Everyone Calls Thomson Reuters a News Company. The Newswire Is a Small Fraction of It.

A lawyer in a glass tower in Chicago types a query into Westlaw and pulls a precedent in eleven seconds. She will never see an invoice for that search. Her firm pays a subscription, year after year, because the alternative - not having instant, authoritative access to case law - is unthinkable for a firm that bills by the hour. Multiply that quiet dependence across the legal, tax, and compliance professions of dozens of countries, and you get a business almost nobody describes correctly. People hear 'Reuters' and picture a newswire. They are looking at the wrong part of the company.

The official story is that Thomson Reuters is a news and financial-data company facing an existential AI threat. Strike most of that out. Reuters News is a small minority of the recurring base. The financial-terminal business was sold off: Thomson Reuters closed the sale of a 55% interest in its Financial & Risk unit to Blackstone in October 2018, with the remaining stake exiting when Refinitiv was sold to London Stock Exchange Group in January 2021.⁹¹⁰ What's left, and what generates the cash, is something far less glamorous and far more durable: a subscription machine for the tools professionals are legally and practically unable to work without.

Here is the thesis a smart friend could repeat at dinner: Thomson Reuters isn't a news company that also sells data. It's a legal-and-tax subscription business that happens to own a famous newswire - and the recurring nature of those subscriptions, not the brand, is the entire machine.

The wire everyone remembers is barely a rounding error

Walk through the actual revenue and the misunderstanding falls apart. In the full year 2024, recurring revenues grew 7% and made up 83% of total revenues, while the transactional and print pieces - the lumpy, declining bits - shrank.¹ The center of gravity is the Legal Professionals segment, where recurring revenues hit 97% of segment total in the fourth quarter of 2024, growing 8% organically on the back of Westlaw, Practical Law, and a newer AI assistant called CoCounsel.² That is not a media company's revenue profile. It is the profile of an enterprise software business wearing a press lanyard. The newswire still carries the famous name; the subscriptions carry the company.

Why a subscription beats a sale, every single year

The mechanism is worth working all the way down, because the magic isn't the data - it's the renewal. A one-time sale earns once and then begins decaying; you have to win the customer again next quarter. A subscription earns the same money next year as a starting point, before you've sold anything new. So a business that converts its revenue to recurring isn't just steadier - it compounds, because each year's growth stacks on a base that didn't reset. When 97% of a segment renews and the renewals themselves grow 8%, you are not running on a treadmill. You are climbing stairs.² And the reason customers renew is switching cost: a litigator who has built ten years of habit, saved searches, and trust into Westlaw doesn't price-shop a legal database the way she shops a stapler. The product is woven into how the work gets done. That weave is the moat, and every renewal pulls it tighter.

The owner who doesn't have to flinch

There's a second, less obvious gear, and it sits in the ownership structure. When Thomson Corporation completed its acquisition of Reuters on April 17, 2008 - and it was an acquisition, with Thomson the acquiror and Reuters the acquiree, not the 'merger of equals' the legend prefers - the Thomson family's vehicle, The Woodbridge Company, held about 54% of the voting interest.⁴⁵⁶ That stake has not shrunk. As of March 31, 2026, Woodbridge alone held 67.91%, and the broader Thomson family group held 70.56% of the shares outstanding - a position that grew not by buying more but because the company bought back its own stock, concentrating the family's grip without the family spending a dime to acquire it.¹¹ A controlling owner with a multi-decade horizon can do something a quarterly-driven board cannot: absorb a margin hit today - say, heavy investment in AI tools like CoCounsel - and wait calmly for it to pay off, while the market frets. The family's patience is itself a competitive asset.

That quote matters because it punctures a common comfort. People assume the storied 'Reuters First Principle' - no party may own more than 15% - constrains the Thomson family's control. It doesn't. The principle was explicitly waived for this very purchase, with the Reuters Founders Share Company's chairman stating publicly that the financial circumstances of Reuters left no alternative.⁸ The Founders Share Company still guards editorial independence at the newswire, but it places no ceiling on Woodbridge's economic or voting power. The family controls the company. The 15% rule guards only the journalism - the smallest, least lucrative corner of the empire.

But isn't AI about to eat the legal database?

The honest objection is the one everyone leads with: large language models can now read case law and draft memos, so why pay for Westlaw at all? It's a real threat and worth taking seriously rather than waving away. But notice two things. First, an LLM that confidently invents a citation is worse than useless to a lawyer who can be sanctioned for filing it - and the asset Thomson Reuters sells is not raw text but vetted, authoritative, liability-bearing content. Trust is the product, and trust is exactly what a hallucinating model lacks. Second, watch what the company is actually doing: CoCounsel is a Thomson Reuters AI assistant, sold into the Legal segment alongside Westlaw and growing on the same recurring-revenue base.² The incumbent isn't being disintermediated by AI; it's bolting AI onto a base of customers it already owns. That is a far better position than the disruption narrative assumes. The fair counter is that this could still erode pricing power over time - but eroding a 97%-recurring base is slow, and a patient owner has the years to manage it.²

Thomson Reuters makes its money the way a tollkeeper on a single mountain pass does - not by being interesting, but by being unavoidable, and by collecting again every time the seasons turn. The newswire that gave it the famous name is the small fraction of the company everyone watches; the legal and tax subscriptions nobody outside the professions ever sees are the bulk that pay. And behind it all sits a family that owns 70% of the thing and answers to no quarter. The real question was never whether AI will disrupt the database. It's whether anyone betting against this machine understood, before they placed the bet, what they were actually betting against.