Morning Call & AI12 Feb 2026 13:51
Subjective, as always
I thought the CEO did a good job to dispel AI fears, Hoping that the understanding will start to sink in to market generally, as some market participants clearly dont understand how AI works with enerprise environment
My take on things...
We are a "Picks and Shovels" play, in that we provide essential infrastructure for important professions. OpenAI and Anthropic may provide the reasoning capability but we provide the verifiable truth In legal, medical, and risk sectors. AI's "opinion" is pretty worthless without a citation (its a trust thing).
Rel has successfully moved from publisher to Vertical AI Software leader. They aint just selling data, they are selling "AI-ready" data (M&S..:-) that fuels the services stack.
Investors should view RELX’s Data Layer as the "go to point" for AI, which is technology agnostic. Whether a business uses GPT-4o, Claude 4 or any other model, the output is only as good as the input. Because our data is clean, structured, copyright-protected etc ensures that any model plugged into it performs at a professional level, whereas models relying on the open internet are prone to "hallucinations" and legal inaccuracies.
Also from a rev perspective, because Rel owns the data library, every AI query processed through their library is at virtual zero cost. Leading to the significant margin expansion we are now seeing (34.8%)
The market is still panicked that "General AI" will replace everything. Well it cannot replace propriatary data, it needs it. So to flip this around Erik was quiet specific, it does not matter what model layer (LLM) Rel plug into. Its the Orchestration (worksteam curration), data returns and application interface that the customer uses, that are of importance and where the customer gets real value
In addition we have Professional accountability preventing lawyers and doctors from using unverified AI.
In the context of Harvey , Rels Protégé platform proves that the "Model Layer" needs the "Data Layer" to survive within the enterprise. In this commercial model Rel collects rent from Harvey. My understanding is Harvey is a 1 off and we own a good % + a seat on their BoDs
So in Summary - Rel owns the "Primary Source" material that AI companies cannot legally scrape or replicate, Rels Orchestration helps govern this
Workflows are locked in by embedding AI into Lexis+ etc, Rel owns the user’s daily "work interface," making services "sticky."
M2M Revenue is c90% of Risk revenue and machine-led ie Rels income is almost entirely generated by automated "API calls" rather than subscriptions
Enough whaffle ......
Re the market reaction looks like we have been re priced for the time being. May take more time and evidence to prove we are an AI winner, based on the speed at which things are moving 12 months should provide clear evidence. That said todays results / guidance along with a robust Q&A should help start the realisation process for Rel