that the path was its to keep. The opinions stay free. The way in costs $399.75 a month, and the renewal notice doesn't list a price.