Is LexisNexis Losing Ground to Harvey? RELX's Legal Growth Says Not Yet

Harvey went from 42% of the AmLaw 100 to 80% in about 13 months, and over the same stretch RELX's Legal division sped up rather than slowed down. Underlying revenue growth at LexisNexis Legal & Professional rose to 10% in the first half of 2026. If Harvey is the LexisNexis competitor that investors fear, it has not yet shown up in the incumbent's filings.

The LexisNexis competitor question has two ledgers

Every argument about legal AI eating the incumbents rests on a claim about money moving. Money that used to go to a research subscription now goes to an AI platform instead. That claim can be checked, at least partly, because one side of the trade publishes audited segment results every six months.

RELX is a listed company. It reports its Legal segment, which trades as LexisNexis Legal & Professional, with revenue, operating profit and an underlying growth rate. Harvey is private. It discloses what it chooses to, usually alongside a funding round.

So the two ledgers are not symmetrical. One is audited and backward-looking. The other is curated and forward-looking. Reading them side by side is still the cleanest test available of whether vertical AI is taking incumbent revenue or adding new spend on top.

This post extends the earlier look at Harvey against Thomson Reuters to the other legal research incumbent. The question is narrower here. It is not whether Harvey is a good product. It is whether RELX's numbers show any damage yet.

Direct answer

No, LexisNexis is not losing revenue to Harvey on any figure RELX has published. RELX Legal's underlying growth rose from 7% in 2024 to 10% in the first half of 2026, while Harvey reached 80% of the AmLaw 100 and a $15.5 billion valuation. So far the evidence points to additive spending, with the caveat that multi-year contracts delay any visible displacement.

What RELX Legal results actually reported

RELX's interim results for the six months to 30 June 2026 put Legal revenue at £959 million, up from £900 million a year earlier. Reported growth was 7%. At constant currency it was 9%. On RELX's underlying measure, which strips out currency, acquisitions, disposals and print, it was 10%.

Adjusted operating profit rose from £188 million to £208 million. The adjusted operating margin moved from 20.9% to 21.7%. Underlying profit growth was 13%, which means costs grew more slowly than revenue.

RELX described this as "a further step up in growth from full year 2025". That phrase matters, because the full-year figure it refers to was itself an improvement. RELX reported Legal underlying growth of 7% for 2024 and 9% for 2025, according to its 2025 highlights and its 2024 annual results filing.

The sentence that names the driver

The interim filing attributes double-digit growth in Law Firms & Corporate Legal directly to AI products. In its words, that growth "is being driven by the continued adoption of our core AI-enabled legal research and analytics platform with its integrated agentic legal assistant (Lexis+ with Protégé)".

That is the segment where Harvey competes hardest. Large law firms and in-house teams are Harvey's stated customer base. RELX is telling investors that its fastest growth sits in exactly the same buyers.

How big Legal is inside RELX

Legal is a minority of the group. Using the same interim table, Legal was about 19.7% of RELX's £4,871 million first-half revenue. It was about 12.0% of group adjusted operating profit of £1,727 million, because its margin sits well below the group's 35.5%.

Those two shares are my arithmetic from RELX's own figures, not a number RELX publishes. They explain one thing about the market reaction to legal AI. A full collapse in Legal would hurt RELX, but it would not break it.

Source: RELX 2026 interim results, Form 6-K, 23 July 2026. The share-of-group row is calculated from the same filing and is not a RELX disclosure.

What the Harvey AI valuation disclosures contain

Harvey announced a $550 million round at a $15.5 billion valuation on 9 September 2026. Diffusion and Lightspeed Venture Partners co-led it. Sequoia, Kleiner Perkins, a16z, Coatue and GIC were among the existing investors that followed on.

The valuation path is steep. TechCrunch reported that the round followed an $11 billion valuation in March 2026 and an $8 billion valuation in December 2025. Legal Business put the mid-2025 valuation at $5 billion. That is roughly a tripling in about 15 months.

What the round announcement left out

Harvey's own post gives investor names and one adoption claim: 80% of AmLaw 100 firms use Harvey, along with five of the Fortune 10. It gives no total annual recurring revenue figure. ARR, short for annual recurring revenue, is the annualised value of current subscription contracts.

The ARR totals in circulation came from trade reporting before and around the round, not from Harvey's announcement. Those reports put Harvey's ARR at around $300 million ahead of the raise, with later coverage citing figures above $400 million. I could not trace either to a document Harvey published, so I treat both as directional.

What Harvey did say about revenue

Harvey has published two revenue facts of its own. In August 2025 it told Artificial Lawyer it had reached $100 million of ARR and over 500 customers in 54 countries, including 42% of the AmLaw 100. In August 2026, announcing a new chief revenue officer, it said Q2 was a quarter in which it added more than $100 million in ARR.

Put those together and Harvey added in one quarter of 2026 roughly what it had built in its first three years. That is a fast-growing business by any standard. Legal Business, covering the September round, reported Harvey working with over 200,000 lawyers and nearly 2,500 organisations.

The valuation multiple depends on which ARR you believe. At $300 million of ARR, $15.5 billion is about 52 times. At $400 million, it is about 39 times. Neither is a disclosed figure, which is why the range matters more than the point. The broader pattern of private AI multiples is covered in the analysis of vertical AI against horizontal SaaS.

The legal AI market displacement test, run on the filings

A displacement thesis makes a specific prediction. If Harvey were winning budget that would otherwise go to LexisNexis, RELX Legal growth should slow as Harvey scales. The slowdown should be sharpest in law firms and corporate legal, where both sell.

The filings show the reverse. Harvey went from $100 million of ARR to adding more than that in a single quarter. Over the same window, RELX Legal underlying growth went from 7% to 9% to 10%. The sub-segment RELX names as growing double digits is the one where Harvey has its densest adoption.

Why both can grow at once

The simplest explanation is that the budgets differ. A LexisNexis subscription buys primary law, citators and secondary sources. A Harvey seat mostly buys drafting, review and document work that a firm previously paid for in associate hours.

In that reading, Harvey is competing with the billable hour more than with the research database. RELX gets paid for authoritative content, and Harvey gets paid for labour substitution. Both lines can grow while the firm's total technology spend rises.

What the absolute scale says

Scale makes the same point. Doubling RELX's first-half Legal revenue gives a crude annual figure of about £1.9 billion. Even the higher press estimate of Harvey's ARR would be a fraction of that. Harvey would need to take share for years before the effect showed in a segment of this size, unless its wins came disproportionately from LexisNexis renewals.

The final row is why this table is a test and not a proof. Every disclosed signal points one way. The undisclosed ones could still point the other.

LexisNexis is Harvey's competitor, supplier and shareholder

The cleanest frame for this rivalry is wrong. RELX does not sit across the table from Harvey. It sits on three sides of it.

The 2025 alliance

In June 2025, LexisNexis and Harvey announced a strategic alliance. Harvey users can query LexisNexis Protégé inside Harvey and receive answers grounded in LexisNexis US case law and statutes, checked against Shepard's Citations. The two companies also agreed to co-develop workflows, starting with motions to dismiss and motions for summary judgment.

That arrangement turns part of Harvey's growth into distribution for LexisNexis content. When a firm buys Harvey and wants verified US case law inside it, the content still comes from RELX. The commercial terms have not been disclosed, so the size of that flow is unknown.

The investment link

Coverage of the alliance by Legal IT Insider noted that Rev Ventures, a venture fund backed by RELX, took part in Harvey's $300 million raise earlier in 2025. RELX does not report the size of that stake. It does mean Harvey's rising valuation is not purely a loss for RELX shareholders.

Why the three roles matter for the test

A company that supplies, invests in and competes with a challenger has hedged the outcome. If Harvey wins the interface, RELX still sells the content underneath. If Harvey stalls, RELX keeps the customer directly. I think this hedge is the most underrated fact in the whole LexisNexis competitor debate.

The same pattern, an incumbent owning the data a challenger needs, is the core of the argument about data moats against workflow moats. Legal research is one of the few markets where the data moat looks real, because citator coverage and editorial enhancement cannot be scraped into existence.

LexisNexis Protégé is the incumbent's real answer

RELX did not wait to be displaced. Lexis+ AI rolled out through 2024, and Protégé, described by RELX as its agentic legal assistant, followed. By the 2026 interim results the company was describing the combined product as "Lexis+ with Protégé" and adding a layer called Protégé Work.

The product strategy is visible in the filing language. RELX stresses "comprehensive, verified legal content" and "trusted legal AI". It is not selling a better chatbot. It is selling an answer grounded in a corpus that general model providers cannot license.

Price, not seats, is probably doing the work

Here is my read, and it is an inference rather than a disclosed fact. Much of the growth acceleration likely comes from moving existing subscribers onto AI tiers at higher prices. RELX's language about a "shift in business mix towards higher growth, higher value legal analytics and tools" fits that reading.

If that is right, the incumbent's AI strategy is a repricing strategy. It works for as long as customers believe the AI tier is worth more than the base subscription. Harvey's growth makes that belief easier to sustain, not harder, because it normalises paying for legal AI at all.

The quality caveat both sides share

Grounded content does not mean error-free output. A Stanford RegLab study found that purpose-built legal research tools, including LexisNexis's, still produced hallucinated answers on a meaningful share of queries. That evidence is covered in the comparison of Harvey and Legora against general AI. Neither incumbent nor challenger has a clean accuracy story.

Where this argument is weakest

The additive reading fits every disclosed number. It could still be wrong, and these are the four ways it fails.

Contracts hide displacement for years

Large law firm research contracts commonly run for several years. A firm that decided in 2025 to rely on Harvey for most drafting work might not cut its LexisNexis spend until renewal. RELX's reported growth today reflects decisions made at earlier renewals. This is the strongest objection, and the data available cannot rule it out.

Underlying growth is a curated measure

RELX's underlying growth excludes print revenue, among other things. Print is the declining part of the Legal business. Underlying growth of 10% sat next to reported growth of 7%, partly on currency. The headline figure is honest about its method, and it is still the most flattering of the three rates published.

Price can mask volume loss

If RELX is raising prices on fewer seats, revenue can grow while usage shrinks. RELX does not publish seat counts or active user numbers for Lexis+ AI. Without them, you cannot separate a healthy franchise from one harvesting its installed base before a decline.

The market has already voted the other way

RELX shares fell sharply in early February 2026, alongside Thomson Reuters and Wolters Kluwer, after Anthropic released a legal plug-in for Claude. That episode is examined in the breakdown of the SaaSpocalypse sell-off. Investors were pricing a threat the filings do not yet show. I think the market was early rather than wrong, and the general-model threat it priced is different from the Harvey threat this post tests.

What would change the verdict on LexisNexis

A verdict of "additive so far" is only useful if you name what would overturn it. These are the readings I would treat as evidence of real displacement.

The alliance is the leading indicator

The third row is the one I would watch most closely. As long as Harvey routes verified case law through LexisNexis, the two companies are partners in the largest firms. If Harvey ever builds or buys its own primary law corpus with citator coverage, the relationship becomes a straight fight. That move would tell you more than any growth rate.

Founders watching this market should note the lesson in it. The incumbent advantage in AI is strongest where the incumbent owns content the challenger cannot replicate. Legal research is close to the textbook case.

Frequently asked questions

How to use this before your next legal tech decision

If you run legal operations or a legal tech company, there is one practical point. Research content and AI workflow are being bought as separate things. Treat them that way in your own planning.

This week, pull your current research subscription and your AI tool contracts into one sheet. Write down what each one is actually used for, and which renewal comes first. If the AI tool already does work you were paying the research vendor to support, you have a negotiating position at that renewal. If it does not, you are running two budgets, and the incumbent's growth figures suggest most firms are doing exactly that.

If you are building in this market, stop pitching "replace Lexis". The filings say buyers are not replacing it. Pitch the work a research database never did. Pricing that pitch is a separate problem, explored in the review of Harvey's pricing and customers.

Related reading