RECompany report
RELX Plc RELX
The bet you're really making is that RELX keeps selling its proprietary data and analytics tools to the same lawyers, scientists, bankers and insurers every year, and that those customers keep paying more as RELX builds AI into the products. You're betting AI helps RELX instead of killing it: that a lawyer still needs LexisNexis to be sure an answer is right, and an insurer still needs RELX's data to price a policy. Right now it is going well: first-half sales hit a record, up 7-8% before currency, with legal growing fastest in years, even as the market keeps worrying AI will hollow the moat. You pay 21 times earnings, the least the stock has cost in twelve years, down from 35 times at its decade high.
Key data
RELX · price with moving averages
Source: market data.
The business
RELX sells information and analytics tools to four groups: risk and fraud teams at banks and insurers through LexisNexis Risk Solutions, scientists and doctors through Elsevier and ScienceDirect, lawyers through LexisNexis Legal, and trade-show attendees through RX. The product a customer holds is a subscription seat into a proprietary database plus the decision tools wrapped around it. An underwriter runs an identity through RELX's data to price a policy in seconds. A researcher searches millions of peer-reviewed articles. A lawyer checks whether a precedent still stands and needs to be right, not roughly right. Revenue is predominantly electronic and recurring, renewed annually. The moat is the data itself: contributory and proprietary datasets rivals cannot rebuild, plumbed into the customer's daily workflow so switching means retraining people and re-wiring systems. Risk is now the biggest division at £1.8B of first-half sales, 37% of the group, built on identity, fraud and financial-crime-compliance data that gets more valuable exactly as fraud gets cheaper to commit. Exhibitions, 12% of sales, is the one physical, cyclical piece. The other three segments are subscription software sitting on a data spine.
The numbers
Five years of filed accounts show a compounder, not a cyclical.
| Fiscal year | Revenue | Net income | Diluted EPS |
|---|---|---|---|
| 2021 | £7.24B | £1.47B | £0.76 |
| 2022 | £8.55B | £1.63B | £0.85 |
| 2023 | £9.16B | £1.78B | £0.94 |
| 2024 | £9.43B | £1.93B | £1.03 |
| 2025 | £9.59B | £2.06B | £1.12 |
| 2026, 1H to Jun | £4.87B | £1.17B | £0.65 |
Revenue compounded 7.3% a year from 2021 to 2025, net income 9.0%, and operating income 12.6% as the mix tilted toward higher-margin analytics. Operating margin widened from 26.0% to 31.6%. Every £1 of new revenue over that span dropped 49p to operating profit, the operating leverage the whole case rests on.
| Period | Revenue | Net income | Diluted EPS |
|---|---|---|---|
| 2024, H1 | £4.64B | £985M | £0.52 |
| 2024, H2 | £4.79B | £949M | £0.51 |
| 2025, H1 | £4.74B | £976M | £0.53 |
| 2025, H2 | £4.85B | £1.09B | £0.59 |
| 2026, 1H to Jun | £4.87B | £1.17B | £0.65 |
The first half of 2026 set a record, with adjusted earnings of $0.91 an ADR, up 6% on the prior first half, even as a strong pound held reported revenue growth to 3%. Underlying growth ran 7-8%.
| Segment | H1 2026 rev, £B | Underlying growth |
|---|---|---|
| STM | 1.37 | +6% |
| Legal | 0.96 | +10% |
| Exhibitions | 0.58 | +6% |
| Group | 4.87 | +7% |
The divisional picture is where the story turned. Legal underlying growth stepped up to 10%, its fastest in years, STM to 6%, and Risk kept compounding in the high single digits. Only Exhibitions is soft. What the market does not believe is priced right here: at 21 times earnings, the lowest multiple in twelve years, the stock reads as if AI erodes these franchises, while the numbers say AI is lifting growth. The print that settles it is the underlying rate in STM and Legal at the February full-year.
Management
Erik Engstrom has run RELX for over a decade, and the record is in the capital allocation, not the biography. No insider bought or sold a share in the open market over the past year, so the signal is the buyback. RELX will spend £2.25B on repurchases in 2026, about 3.6% of the company, executed daily and without discretion around 2,300 to 2,850 pence through late August, right around today's price and at the cheapest multiple in over a decade. Share count fell from 1.77 billion in late May to 1.75 billion by late August, roughly a 4% annual reduction on top of a 20.9p interim dividend. Bolt-on acquisitions ran £262m in the half against £11m of disposals. This is a team that buys small, sells the fading print assets, and returns the rest. Pay is tied to adjusted revenue, profit and constant-currency EPS growth, which is why the reported-versus-underlying gap matters to how they are scored. The discipline that counts is buying back hardest when the multiple is lowest, which is now.
How it fails or surprises you
The AI substitution risk. If large language models let a lawyer or researcher get a trustworthy answer without RELX's database, renewals stall and pricing power fades. This is the fact the numbers explain least well: the stock de-rated from 35 to 21 times earnings over two years while earnings rose, so the market is paying for erosion the income statement does not yet show. It would appear first as STM and Legal underlying growth slipping from 6-10% toward low single digits within a year.
The strong-pound drag. RELX earns most of its money in dollars but reports in sterling, and a pound that stays strong turns 7-8% underlying growth into 3% reported and dents the dividend and EPS the ADR pays. It is not fundamental, but it caps the headline numbers the market anchors to, and another year of 3% reported prints could keep the multiple pinned while the business compounds underneath.
Risk becomes the whole story (right tail). LexisNexis Risk, £1.8B and 37% of sales, sells fraud, identity and compliance data that gets scarcer and more valuable as AI makes fraud cheaper. If that division re-accelerates into double digits and the market re-rates it as a data-analytics compounder rather than a publisher, RELX earns faster growth and a higher multiple at once. The tell is Risk underlying growth breaking above 10% at the full year.
Closing thoughts
The underlying growth rates in STM and Legal at the February full-year will settle this: hold at 6-10% and the moat survived AI, fade to low single digits and the de-rating was correct. The market decided AI is a threat to information services and took RELX from 35 times earnings to 21, the cheapest in twelve years, while the business grew straight through it. The other side of that decision owns the view that models commoditize the data. Your read is that liability-grade answers still need RELX's proprietary spine, and the first-half divisional step-up is early evidence you are right. The fatter tail is up. The downside is a slow fade in a recurring-revenue base that still throws off cash and shrinks its own share count 4% a year, so what is at risk is years of a flat multiple, not permanent loss. The upside is growth reacceleration and a re-rating arriving together.
The bet is still that lawyers, scientists, bankers and insurers keep paying RELX every year and pay more as AI gets built into the tools, not around them. What breaks it is one pair of numbers: STM and Legal underlying growth. Hold at 6% and 10% and the moat is intact and the price is a gift. Drift toward low single digits and the bears were slow but correct. Watch those two lines in February, not the group multiple.
Methodology
RELX reports semiannually, not quarterly, so the near-term table shows five consecutive half-year periods; H2 figures for 2024 and 2025 derived as fiscal year less H1 from annual and interim results.
Data: FY2021-FY2025 revenue, operating income, net income and cash flow from RELX filed accounts (XBRL); H1 2026 revenue, segment and underlying-growth figures from the H1 2026 interim results 6-K (reported Jul 23, 2026); buyback and treasury figures from daily repurchase 6-Ks through Sep 1, 2026.
Adjusted EPS is stated per ADR in USD, group and segment revenue in GBP as filed.
Risk segment revenue (£1.8B, 37% of group) derived by subtraction: H1 2026 total revenue £4.871B minus disclosed segments STM £1.370B, Legal £0.959B, Exhibitions £0.575B, and Print £0.157B equals £1.810B.
Valuation: trailing P/E 20.8x and EV/EBITDA 13.5x from live quote and vendor ratios; forward FY2027 P/E 17.0x from consensus GBP EPS of £1.56 (8 revenue, 2 EPS estimates), converted at GBP/USD 1.34; twelve-year P/E range 21x-35x, current 21x is the low.
Insiders: no open-market buys or sells in the past twelve months; plan status not applicable. Capital return is the signal, £2.25B of buybacks planned for 2026.
Judgment (which tail is fatter, the AI read, the resolving print) is mine, not sourced fact. Figures current to the Sep 1, 2026 filing; GBP/USD moves affect all reported and multiple work.
Fact check: All financial metrics cross-checked against RELX 6-K filings (Jul 23, 2026 interim results, daily repurchases through Sep 1, 2026) and FMP vendor data; segment percentages mathematically derived from filed revenue totals; CAGRs and margins recalculated from source figures; repurchase price range approximated from daily filings (actual range 2,260p-2,885p, typical midpoint stated). Final analysis verified as of Sep 6, 2026.
Bid Cap
Daily ideas, a 390-name database, and a model long/short book from an investor who mostly covers financials. $70 a month or $700 a year.
Subscribe on Substack


