Thomson Reuters Corporation (TSX, Nasdaq: TRI) reported third-quarter 2026 revenue of $1.83 billion, up 10% year over year, while organic revenue also increased 10% across both the total company and its Legal Professionals, Corporates and Tax, Audit & Accounting businesses. Management raised full-year organic revenue growth guidance to 8.5%–9.0% from approximately 8%, yet Thomson Reuters shares fell about 7% to $113.46 after investors focused on the growing cost of artificial intelligence investment.
The selloff is notable because there is little evidence that generative AI has damaged customer demand. Legal Professionals recurring revenue increased 9%, Thomson Reuters said CoCounsel and Westlaw contributed materially to organic growth, and management disclosed that approximately 34% of underlying contract value now relies on generative artificial intelligence, up from 32% during Q2 and 24% a year earlier.
The problem is profitability. Legal Professionals adjusted EBITDA margin fell to 46.6% from 49.0%, while the combined margin across the three largest segments declined to 40.3% from 41.7%. Investors therefore appear less concerned about whether Thomson Reuters can sell AI products than about whether it can do so while maintaining the steady margin expansion historically expected from a premium professional-information company.
Why is Thomson Reuters spending so aggressively on artificial intelligence?
Professional information markets are unusually exposed to generative AI because much of the value historically came from helping lawyers, accountants and corporate professionals find and interpret complex information. Large language models can now perform portions of search, summarisation, drafting and analysis rapidly, potentially threatening products built around conventional research workflows.
Thomson Reuters cannot defend Westlaw or its tax platforms merely by preventing customers from using AI. It instead needs to ensure that lawyers and accountants use AI inside Thomson Reuters products, powered by trusted proprietary content and professional workflows.
This explains the investment behind CoCounsel Legal and CoCounsel Tax, Audit & Accounting. The products increasingly operate as agents capable of conducting research, drafting documents and assisting professionals through multi-step tasks rather than simply returning search results.
That evolution requires spending on models, software engineering, content structure and product integration. The result may be a stronger competitive moat, but shareholders absorb the cost before the full revenue benefit arrives.
How much margin did the AI investment cost Legal Professionals?
Legal Professionals revenue reached approximately $788 million during the quarter, while adjusted EBITDA increased only 4% to $368 million. The segment’s margin declined 240 basis points to 46.6%, compared with 49.0% a year earlier.
Management attributed the pressure partly to technology and other costs, while Reuters also reported approximately $22 million of severance expense and costs associated with the Noetica acquisition. Not every basis point of the decline should therefore be described purely as AI expenditure.
The important point is that the largest and most profitable core business is deliberately investing faster than revenue. Thomson Reuters believes higher prices for premium agentic products and deeper customer adoption will ultimately offset the increased expense.
Investors are effectively being asked to accept lower near-term margin in exchange for stronger long-term competitive positioning. The 7% share-price decline shows that many shareholders want clearer evidence of how quickly that payback arrives.
Why is 34% of contract value tied to generative AI such an important number?
Chief Financial Officer Gary Bischoping said approximately 34% of Thomson Reuters underlying contract value depended on generative AI during Q3. The figure increased from 24% only a year earlier, demonstrating how rapidly AI features are becoming embedded in commercial contracts rather than remaining optional experimental products.
This is an important distinction from technology companies that report user activity without meaningful monetisation. Thomson Reuters is increasingly attaching AI to actual recurring customer agreements.
The percentage does not mean 34% of company revenue is directly generated by AI. Underlying contract value measures the contractual economics associated with products containing or depending on generative AI, and those products can include substantial conventional content and software functionality.
Still, the direction is powerful. If more contract value migrates toward AI-enabled tiers, Thomson Reuters may gain greater pricing power while making it harder for customers to replace individual tools with stand-alone consumer chatbots.
Why are professional customers still paying for Westlaw when generic AI is everywhere?
Legal and tax work places unusually high value on accuracy, provenance and professional liability. A lawyer cannot safely rely on a model that invents a judicial precedent or cites an outdated statute, particularly when the output enters a court filing or client advice.
Thomson Reuters owns highly structured legal, tax and regulatory datasets and can connect generated answers directly to authoritative materials. That combination provides an advantage over generic models whose underlying sources may be unclear or incomplete.
Workflow integration strengthens the moat further. A law firm may pay not simply for an AI response but for research, document drafting, citation checking, case-law history and collaboration inside one professional environment.
Scotia Capital analysts cited by Reuters said there was no evidence of demand slowing because of AI-native competitors. The larger threat may therefore be margin pressure from the investment required to make Thomson Reuters’ own products sufficiently advanced that customers have little reason to leave.
Why did Reuters revenue jump 12%?
The Reuters news division produced 12% organic revenue growth, partly because licensing revenue linked to generative AI increased sharply. Transaction revenue in the news business rose 56% to $44 million.
This creates an intriguing inversion of the copyright concerns surrounding generative AI. News organisations fear that AI systems can reduce traffic or reuse journalism without adequate compensation, but high-quality verified news content can also become increasingly valuable as model developers seek reliable data.
Thomson Reuters has one of the largest global news-gathering operations, giving it a dataset that cannot be replicated cheaply from synthetic content. Licensing that material to artificial intelligence companies could become a meaningful secondary revenue stream.
The economics remain different from recurring professional subscriptions because AI licensing can be transactional and potentially uneven. Investors should therefore avoid assuming Reuters’ current transaction growth repeats every quarter at the same rate.
Why does Thomson Reuters have $9bn to $10bn available for investment?
Management says the company has approximately $9 billion to $10 billion of investment capacity and sees a pipeline across AI, legal, tax and accounting assets. That balance-sheet flexibility follows years of portfolio reshaping, including the sale of a 51% interest in Global Print to KKR-managed funds for approximately $500 million.
Exiting majority ownership of print sharpens the group around digital professional information and software. It also removes a slower-growth business whose economics were increasingly different from AI-enabled subscription platforms.
The danger is acquisition discipline. Artificial intelligence assets are commanding high valuations, and professional-data companies can easily overpay for businesses justified by strategic buzzwords rather than durable revenue.
Noetica demonstrates the intended approach on a smaller scale, adding specialised AI data and analytics capabilities to legal products. Larger acquisitions would require far more scrutiny because the company’s recent share-price weakness suggests investors are already concerned about returns on incremental AI spending.
Why did free cash flow fall despite stronger revenue?
Quarterly free cash flow declined to approximately $408 million from $526 million a year earlier, even as revenue and adjusted EBITDA increased. Capital expenditure rose to $176 million and operating cash flow fell to $599 million from $704 million.
The company still expects roughly $2.1 billion of full-year free cash flow, so one quarter does not imply a structural cash problem. The decline nevertheless reinforces the market’s central concern that investment is absorbing more of the operating improvement.
Thomson Reuters historically appealed partly because recurring professional subscriptions supported reliable cash generation. AI investment changes that profile temporarily by requiring larger technology spending before products mature.
The quality of the strategy will eventually be measured by whether contract growth and pricing accelerate faster than the investment burden. If not, the company risks becoming a faster-growing business with structurally lower margins.
Could proprietary professional data become more valuable as AI models improve?
There is a strong argument that the answer is yes. Artificial intelligence can commoditise generic analytical capability while increasing the value of information that is difficult to reproduce, legally sensitive or continuously updated.
This dynamic resembles London Stock Exchange Group’s recent decision to buy Octus for proprietary private-credit information. In both cases, the defensive strategy is to own data that an outside model cannot simply obtain from public internet sources.
Thomson Reuters’ advantage lies in combining that proprietary information with workflow tools used by professionals every day. The company does not need to build the world’s largest foundation model if it can provide the most reliable context for specialised legal, tax and compliance tasks.
Management’s own Thomson AI model and strategic partnerships with major law firms represent another route. Chief Executive Officer Steve Hasker expects Skadden, Arps, Slate, Meagher & Flom to become the first of several major strategic AI partners, giving Thomson Reuters feedback from some of the highest-value users in the legal market.
What should investors watch after the 7% selloff?
Legal Professionals margin is the first number because it provides the clearest measure of whether AI revenue is beginning to outrun investment costs. Continued contraction would strengthen fears that premium products require permanently higher technology spending.
Generative-AI contract value provides the second test. A move materially above the current 34% would demonstrate continuing commercial adoption, particularly if it coincides with higher revenue retention and pricing.
Free cash flow and acquisition spending provide the third. Management has considerable investment capacity, but investors increasingly want disciplined returns rather than an open-ended AI shopping budget.
Thomson Reuters has already answered one question that once worried professional-information investors: generative AI has not destroyed demand for Westlaw, CoCounsel or its tax platforms. The new question is harder because success itself costs money. Shareholders now need evidence that owning the professional AI workflow can eventually be even more profitable than owning the databases that preceded it.
Discover more from Business-News-Today.com
Subscribe to get the latest posts sent to your email.
