LSEG — London Stock Exchange Group plc

Executive summary

LSEG is a global financial markets infrastructure and data business operating across three main franchises: Data & Analytics (c.45% of revenue — Workspace, real-time feeds, pricing/reference), FTSE Russell + Risk Intelligence (c.17% subscription), and Markets (c.38% — Tradeweb, FXall, LCH/SwapClear, LSE equities). Across the period covered, the group has consistently accelerated organic revenue growth (from +4-5% post-Refinitiv in 2021-23, to +7.1% in 2025 and +8.4% in H1 2026) while expanding adjusted EBITDA margin from 47% to 52.7% in the latest half, delivering adjusted EPS CAGR of ~15-17%. The single most important point for valuation today is that LSEG is monetising a genuine AI-receiver thesis — proprietary, licensed data distributed via MCP infrastructure to Anthropic/OpenAI/Microsoft/Databricks/Snowflake — on top of a high-recurring-revenue base trading at a meaningful discount to US data peers.

Fair value estimate

• Fair value range: 9,500p – 10,800p per share, implied market cap £45.7bn – £51.9bn
• Methodology: blended forward P/E and FCF yield, cross-checked against DCF
  • 2025 adjusted EPS 420.6p; H1 2026 adjusted EPS 244.9p (+17.2% YoY) 2026-07-30 interim. Extrapolating full-year gives FY26E adjusted EPS of ~490-500p.
  • 2026 equity FCF guided to at least £2.7bn (≈562p/share on 480.8m shares) 2026-07-30 interim — FCF yield 6.6% at current price.
  • Fair P/E range 19x–22x FY26E adj EPS reflects: double-digit FCF/share CAGR targeted for 2027-2029 2026-02-26 final, 70%+ recurring revenue, and a modest discount to S&P Global/MSCI/Moody's (24-32x) justified by markets/clearing cyclicality.
  • Midpoint fair value: ~10,150p / £48.8bn.
• vs current market cap £41.8bn: upside ~17% at midpoint (range: +12% to +27%).

Sector context

• Sector: Financial Services — specifically Financial Markets Infrastructure & Data. ICB "Financials" classification is technically correct but understates the data/software mix (over 70% recurring subscription-like income).
• Quality/growth/leverage vs peers: Above-average quality (fortress-like franchises in indices, interest-rate swap clearing, pricing data); growth now matches peers (~8%); leverage in line (1.5-2.5x net debt/EBITDA target). Margin trajectory is a positive outlier.
• Listed peers: S&P Global (SPGI), MSCI, Intercontinental Exchange (ICE), Moody's (MCO), Nasdaq (NDAQ), Deutsche Börse (DB1), FactSet.

Investment thesis

1. Material AI-receiver with proof points, not slogans. LSEG's proprietary data (real-time, Reuters news, Lipper, FTSE Russell indices, World-Check, Tradeweb trade data) is being licensed into AI tooling via MCP infrastructure, with 200+ customers engaged and partnerships signed with Anthropic, OpenAI, Microsoft Copilot, Databricks, Snowflake, Google Gemini, Amazon Quick, and Rogo — plus 17,000 active Workspace AI Search users 2026-07-30 interim; 2025-10-23 Q3 update. Unlike buyers of AI, LSEG's dataset gets more valuable as agentic AI scales.
2. Operating leverage visibly compounding. Adjusted EBITDA grew +14.1% constant-currency on +8.4% organic revenue in H1 2026, margin +260bps CC; 2027-2029 medium-term framework explicitly targets +150bps cumulative margin + double-digit FCF/share CAGR 2026-02-26 final. The SwapClear revenue-share restructuring (completed H2 2025) added a further ~100bps structural margin uplift.
3. Available at a fair-to-cheap price vs US data peers. At ~17x FY26E adjusted EPS and ~6.6% FCF yield, LSEG trades at a 20-30% discount to SPGI/MSCI/MCO despite comparable (now matching) organic growth, better recurring mix (73%), and record £3bn buyback programme running through Feb 2027 with £1.4bn still to go 2026-07-30 interim.

Key risks

1. Market-sensitive revenues (~27% of total) could reverse. Tradeweb, FXall, SwapClear and LSE equities benefit from current volatility; a prolonged risk-off collapse in trading volumes would hit the fastest-growing bits. Tradeweb Q1 2026 ADV +31% is unlikely to repeat annually 2026-04-23 Q1 update.
2. AI distribution could cannibalise Workspace seats. If MCP-based data access substitutes for Workspace licences faster than it adds net revenue, the subscription base could decelerate. Management acknowledges "refining commercial strategy" for this channel 2026-04-23 Q1 update — the economic design is not yet fully disclosed.
3. Material amortisation of Refinitiv intangibles obscures reported earnings. Non-underlying amortisation runs ~£500m per half, so IFRS EPS looks very different from adjusted (H1 2026 reported EPS 163.8p vs adjusted 244.9p) 2026-07-30 interim. Investor reliance on adjusted figures is a quality-of-earnings watch item.

Operating leverage

LSEG is a textbook high-operating-leverage platform. The adjusted EBITDA margin has expanded from 47.2% in 2023 to 48.8% in 2024, 50.3% in 2025, and 52.7% in H1 2026 — a cumulative ~550bps over 3 years, delivered on top of accelerating revenue growth 2026-02-26 final; 2026-07-30 interim. The cost base is dominated by fixed R&D, platform engineering, central technology and real estate: adjusted operating expenses grew just +4.6% organically in H1 2026 vs +8.4% revenue growth 2026-07-30 interim. Capex is set to decline from 10% of revenue in 2025 to ~8% by 2029. On a 10-20% revenue upside case driven by AI-channel acceleration, incremental contribution margin is likely 60-70%+ — meaning adjusted EBITDA could realistically grow 20-30% on a 10-15% revenue beat, which would translate into materially outsized EPS and FCF/share compounding. The observable inflection points are already visible: in H1 2026, reported operating profit was +34.6% on +6.9% reported revenue 2026-07-30 interim.

Value-trap signals

None identified. The reverse — this is a quality compounder with multiple consecutive guidance raises, accelerating organic growth, expanding margins, record buybacks, and growing dividend (+17% interim in 2026). The only yellow flags are (a) heavy adjustment items (purchased-intangible amortisation) and (b) the Blackstone/Thomson Reuters overhang now effectively eliminated by 2024 directed buybacks.

Earnings vs expectations

Across the period, LSEG has established a consistent pattern of meeting or beating guidance and raising expectations:

• FY2024: delivered total income growth at the upper end of the 6-8% guidance band set at Nov 2023 CMD 2025-02-27 final.
• FY2025: raised EBITDA margin guidance twice (from +50-100bps to +75-100bps at H1, then to +100bps at top-of-range at Q3), ultimately delivering +150bps reported, +210bps CC 2025-07-31 interim; 2025-10-23 Q3; 2026-02-26 final.
• FY2026: raised revenue growth guidance to 7.0-7.5% (from 6.5-7.5%) and margin to +100bps (from +80-100bps) at H1 2026; Q1 trading update called "upper half" of range before the H1 raise 2026-04-23 Q1; 2026-07-30 interim.

The pattern is unambiguously one of consistent beats and sequential raises.

Conviction

Conviction: 4 (high). Anchors: (1) exceptionally detailed, line-by-line segmental disclosure with consistent constant-currency and organic cuts; (2) multiple valuation approaches (forward P/E, FCF yield, peer-relative) converge in a tight range around £46-52bn fair value; (3) explicit medium-term framework from management with a track record of meeting it. Caveats: (1) sensitivity of adjusted EPS to the large purchased-intangible amortisation adjustment means IFRS EPS tells a different story; (2) peer multiples themselves are elevated — if US data peers derate, LSEG's "discount to peers" narrative weakens.

Driver scoring

• ai_beneficiary (75): Proprietary data business with demonstrable AI monetisation — 200+ MCP customers, major foundational-model partnerships, Workspace AI Search at 17k users, LSEG DaaS flowing data into Databricks/Snowflake. Not just press-release AI; genuine addressable-market expansion.
• operating_leverage (75): 70%+ recurring revenue, 52.7% EBITDA margin and climbing, EBITDA growth consistently outpacing revenue by ~2x. Fixed R&D and central cost structure. Classic specialist-software/data operating leverage profile.
• earnings_surprise_trend (75): Multiple sequential guidance raises in 2025 and 2026; H1 2026 delivered 8.4% organic growth vs original full-year guidance of 6.5-7.5%.
• cyclicality (30): ~27% of revenue is market-activity-sensitive (Tradeweb, FX, equities, OTC clearing), but these businesses are counter-cyclical to volatility, and 73% is highly stable subscription. Net cyclicality is low.
• moat (85): Network effects in clearing (SwapClear ~90% share of global IRS clearing), proprietary data stacks (FTSE Russell indices, Reuters news, Pricing & Reference), regulatory infrastructure, 98% customer retention. One of the wider moats in financials.
• leverage (45): Net debt/EBITDA around 1.9-2.0x — mid-range of 1.5-2.5x target, well-supported by investment-grade ratings (A/A3), ample liquidity, strong FCF.
• earnings_quality (70): Cash conversion very strong (£2.4bn equity FCF 2025, guided ≥£2.7bn 2026); adjusted metrics are the appropriate lens but reconcile to robust underlying cash. Deduction for the large purchased-intangible amortisation line that obscures IFRS P&L.
• management_quality (78): Refinitiv integration delivered cost synergies ahead of schedule, revenue synergies on track, Microsoft partnership secured on favourable terms, Blackstone overhang resolved via directed buybacks at attractive terms, consistent dividend growth, disciplined M&A (Acadia, Quantile, Tora, ICD).
• growth_momentum (74): Organic growth has accelerated each year — 7.1% FY25, 9.8% in Q1 2026, 8.4% in H1 2026 — with the 2027-2029 framework targeting mid-to-high-single-digit annually. Clear positive momentum.

Overall score: 720

LSEG is a strong fit for all three pillars of the investor profile. It is a genuine AI-receiver through its proprietary-data moat and MCP distribution infrastructure (not a hyped buyer of AI); operating leverage is high and visibly compounding; and valuation at ~17x FY26E adjusted EPS / 6.6% FCF yield is fair-to-cheap relative to US data peers (SPGI/MSCI/MCO at 24-32x). Downside protection is strong (fortress balance sheet, systemically important clearing franchise, 70%+ recurring revenue). The reason it doesn't score higher is that (a) the AI monetisation economics are still early and management is "refining commercial strategy" for the MCP channel — the explicit AI-driven revenue line is not yet quantified, and (b) a quarter of revenue is market-volume-sensitive, so it's not a pure "picks and shovels" play on AI like a semiconductor name.