S&P Global Inc. (NYSE: SPGI) — The Toll Collector on Capital, On Sale for Fear of a Robot

An independent equity research note Date: June 11, 2026 Price at analysis: ~$426.38 (52-wk range $381.61–$579.05; ~26% off the high, ~12% off the low) · Market cap: ~$124–126B · Net debt: ~$11.3B · EV: ~$135B Sector: Financials — Capital Markets / Financial Exchanges & Data (GICS Financials / Capital Markets / Financial Exchanges & Data) Fiscal year-end: December 31 · CIK: 0000064040 · Founded: 1860 · Index membership: S&P 500 (it both calculates and is a member)

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice, and not a recommendation to buy or sell any security. The analysis that follows takes no position and sets no price target; the only opinion expressed in this article is in this clearly-labeled block.

Verdict: BUY / accumulate here, add aggressively on further weakness — a portfolio of some of the best moats in public markets, de-rated to the cheapest third of its own decade on a fear aimed at the wrong segment. Conviction: MEDIUM-HIGH. Directional value zone: I find ~$420–470 an attractive entry; base-case fair value ~$530–590 over 12–24 months (~27–28× forward adjusted EPS, roughly the street’s ~$533 and the low end of SPGI’s historical multiple), with a bull case toward ~$650 if AI data-licensing becomes a visible revenue line; a real downside cushion only breaks toward ~$360–390 if the ratings cycle rolls over and the multiple stays compressed.

S&P Global is not a “financial data company” — it is a tollbooth on the two largest flows in finance: the issuance of debt and the indexing of equity. Roughly 65% of segment operating profit comes from Ratings (64% margin) and S&P Dow Jones Indices (69% margin) — a regulatory-and-reputation ratings duopoly and a brand-and-network index oligopoly, both with ~50–70% margins sustained for decades, the textbook financial signature of a genuine moat. The market is treating SPGI like FactSet (down ~43% from its high) and Verisk — a data vendor whose terminal an LLM might hollow out. That fear is mis-located: it lands hardest on Market Intelligence, which is ~37% of revenue but only ~16% of profit and the one segment with a contestable 20% GAAP margin. You cannot prompt your way around a credit rating the bond market has collectively agreed to require, or around the S&P 500 that trillions of dollars contractually track. The crown jewels are nearly AI-immune; the de-rating priced them as if they weren’t. On top of that, the near-term tape is noisy in a way that creates the entry: an Iran energy shock denting the Energy segment, ratings issuance set to “turn negative in Q4” on brutal 2025 comps, and private-credit jitters — all cyclical, all already in guidance.

The framing is quality-compounder-at-a-discount, the mirror image of MSCI (an equally elite franchise that never de-rated and trades ~50% richer). SPGI sits at the ~28th percentile of its own ten-year valuation range, ~21.8× forward adjusted EPS versus a historical ~28–30×, while management compounds adjusted EPS at a guided +9–10% for 2026 (a trough year), targets 7–9% organic revenue growth and double-digit EPS growth over 3–5 years, returns 100%+ of free cash flow (≈$4.5B of buybacks into the weakness), has raised its dividend for 53 consecutive years (a Dividend King), and is spinning off the lower-multiple Mobility/CARFAX business to sharpen the franchise. Conviction triggers: turns more bullish (high) if AI data-licensing/MCP monetization shows up as a measurable ACV uplift or if private-credit ratings demand inflects up; turns bearish if S&P Ratings loses NRSRO-grade share to KBRA/DBRS in private credit at scale, or if MI organic growth slips below ~5% with retention cracking (the AI bear thesis made real). Catchy version: “They charge a toll every time the world borrows or indexes a dollar — and it just went on clearance because a chatbot scared the wrong segment.”

1. Executive Summary

S&P Global is the infrastructure layer of global capital markets — it rates the debt, calculates the benchmarks, prices the commodities, and sells the data that the financial system runs on. In FY2025 it generated $15.34B of revenue (+8% reported, ~7% organic), $6.48B of GAAP operating profit (42% margin), $4.47B of GAAP net income, and ~$5.1B of adjusted free cash flow, with adjusted diluted EPS of ~$17.80 (+14% YoY). The business is a collection of five franchises of sharply unequal quality: S&P Global Ratings (~$4.72B revenue, 64% GAAP margin), Market Intelligence (~$4.92B, 20% margin), S&P Dow Jones Indices (~$1.85B, 69% margin), S&P Global Energy (formerly Commodity Insights/Platts; ~$2.30B, 41% margin), and Mobility (~$1.75B, 22% margin — being spun off as a separate public company, completion targeted ~July 2026).

The investment question is not whether SPGI is a great business — it plainly is — but why a great business is trading at the cheap end of its own history, and whether that is a gift or a warning. SPGI sits at roughly the 28th percentile of its ten-year valuation range (~21.8× forward adjusted EPS, ~24× trailing), ~26% below its 52-week high, having fallen with the broader “GenAI will hollow out financial-data vendors” de-rating that took FactSet to the 3rd percentile of its own decade and Verisk down ~44% . The bear case bundles three fears: (1) generative AI commoditizes the data/terminal layer; (2) private credit disintermediates the public-bond ratings annuity; (3) a cyclical air-pocket in 2026 (energy shock, ratings comps, soft Energy guidance).

This memo’s analysis is that the de-rating is disproportionate to the franchise actually at risk. The profit pool is overwhelmingly concentrated in the two segments least exposed to AI substitution: Ratings and Indices together are ~65% of segment operating profit and are protected not by computation (which AI cheapens) but by coordination standards — a regulatory NRSRO license plus a century of reputation in Ratings, and the embedded S&P 500/Dow benchmark in Indices — that no model can replicate. The AI fear is most valid for Market Intelligence, which is the largest revenue segment but the smallest profit contributor (~16%) and the only one with genuinely contestable economics. Meanwhile management is leaning into the weakness with the strongest capital-return posture in years (100%+ of FCF returned, ~$4.5B buyback), a portfolio-sharpening spin-off, and a 53-year dividend-increase streak.

The honest risks are real and specified later in this article: ratings revenue is cyclical and faces a hard Q4-2026 comp; private credit is a genuine structural swing factor; Market Intelligence must prove it can monetize AI rather than be commoditized by it; and the IHS Markit deal left a goodwill-laden balance sheet that makes reported ROE/P/B uninformative. This memo takes no recommendation and sets no price target. Its conclusion is that SPGI is a wide-moat, capital-light, cash-generative compounder whose multiple has compressed to a level that discounts a structural threat to the parts of the business least subject to it.

2. Business Overview

S&P Global makes money by selling trust, standards, and proprietary data into the plumbing of capital markets. It reports five segments; the company itself frames >95% of revenue as tied to “proprietary benchmarks, differentiated data and critical workflow tools,” with under 5% from “undifferentiated” public data.

S&P Global Ratings (~$4.72B FY2025 revenue, ~31% of total, 64% GAAP operating margin). The crown jewel. S&P assigns credit ratings to bond issues and issuers, earning a transaction fee when a new bond is rated (≈52% of Ratings revenue in FY2025) and a non-transaction/recurring fee for surveillance, annual relationship fees, and the CRISIL majority-owned Indian franchise (≈48%). The issuer pays. The model is exquisitely capital-light: the marginal cost of rating one more bond is an analyst’s time, while the value (market access, lower borrowing cost via a recognized rating) is enormous. Transaction revenue is cyclical — it tracks debt issuance volumes — but the non-transaction half is a sticky annuity.

Market Intelligence (~$4.92B, ~32% of total, 20% GAAP margin). The largest segment by revenue, the weakest by margin. MI sells multi-asset data, analytics, and workflow software: the Capital IQ Pro desktop, Compustat and SNL data, Visible Alpha (sell-side estimate granularity), Ratings distribution (RatingsDirect/RatingsXpress), and Enterprise Solutions (mission-critical workflow software like Wall Street Office and ClearPar). It is overwhelmingly subscription (~85% of MI). This is the segment in the AI crosshairs — but management’s reframing is that the value is not the raw data (12% of MI revenue is “undifferentiated”) but the proprietary content, business logic, and embedded workflows.

S&P Dow Jones Indices (~$1.85B, ~12% of total, 69% GAAP margin). Economically the best business SPGI owns. It owns and licenses the S&P 500, the Dow Jones Industrial Average, and thousands of other benchmarks. Revenue is (i) asset-linked fees — a royalty (a few bps) on the AUM of ETFs and funds tracking S&P indices; (ii) exchange-traded derivatives royalties on SPX/VIX options (CBOE) and E-mini futures (CME); and (iii) data subscriptions. Important structural caveat: S&P DJI is a joint venture in which CME Group owns 27%, so a quarter of the segment’s economics leaks out as non-controlling interest (~$322M deducted in FY2025).

S&P Global Energy (formerly Commodity Insights / Platts; ~$2.30B, ~15% of total, 41% GAAP margin). Price assessments and benchmarks for energy and commodities — most importantly Dated Brent, the reference for roughly two-thirds of globally traded waterborne crude — plus the CERAWeek conference, energy data/insights, and Global Trading Services. A benchmark monopoly facing a slow, multi-decade energy-transition question rather than a competitive one.

Mobility (~$1.75B, ~11% of total, 22% GAAP margin) — being spun off. CARFAX (vehicle-history reports, ~65% of Mobility), automotiveMastermind, and Polk/automotive data. On April 29, 2025 SPGI announced it would separate Mobility into an independent public company (“Mobility Global”); the spin is targeted for mid-2026 (≈July 1, distribution to SPGI holders), with ~$2B of debt issued at Mobility and the cash dividended back to SPGI for buybacks and debt reduction. Post-spin, SPGI becomes a cleaner four-segment ~$13B-revenue capital-markets-information company.

Revenue model summary. SPGI is ~75%+ recurring/subscription at the enterprise level, with the cyclical kicker concentrated in Ratings transaction revenue and Indices asset-linked/derivatives fees. It is asset-light (capex ~$195M FY2025, ~1.3% of revenue), converts a high share of operating profit to cash, and reinvests little physical capital — the hallmark of a royalty/toll business.

Verdict: A portfolio of mostly toll-road economics with one contested data franchise. The mental model is not “diversified data conglomerate” but “two world-class annuities (Ratings, Indices) + a benchmark monopoly (Energy/Platts) + a contested data-and-software business (MI) + a soon-to-depart auto-data unit (Mobility).” The quality of the whole is carried by the parts that throw off 60–70% margins.

3. Industry Dynamics

SPGI spans four distinct industry structures (five pre-spin). They are not equally good, and the analysis must weight them by profit, not revenue.

Credit ratings — a near-impregnable regulatory oligopoly

The ratings industry is a textbook duopoly-plus-one. By outstanding ratings, S&P (~48.9% US share) and Moody’s (~34.2%) together hold ~80%+, with Fitch (~13–15%) a distant third and DBRS Morningstar, Kroll (KBRA), and AM Best smaller still (SEC NRSRO data, compiled via public sources, 2025). Three reinforcing barriers make this one of the most defensible structures in all of finance:

1. Regulatory license (NRSRO). SEC registration as a Nationally Recognized Statistical Rating Organization is required for ratings to count toward regulatory capital, mandate eligibility, and index inclusion. Counterintuitively, post-2008 regulation (Dodd-Frank’s ~517 pages of Rule 17g requirements) raised barriers to entry — an SEC commissioner called the rules potentially “life-threatening to smaller agencies.” Regulation polices the incumbents but moats them against entrants.
2. The two-rating convention. Most large bond issues carry two ratings, and S&P + Moody’s are the default pair. A new entrant must displace an incumbent on a syndicate desk’s checklist, not merely be competent — a coordination barrier, not a quality one.
3. Reputation as accumulated intangible capital. A century-plus track record is the product; investors and regulators mandate the brands by name.

The financial proof: S&P Ratings has run a 56–64% GAAP operating margin (FY2023–25); Moody’s Investors Service runs ~50%+ adjusted. Two firms earning ~50–65% margins for decades, through cycles, is precisely what a durable oligopoly looks like. The genuine structural debates are issuer-pays conflicts of interest (live since 2008, never resolved, never dislodged) and private credit — not competitive displacement.

Indices — the asset-linked tollbooth on passive investing

The institutional equity-index market is a three-firm oligopoly with largely non-overlapping franchises: S&P Dow Jones Indices owns US large-cap (the S&P 500), MSCI owns global/international/emerging-markets, and FTSE Russell (LSEG) owns US small-cap (Russell 2000) and much of the UK/Asia. They rarely dislodge each other’s embedded standards. The economics are extraordinary (S&P DJI 69% GAAP margin; MSCI’s index segment runs a 76% EBITDA margin — prior the author work) for the same reason: an index is a rules-based intangible with near-zero marginal cost to license, while a fund named for it cannot change benchmark without changing its identity, tracking history, and prospectus.

The structural tailwind is the multi-decade shift from active to passive management, which compounds the asset-linked royalty base without adding cost. SPGI’s Indices asset-linked fees grew +18% in Q1-2026. The cyclical kicker is exchange-traded derivatives (SPX/VIX), which rise with volatility — a natural hedge inside an otherwise market-beta-sensitive annuity. The honest qualifier specific to SPGI: 27% of S&P DJI’s economics belong to CME, and the asset-linked line falls in an equity-market drawdown.

Financial data & analytics — the contested arena

Market Intelligence competes in the most crowded, least-moated of SPGI’s arenas. Bloomberg’s terminal (~33–36% share, ~$13B+ revenue) anchors the high end; LSEG/Refinitiv (~20%) has scale; FactSet, Moody’s Analytics, MSCI, and Morningstar fight for the rest. This is why MI earns only a 20% GAAP margin against Ratings’ 64%. It is also the segment where the GenAI substitution thesis is most credible and where the private-markets-data land grab is being fought — and SPGI is a follower, not the leader (BlackRock acquired Preqin for ~$3.2B in March 2025; Morningstar’s PitchBook, MSCI/Burgiss, and others are entrenched). SPGI’s answer is With Intelligence (acquired Q4 2025), Visible Alpha, and partnerships with Cambridge Associates and Mercer.

Energy / commodity benchmarks — a fading-market monopoly

Platts owns Dated Brent and a suite of price assessments that underpin trillions in physical and derivative contracts. Switching the reference price for the global crude market is a collective-action problem no competitor (Argus, OPIS, ICE) has solved. The 41% margin reflects genuine benchmark power. The long-term risk is not competitive but existential-to-the-market: the energy transition slowly shrinks the hydrocarbon pool the benchmarks price — though the same transition creates new demand (critical minerals, power, renewables data) that SPGI is positioning to capture.

Macro setup for 2026

Cyclically favorable for the biggest profit driver, with one structural swing factor. The 2026 maturity wall is large — over $8T of debt comes due through 2028, ~9% above the prior year — generating refinancing-driven ratings demand. Hyperscaler AI-infrastructure debt (~$600B of announced capex, increasingly debt-funded) is a fresh, large pool of investment-grade issuance that lifted Q1-2026 IG volumes. Against this, private credit (~$1.7T, forecast toward ~$5T by 2029) is the double-edged structural question discussed above.

Verdict: Structurally excellent where it matters most, contested where it matters least. Ratings (regulatory oligopoly) and Indices (brand/network oligopoly) are two of the best industry structures in financial services and produce ~65% of profit. Energy/Platts is a benchmark monopoly facing a slow secular fade. Market Intelligence is a genuinely competitive, lower-margin arena. The blended industry quality is high precisely because the profit is concentrated in the best-structured pools.

4. Competitive Position

The moat question must be answered segment by segment, because SPGI owns advantages of very different durability — and the entire variant-perception debate turns on where the profit sits relative to where the AI threat sits.

Ratings — regulatory license + reputation + network/standard (Greenwald’s strongest combination). The advantage is threefold and self-reinforcing: the NRSRO license is a legal barrier; the century-long track record is reputational capital that cannot be bought or built quickly; and the two-rating convention is a coordination standard that locks in the incumbent pair. The Greenwald market-share-stability test is passed emphatically — S&P’s share has been stable for decades — and the ROIC/margin test (64% margin) confirms it. This moat is almost entirely AI-immune. A credit rating’s value is not the analytical computation (which AI can cheapen) but the market’s collective agreement to require and trust it — a coordination good no model can replicate. If anything, AI lowers S&P’s own cost to produce ratings (management cites Ratings as an early AI adopter, expanding analytical capacity), widening the margin.

Indices — brand + demand-side network effect (the ABF flywheel). The S&P 500 is the global default equity benchmark; the more AUM tracks it, the more issuers default to it, the deeper the liquidity, the more investors benchmark to it. This is a demand-side network effect bolted onto a near-zero-marginal-cost intangible. Switching costs are coordinated and brutal (a fund cannot quietly change its benchmark). Also largely AI-immune — you cannot prompt your way around a benchmark the entire industry has contractually agreed to measure against. The marginal AI vector here is additive (clients licensing index content for AI use cases). The honest caveat: low-cost disruptors (Solactive) and self-indexing by mega-issuers (BlackRock, Vanguard) compress fees on new mandates at the margin, and 27% of the economics leak to CME.

Energy/Platts — benchmark standard. Same coordination-standard logic as Indices: Dated Brent is entrenched because everyone references it because everyone else does. Durable against competitors; vulnerable only to the slow shrinkage of the underlying market.

Market Intelligence — switching costs only, no standard (the soft underbelly). MI’s stickiness is real (embedded workflows, mission-critical Enterprise Solutions software, integration costs) but it is switching-cost captivity, not standard/network captivity — weaker and more contestable. It competes head-on with Bloomberg’s terminal lock-in and LSEG’s scale, which is why the margin is only 20%. This is where the GenAI threat genuinely bites: LLMs can commoditize data aggregation, retrieval, and basic modeling — exactly MI’s weaker, lower-differentiation layer. Management’s counter is credible but unproven: that MI’s value is the proprietary/curated content (Compustat, SNL — datasets literally hand-built from microfiche and not replicable), the business logic, and the workflow, and that AI expands the data-licensing TAM because models need licensed, structured ground-truth data. Early signals support this (AI-using MI customers growing 30% faster; clients paying 35–45% more on renewal for AI-ready data; >300 customers on Kensho LLM-ready APIs; API call volume up 5× in a quarter) — but it is early days, and MI is the one franchise where the bear case is not absurd.

The decisive structural fact for the thesis. Pulling the profit and the threat together: in FY2025, Ratings (~46% of segment OP) + Indices (~19%) ≈ 65% of segment operating profit comes from the two nearly-AI-immune franchises, while Market Intelligence — the AI-threatened segment — is ~32% of revenue but only ~15% of segment operating profit. The market has de-rated the whole company on a threat concentrated in its smallest profit contributor. Post-Mobility-spin, the Ratings+Indices share of RemainCo profit is higher still.

Verdict: A top-tier, durable moat in Ratings, Indices, and Energy/Platts (regulatory, reputational, and network/standard advantages that AI does not erode), and a weaker switching-cost position in Market Intelligence (the genuine AI battleground). Because ~65% of profit sits behind the strongest walls, the consolidated competitive position is excellent — and materially stronger than the de-rated multiple implies.

5. Growth History and Forward Opportunities

History — durable compounding, supercharged by IHS Markit. Revenue grew from $8.30B (FY2021) to $15.34B (FY2025), but the jump is distorted by the February 2022 all-stock acquisition of IHS Markit (~$43.5B), which roughly doubled Market Intelligence and added Energy/Mobility scale. On an organic basis, SPGI has compounded mid-to-high single digits: FY2025 organic constant-currency revenue grew ~7%, and adjusted diluted EPS grew +14% (operating leverage + buybacks). Over the period since its 2022 Investor Day, management cites ~9% annual revenue growth, ~50% margins, ~17% EPS growth, and ~$25B returned to shareholders — an elite track record.

The forward algorithm (management’s medium-term targets, 3–5 years, ex-Mobility):

• Total organic revenue growth: 7–9% annually.
• By segment: Market Intelligence 6–8%, Ratings 6–9% (plus cyclical upside), Energy 6–8%, Indices 10–12% (raised, on the active-to-passive shift).
• Adjusted margin expansion: 50–75 bps/year (more in cyclically strong years), with the largest opportunity in MI (where costs are ~65% of revenue).
• Adjusted EPS growth: double-digit.
• Capital return: ~85% of adjusted FCF in a typical year (history: nearly 2×).

FY2026 guidance (a deliberately cautious “trough-ish” year): organic constant-currency revenue +6–8%; margin expansion +50–75 bps (ex-OSTTRA); adjusted EPS $19.40–19.65 (+9–10%). Notably, management reiterated full-year guidance through the Q1-2026 Iran energy shock, trimming only Energy by ~1 point and offsetting with higher buybacks.

Where the growth comes from:

1. Ratings cyclical + secular. The 2026 maturity wall and hyperscaler IG issuance drive near-term volumes; the secular growth of global debt outstanding drives the long term. (Caveat: management expects ratings growth to turn negative in Q4-2026 on the brutal comp against a record 2025 — a known air-pocket, not a thesis break.)
2. Indices/passive. The structural active→passive shift compounds the asset-linked royalty base at a double-digit clip; new product (factor, thematic, fixed income, tokenized/blockchain indices) extends it.
3. Private markets. SPGI ended 2025 with >$600M of enterprise private-markets revenue, growing fast (Ratings private credit +25% in Q1-2026); With Intelligence, Visible Alpha, and the Cambridge/Mercer partnerships build a private-data franchise.
4. AI monetization. The genuinely new vector — LLM-ready APIs, MCP/agent distribution, the Claude for Financial Services plug-in, Kensho — with early evidence of pricing power (35–45% AI-data renewal uplifts).
5. Energy expansion. Power, renewables, and critical-minerals data riding the energy/AI-infrastructure build-out, offsetting the upstream/hydrocarbon fade.

Verdict: High-quality, mostly-organic growth with credible double-digit-EPS algorithm. The growth is real, diversified, and largely organic, with pricing power evident across the differentiated franchises. The quality caveats: Ratings growth is cyclical (and faces a 2026 comp air-pocket), MI must convert AI from threat to driver to hit 6–8%, and the headline EPS algorithm leans on continued aggressive buybacks. This is high-quality growth — not the highest-velocity, but among the most durable in the sector.

6. Financial Quality

Revenue and margins. FY2025 revenue $15.34B (+8% reported, ~7% organic). GAAP operating margin expanded to 42% (from 32% in 2023, as IHS Markit integration costs rolled off and scale built); adjusted operating margin runs ~49–50%, and at the segment level the spread is stark: Indices 69%, Ratings 64%, Energy 41%, Mobility 22%, MI 20% (GAAP). The trajectory is unambiguously up-and-to-the-right on both revenue and margin — operating leverage is working.

The central quality-of-earnings issue: GAAP vs. adjusted, and it is an amortization story, not a gimmick. The gap between FY2025 GAAP EPS ($14.66) and adjusted EPS (~$17.80) is dominated by ~$1.1B/year of acquisition-related intangible amortization — almost entirely the IHS Markit deal (Market Intelligence ~$588M, Mobility ~$303M, Energy ~$130M). This is a genuinely non-cash charge against a fixed asset base that is finite and will roll off; adding it back to reach adjusted EPS is legitimate. Two quality reassurances: (i) stock-based comp is small (~$236M, ~1.5% of revenue) and is NOT added back to adjusted EPS — so the adjustment is not the SBC-flattering kind that should be discounted; (ii) free cash flow (~$5.1B) sits close to adjusted net income (~$5.4B), confirming the add-back is genuinely non-cash and that adjusted earnings are backed by cash. The honest move is to value on adjusted EPS or FCF, treat the GAAP P/E as overstating the multiple, and not over-credit the adjusted figure either.

Cash generation. Operating cash flow $5.65B (FY2025), capex just $195M, NCI distributions $321M → company-defined adjusted FCF ~$5.1B. FCF conversion of adjusted net income is ~95%. This is a near-ideal cash machine: high margins, trivial capex, modest working capital.

Returns on capital — read carefully. Reported ROE (~14%) and P/B (~4×) are uninformative because the $43.5B IHS Markit deal buried $50B+ of goodwill and intangibles on the balance sheet (goodwill $36.5B + intangibles $16.3B > equity $31.2B → negative tangible book). The right metric is ROIC including goodwill, which sits around ~11–13% (adjusted NOPAT ~$5.9B over ~$47–48B of invested capital) — a thin-to-decent spread over an ~8% WACC, reflecting the high price paid for IHS Markit. But the underlying operating business earns spectacular cash-on-cash returns (the franchises need almost no capital); the modest consolidated ROIC is an artifact of acquisition accounting, not weak operations. (Same pattern as Thermo Fisher and Linde in prior the author work — goodwill-heavy roll-ups where ROIC understates operating quality.)

Balance sheet. Total debt ~$13.1B against ~$1.75B cash → net debt ~$11.3B, ~1.5× adjusted EBITDA (management targets 2.0–2.5× gross). Comfortably investment-grade (SPGI is, fittingly, A-rated). No liquidity concern. The redeemable NCI (~$4.9B, CME’s S&P DJI put) sits in mezzanine equity.

Verdict: Economics improve with scale, and the quality of earnings is high. Margins expand as revenue grows, capex is trivial, FCF backs adjusted earnings, and SBC is not used to flatter the adjusted number. The two caveats are presentational, not substantive: the GAAP-to-adjusted bridge is large (but legitimate, amortization-driven) and reported ROE/ROIC are depressed by IHS Markit goodwill (but the operating business is capital-light and high-returning). This is a financially excellent business.

7. Capital Allocation

The record is strong, with one expensive asterisk. Management’s framework is explicit and shareholder-friendly: fund organic growth first, return ~85% of adjusted FCF in a typical year (history: nearly 2×), grow the dividend every year, prefer bolt-on M&A, and explicitly reject “transformational” deals.

Dividends — Dividend King. SPGI has raised its dividend for 53 consecutive years — one of only a handful of S&P 500 companies to exceed 50 years (Dividend King status). The dividend (~$3.84/share FY2025, raised to $0.97/qtr in 2026) is a modest ~22% of adjusted EPS, leaving ample room for buybacks and growth. ~$1.17B paid in FY2025.

Buybacks — and the read on the current weakness. SPGI repurchased $3.3B/year in 2023–24, stepped up to $5.0B in 2025, and — critically — raised the 2026 target from 85% to 100%+ of adjusted FCF (~$4.5B), citing the share price as “an attractive opportunity.” The Mobility spin will fund additional buybacks (the ~$2B Mobility debt is dividended back to SPGI). The Board authorized another 30M shares (~10% of shares out) in November 2025. Diluted share count has fallen steadily from 318.9M (2023) to 305.1M (2025). Management buying back 100%+ of FCF into a 28th-percentile valuation is a meaningful capital-allocation signal — they are voting that the stock is cheap.

M&A — the IHS Markit asterisk. The 2022 IHS Markit deal (~$43.5B all-stock) is the defining capital-allocation act of the decade and the verdict is mixed. Strategically it was sound — it scaled Market Intelligence, added Energy/Mobility, and was all-stock at a moment SPGI’s currency was rich. But it was expensive: it loaded $50B+ of goodwill/intangibles, depressed consolidated ROIC to low-teens, and the segment it most enlarged (MI, 20% margin) is the weakest. Since then management has been disciplined: bolt-ons only (With Intelligence, ORBCOMM AIS), and active divestitures of non-core/lower-quality assets (Engineering Solutions 2023; OSTTRA stake 2025 for a $270M gain; EDM and thinkFolio Jan 2026; the Energy software portfolio to SLB, closing 2H26-early-27). The Mobility spin is the capstone — shedding a 22%-margin auto-data business to sharpen the franchise. This is exactly the right post-deal behavior: integrate, then prune.

Incentive alignment. Compensation is tied to revenue, adjusted EPS/margin, and relative TSR (per the proxy) — reasonable metrics, though as with most data/ratings names the absence of an explicit ROIC target is a mild critique given the goodwill on the books.

Verdict: Management has allocated capital well, with IHS Markit as a defensible-but-pricey exception now being actively cleaned up. The combination of a 53-year dividend streak, disciplined post-deal pruning, an opportunistic step-up in buybacks into weakness, and a portfolio-sharpening spin-off is a strong capital-allocation profile. The only real demerit is the price paid for IHS Markit — and even that was strategically coherent and is being mitigated by subsequent discipline.

8. Changes and Headwinds — Last Two Years

Strategic / corporate actions:

• Mobility spin-off (announced Apr 29, 2025; completion ~July 2026). The defining structural change — separating CARFAX/auto-data into “Mobility Global.” Sharpens SPGI into a four-segment ~$13B capital-markets-information company; ~$2B Mobility debt funds SPGI buybacks/deleveraging; ~$20–25M of standalone “stranded costs” to be addressed.
• Leadership transition. Martina Cheung became President & CEO (succeeding Doug Peterson) effective late 2024; Eric Aboaf joined as CFO. A new CTO/Transformation Officer (Firdaus Bhathena) was hired to scale AI. The November 2025 Investor Day was the new team’s strategic reset.
• Portfolio pruning: Engineering Solutions sold (2023); OSTTRA JV stake monetized (2025, $270M gain); EDM and thinkFolio divested (Jan 2026); Energy upstream software portfolio being sold to SLB (closing 2H26/early-27, with a new SLB distribution partnership). With Intelligence acquired (Q4 2025) for private-markets data.
• Segment rename: “Commodity Insights” → “S&P Global Energy” (FY2025).

Strategic pivots: A hard pivot to AI — Kensho-powered LLM-ready APIs, MCP/agent distribution, the Claude for Financial Services plug-in, Gemini Enterprise and Copilot integrations, and an Enterprise Data Office targeting >20% run-rate cost reduction by end-2027. A push into private markets and decentralized finance (rating stablecoins, tokenized S&P 500 on blockchain, first Bitcoin-backed ABS).

Headwinds (all live in 2026):

• Iran/energy shock: management calls it “the largest energy shock since the 1970s”; it cut Energy guidance ~1 point and pressures near-term subscription decisions in that segment.
• Ratings comp air-pocket: issuance growth expected to decelerate through 2026 and turn negative in Q4 against record 2025 volumes — optically alarming but a known, cyclical comp effect.
• Private-credit jitters: wider spreads, elevated BDC redemptions, and regulatory scrutiny (“ratings shopping”) in late 2025/early 2026.
• GenAI de-rating: the sector-wide multiple compression on AI-substitution fear (the proximate cause of the share-price weakness).

Verdict: The changes strengthen the franchise; the headwinds are mostly cyclical and already in guidance. The Mobility spin and the disciplined divestitures raise average business quality; the new management team’s AI/private-markets strategy is coherent. The headwinds (energy shock, ratings comp, private-credit jitters) are real but cyclical and acknowledged in guidance — they explain the entry point more than they threaten the thesis. The one genuinely structural item — GenAI — is judged to be mispriced relative to the profit actually at risk.

9. Risk Analysis

Catastrophic-loss risk is low. SPGI is not a balance-sheet-levered, single-product, or going-concern story — it is a diversified, cash-generative, investment-grade collection of moats. The realistic bear outcome is multiple-compression-plus-cyclical-EPS-stall (a flat-to-down stock for a year or two), not impairment of intrinsic value. The genuine long-tail risk is a structural one-two punch: GenAI hollowing MI and private credit eroding the ratings annuity simultaneously — a scenario specified below as the bear falsification test.

10. Valuation Discussion — Embedded Expectations

Where it trades. At ~$426, SPGI carries a ~21.8× forward P/E on FY2026 adjusted EPS (guidance midpoint ~$19.52), ~24× trailing adjusted EPS (~$17.80), ~29× trailing GAAP EPS ($14.66), ~16.6× EV/EBITDA, and ~8.6× EV/revenue. The dividend yield is ~0.9%.

The single most important valuation fact: SPGI is cheap relative to itself. On its own ~10-year valuation history, SPGI sits at the ~28th percentile (P/E ~29th, P/B ~24th, P/S ~31st) — the cheapest third of its decade. It has historically commanded ~28–32× forward earnings; today’s ~21.8× is a roughly 25–30% multiple compression. This matters more than any cross-sectional comparison because the franchise quality has not deteriorated — if anything, the Mobility spin improves it.

Peer context (cross-sectional). Within the data/index/exchange complex : MSCI trades ~27× EV/EBITDA / ~35× earnings at the 39th percentile of its range (it never de-rated); Moody’s, FactSet (3rd percentile, −43%), Verisk, ICE, CME, and Nasdaq round out the set. SPGI at ~21.8× forward is a discount to MSCI (~50% richer multiple) and roughly in line with-to-below Moody’s — despite arguably the best diversified franchise mix in the group. FactSet’s collapse shows the de-rating can overshoot; MSCI’s resilience shows quality can hold a premium. SPGI sits in between, which is the opportunity.

Embedded-expectations / reverse-DCF read. At ~$426 and ~21.8× forward EPS, the market is pricing SPGI for roughly mid-single-digit earnings growth in perpetuity — materially below management’s 7–9% organic revenue + double-digit EPS algorithm, and below the ~14–17% adjusted-EPS growth actually delivered in recent years. In other words, the price embeds a meaningful deceleration — consistent with the bear thesis that AI/private-credit structurally slows the business. If SPGI merely delivers its guided +9–10% 2026 EPS and holds a low-double-digit growth rate while the multiple normalizes toward the low end of history (~26–28×), the equity compounds attractively from here.

Scenario analysis (directional, illustrative — not a price target):

• Bear (~$360–390): GenAI hollows MI and private credit erodes Ratings; organic growth slows to low-single-digits; multiple stays compressed at ~19–20× a stalled ~$19 EPS. Return = roughly flat-to-down.
• Base (~$530–590): Management delivers the guided algorithm (7–9% organic, double-digit EPS); the GenAI fear fades as AI monetization shows up in MI; multiple normalizes to ~27–28× a FY2027 adjusted EPS approaching ~$21. Roughly the street’s ~$533 target and the low end of SPGI’s historical multiple.
• Bull (~$650+): AI data-licensing becomes a visible, fast-growing revenue line; private-credit ratings inflect up; Indices passive tailwind accelerates; multiple re-rates toward the historical ~30× on accelerating EPS.

Verdict (embedded expectations): The market is underwriting structural deceleration in a franchise still guiding to double-digit EPS growth, with the de-rating concentrated on a threat (AI) that is most valid for the smallest profit contributor (MI). The price embeds the bear case more than the base case. No price target; no recommendation.

11. Variant Perception

Consensus view. SPGI is a high-quality compounder, but one whose growth is decelerating and whose data business is structurally threatened by generative AI and private-credit disintermediation — hence the de-rating from ~30× to ~22× forward earnings. The Street’s ~$533 average target implies it still sees upside, but the multiple says the market is skeptical the franchise can hold its historical growth premium.

The strongest bull case (the variant perception this memo finds compelling). The market has mis-located the AI threat. It has de-rated the entire company as if it were a data vendor, when ~65% of profit comes from Ratings and Indices — coordination-standard monopolies that AI cannot disintermediate (you cannot prompt your way around a required credit rating or an embedded benchmark) and that AI actually makes cheaper to operate. The AI fear is valid only for Market Intelligence, which is ~32% of revenue but ~16% of profit, and even there early evidence (35–45% AI-data renewal uplifts, 5× API-call growth) suggests AI expands the licensing TAM. Layer on a Dividend King buying back 100%+ of FCF into a 28th-percentile multiple, a portfolio-sharpening spin-off, and a record refinancing wall — and the de-rating looks like a sector-wide fear overshooting onto one of its highest-quality members.

The strongest bear case (taken seriously). Three things could be structurally true at once: (1) GenAI genuinely commoditizes MI’s data/workflow layer, capping it well below the 6–8% target and pressuring its 20% margin; (2) private credit grows toward $5T and an increasing share bypasses public ratings (or routes to cheaper NRSROs via ratings-shopping), eroding the Ratings transaction annuity; (3) the multiple never re-rates because the market permanently lowers its growth assumption for the whole financial-data complex. In that world, SPGI is a low-double-digit-margin-expansion, mid-single-digit-grower worth ~20× — i.e., roughly today’s price is fair, not cheap.

The 3–5 assumptions that matter most:

1. AI is net-additive to SPGI’s data franchises, not substitutive (bull) vs. net-substitutive to MI (bear). Falsifiable by: MI organic growth and ACV trend, AI-data pricing uplifts, retention.
2. Public ratings retain their share of debt issuance as private credit grows. Falsifiable by: S&P Ratings’ private-credit revenue growth vs. NRSRO competitors; the public-vs-private issuance mix.
3. The Indices passive tailwind and asset-linked fees keep compounding double-digit. Falsifiable by: ETF AUM flows, asset-linked fee growth, fee-rate compression.
4. The multiple normalizes toward historical levels as fear fades. Falsifiable by: the forward P/E percentile over the next 12–24 months.
5. Management sustains the capital-return + buyback algorithm. Falsifiable by: buyback pace, leverage, dividend growth.

What would falsify each side. The bull case breaks if MI organic growth slips below ~5% with cracking retention and Ratings loses measurable private-credit share — the AI/disintermediation thesis made real in the numbers. The bear case breaks if AI data-licensing shows up as a visible ACV uplift, MI sustains 6%+ organic, and the multiple re-rates — at which point the de-rating is revealed as a fear that overshot.

12. Fact vs. Interpretation Table

13. Open Questions

1. Exact post-spin RemainCo financials — recast 2025 quarterly figures ex-Mobility (revenue, margin, EPS) and the precise stranded-cost drag; management will provide upon completion. (FY2026 guidance still consolidates Mobility.)
2. MI organic growth and AI monetization trajectory — is the 35–45% AI-data pricing uplift broad-based or anecdotal? Does ACV growth sustain at ~6%+? This is the crux of the bull/bear split.
3. Private-credit ratings economics — how much of the >$600M private-markets revenue is durable, and is S&P holding share against KBRA/DBRS as the pool grows?
4. Insider activity — Form 4 bodies were not mirrored (corpus fetched without --all-form4); 335 Form 4s + 44 Form 144s over five years is consistent with routine grant/10b5-1-sale behavior typical of a mega-cap, but no open-market-purchase signal could be confirmed. Open.
5. The exact FY2025 adjusted operating margin and adjusted EPS — the earnings-release exhibit (EX-99.1) was not in the local corpus; figures here are derived from guidance growth rates and transcript statements and should be reconciled to the release.
6. Indices fee-rate compression — how fast are asset-linked basis-point fees eroding as mega-issuers gain leverage, and does volume growth keep swamping it?
7. CME relationship — any path to SPGI consolidating the remaining 27% of S&P DJI, and at what cost?

14. What Must Be True

For the bull case to be right (and the de-rating a gift):

1. Ratings and Indices remain AI-immune coordination monopolies — credit ratings stay market-required and S&P benchmarks stay embedded, so ~65% of profit keeps compounding regardless of GenAI.
2. Market Intelligence converts AI from threat to driver — MI sustains ≥6% organic growth with stable/rising retention and visible AI-data monetization, proving the data-licensing TAM expands rather than collapses.
3. The multiple normalizes — as the cyclical fears (energy, ratings comps, private-credit jitters) pass and AI monetization becomes visible, the forward P/E re-rates from ~22× toward the low end of its historical ~28–30× range.
   • Falsification test (bull): MI organic growth falls below ~5% for two-plus consecutive quarters with retention deteriorating, OR S&P Ratings demonstrably loses private-credit share to competing NRSROs — either would confirm the structural-deceleration thesis the price embeds, and the bull case is wrong.

For the bear case to be right (and today’s price fair-to-rich):

1. GenAI structurally commoditizes MI — data aggregation and basic analytics migrate to LLMs, capping MI below 5% organic and pressuring its 20% margin.
2. Private credit disintermediates the ratings annuity — a growing share of the ~$1.7T→$5T private-credit pool bypasses public ratings or routes to cheaper NRSROs, eroding Ratings transaction revenue.
3. The growth premium is permanently lost — the market lowers its long-run growth assumption for the whole financial-data complex, so the multiple stays at ~20× even as EPS grinds higher.
   • Falsification test (bear): AI data-licensing shows up as a measurable, fast-growing revenue line (or a clear ACV inflection), MI holds ≥6% organic, and the forward multiple re-rates above ~26× — any of which would show the de-rating overshot and the bear thesis is wrong.

15. Source Appendix

See the separate Source Appendix (SPGI_source_appendix.md) and Diligence Questionnaire (SPGI_diligence_appendix.md) accompanying this memo. Primary sources: S&P Global FY2021–FY2025 10-Ks and FY2026 10-Qs (SEC EDGAR, CIK 0000064040); Q4-2025 (Feb 10, 2026) and Q1-2026 (Apr 28, 2026) earnings calls; the November 13, 2025 S&P Global Investor Day and May 12, 2026 Mobility Global Investor Day transcripts; SEC NRSRO statistics; and SPGI investor-relations materials. Quantitative figures reconciled to EDGAR XBRL and the AZI fundamentals/valuation feeds. Peer cross-reads drew on prior the author reports for MSCI, FactSet, ICE, CME, and Verisk.

The institutional body of this memo takes no position, sets no price target, and makes no buy/sell recommendation. The only position expressed in this document is in the clearly-labeled “Claude’s Take” block at the top, which is Claude’s own subjective opinion and general information only, not investment advice.

APPENDIX A — Standard Diligence Questionnaire

A diligence checklist on the company. Facts, interpretations, and assumptions are labeled where it matters.

General

What thoughtful questions have other investors asked about this company? The sell-side Q&A (Q1-2026 call) clusters tightly on the AI question: how SPGI monetizes data through MCP/plug-in/LLM channels vs. its own desktop; whether AI is net-additive or substitutive to Market Intelligence; how much margin expansion is AI-driven efficiency. Secondary clusters: the cadence of Ratings/billed-issuance through 2026 (and the Q4 negative comp), private-credit trends and SPGI’s exposure, the Energy/Upstream transformation and SLB software divestiture, and the Mobility spin mechanics/recast financials. The deepest investor debate (INTERPRETATION) is whether SPGI deserves its historical ~30× multiple or has structurally de-rated — i.e., is the GenAI threat to the data business real enough to permanently lower the growth premium.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Mixed and segment-dependent (INTERPRETATION). Ratings transaction revenue is near a cyclical high (record $4.3T+ billed issuance in 2025; management guides Q4-2026 issuance negative on comps) — so the Ratings annuity is closer to a peak than a trough. Conversely, the multiple is near a cyclical low (28th percentile of its decade), and Energy is being depressed by the Iran shock and sanctions (a trough). Net: operating earnings are healthy-to-high, the valuation is low.

Driven by external environment or internal actions? Both. Ratings/Indices transaction/asset-linked lines are externally driven (issuance volumes, equity markets, volatility); margin expansion, AI efficiency, buybacks, and the portfolio reshaping (spin/divestitures) are internal.

How stable are revenues? ~75%+ recurring/subscription at the enterprise level (FACT). The cyclical slugs are Ratings transaction revenue (~52% of Ratings) and Indices asset-linked/derivatives fees. Even the cyclical lines are diversified across debt and equity markets, providing natural offsets (e.g., SPX/VIX derivatives rise with volatility).

Outlook for products/services & market size? Growing. Management targets 7–9% organic revenue growth over 3–5 years (Indices 10–12%, MI 6–8%, Ratings 6–9%, Energy 6–8%). The MI TAM is framed at ~$80B+ growing 5–6%; global debt issuance and passive AUM provide multi-decade tailwinds.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Bifurcated. Ratings and Indices remain stable oligopolies (no entrant has dented them in 15+ years). Market Intelligence is getting more competitive (Bloomberg/LSEG plus AI-native entrants plus the private-data land grab). Energy/Platts is stable competitively but faces a shrinking long-run hydrocarbon market.

How profitable (ROIC/ROE)? ROE ~14% and ROIC ~11–13% as reported — but both are depressed by ~$50B of IHS Markit goodwill/intangibles (FACT). The operating business is capital-light and earns spectacular cash-on-cash returns; the modest consolidated ROIC is an acquisition-accounting artifact (INTERPRETATION). Segment margins: Indices 69%, Ratings 64%, Energy 41%, Mobility 22%, MI 20%.

How profitable is the industry — competitors, barriers? Ratings: ~2 dominant firms (S&P, Moody’s) at ~50–65% margins, near-impregnable (NRSRO license + two-rating convention + reputation). Indices: ~3 firms (S&P DJI, MSCI, FTSE Russell) at 69–76% margins, brand/network barriers. Both are among the best industry structures in finance.

Can the business be easily understood? Yes at the franchise level (tolls on debt issuance and equity indexing), though the five-segment structure and GAAP-vs-adjusted bridge require work.

Undermined by foreign low-cost labor? No — the moats are regulatory, reputational, and network/standard, not labor-cost. AI is the relevant disruption vector, not offshoring.

Do brands matter? Critically. “S&P 500,” “S&P,” and “Platts/Dated Brent” are the brands — they are the moats (coordination standards the market has agreed to use).

Nature of competition / switching costs? Ratings/Indices/Energy: switching costs are coordination-level (an entire market must move) — effectively prohibitive. MI: switching costs are workflow-integration-level (high but contestable, the FactSet/Bloomberg battleground).

Financial Condition & Balance Sheet

Assets not fully on the balance sheet? Yes — the brands (S&P 500, Platts), the NRSRO franchise, and the proprietary datasets (Compustat, SNL) are worth far more than their carrying value; conversely the IHS Markit goodwill overstates tangible book.

Off-balance-sheet liabilities? Nothing material flagged; the CME put on S&P DJI sits as redeemable NCI (~$4.9B, mezzanine). The IHS-related deferred tax liability (~$1B, mostly migrating to Mobility) is on-balance-sheet.

How conservative is the accounting? Reasonable. The key non-GAAP adjustment (~$1.1B/yr acquisition amortization) is legitimate and non-cash; SBC is small and not added back to adjusted EPS (a conservative tell). FCF backs adjusted earnings (~95% conversion).

How CapEx-hungry? Trivially light — capex ~$195M (~1.3% of revenue). A near-ideal asset-light royalty model.

Capital Allocation & Management

FCF generation and use? ~$5.1B adjusted FCF (FY2025). Framework: fund organic growth → return ~85% of FCF (history ~2×) via dividends + buybacks → bolt-on M&A only. 2026 buybacks raised to 100%+ of FCF (~$4.5B). 53 consecutive years of dividend increases (Dividend King).

Significant acquisitions? IHS Markit (2022, ~$43.5B all-stock) — transformative, strategically sound but expensive (loaded the goodwill that depresses ROIC). Since: bolt-ons only (With Intelligence, ORBCOMM AIS) and active divestitures (Engineering Solutions, OSTTRA, EDM/thinkFolio, Energy software to SLB). Mobility spin is the capstone pruning.

Buying back shares? Yes, aggressively — diluted shares 318.9M (2023) → 305.1M (2025); $5.0B in 2025, ~$4.5B guided 2026, plus a new 30M-share authorization. Buying into a 28th-percentile multiple is a positive signal.

Issuing shares to insiders? No — SBC is small (~$236M, ~1.5% of revenue); share count is falling.

Compensation / incentives? Tied to revenue, adjusted EPS/margin, and relative TSR (proxy). Reasonable; the mild critique is no explicit ROIC target given the goodwill on the books.

Motivations of management? New team (CEO Cheung, CFO Aboaf, CTO Bhathena since ~2024–25) executing a coherent AI/private-markets/portfolio-sharpening strategy. INTERPRETATION: well-aligned, shareholder-friendly capital return.

Valuation & Market Data

ADR/MLP/K-1? No — US-domiciled C-corp, files 10-K/10-Q, common stock on NYSE. Standard 1099 dividend.

Dividend policy? ~$3.84/share (2026: $0.97/qtr), ~0.9% yield, ~22% of adjusted EPS — low payout, 53-year growth streak. Buybacks are the primary return vehicle.

How profitable? Very — 42% GAAP / ~49–50% adjusted operating margin; 30% net margin; ~95% FCF conversion.

Net income vs. cash from operations? Aligned. GAAP NI $4.47B vs. OCF $5.65B (FY2025); the OCF > NI gap is the non-cash amortization add-back — a quality positive, not a divergence concern.

Risks & Downside

What would cause the stock to decline? A ratings-cycle downturn (issuance freeze); GenAI demonstrably commoditizing Market Intelligence; private credit disintermediating the ratings annuity; an equity-market drawdown hitting Indices asset-linked fees; or simply the multiple staying compressed as the growth premium is permanently re-rated lower.

Risk of catastrophic loss? Low. Diversified, investment-grade, cash-generative collection of moats with no balance-sheet fragility. The realistic bear outcome is a flat-to-down stock (multiple compression + cyclical EPS stall), not impairment of intrinsic value.

Chance of total loss? Negligible — A-rated balance sheet, multiple durable franchises, no going-concern risk.

Recent News & Events

Has the business environment changed recently? Yes — (1) GenAI sector-wide de-rating of financial-data names (the proximate cause of SPGI’s ~26% drawdown); (2) Iran/energy shock (“largest since the 1970s”) pressuring the Energy segment; (3) private-credit stress (wider spreads, BDC redemptions, regulatory scrutiny) in late-2025/early-2026.

Significant acquisitions/divestitures? With Intelligence acquired (Q4 2025, private-markets data); EDM and thinkFolio divested (Jan 2026); Energy software portfolio sale to SLB (closing 2H26/early-27); OSTTRA stake monetized (2025).

Change in accounting policies? None material; “Commodity Insights” segment renamed “S&P Global Energy” (FY2025).

Recent changes — markets, facilities, management? New CEO/CFO/CTO (~2024–25); the Mobility/CARFAX spin-off (completing ~mid-2026); a strategic pivot into AI (LLM-ready APIs, MCP, Kensho, Claude/Gemini/Copilot partnerships), private markets, and decentralized finance (stablecoin/tokenization ratings).

APPENDIX B — Source Appendix

Primary sources prioritized over secondary; all figures reconciled to filings where possible. Accessed June 11, 2026.

Primary — SEC filings (EDGAR, CIK 0000064040)

• FY2025 10-K (filed 2026-02-11; period ended 2025-12-31) — segment revenue/operating profit/margins; consolidated income statement, balance sheet, cash flow; acquisition-intangible amortization disclosure (~$1.1B/yr); FCF reconciliation; dividend track record; share counts.
• FY2021–FY2024 10-Ks (filed 2022-02-08, 2023-02-10, 2024-02-09, 2025-02-11) — multi-year revenue/NI/margin history; IHS Markit purchase accounting (FY2023 10-K Note 2: closed Feb 28, 2022, all-stock, ~$43.5B, exchange ratio 0.2838, goodwill $31,456M + intangibles $18,620M).
• Q1 2026 10-Q (filed 2026-04-28; period ended 2026-03-31) — latest quarterly financials.
• DEF 14A proxy (2026) — executive compensation metrics; 53-year dividend-increase disclosure.
• 8-Ks — quarterly earnings releases; Mobility separation announcements; buyback authorizations.
• Form 4 corpus (335 filings over trailing 5 yrs) — reviewed for insider-transaction signal.

Primary — earnings calls & investor events (transcripts)

• Q1 2026 earnings call (Apr 28, 2026) — revenue +10%, adj EPS +14%; segment detail; Iran energy shock; AI/MCP monetization; Mobility spin timeline; buybacks raised to 100%+ of FCF (~$4.5B).
• Q4 2025 earnings call (Feb 10, 2026) — FY2025 results; initial FY2026 guidance (adj EPS $19.40–19.65, +9–10%); billed issuance $4.3T record; maturity-wall outlook.
• S&P Global Investor Day (Nov 13, 2025) — 3 strategic pillars; medium-term targets (7–9% organic revenue, 50–75 bps margin/yr, double-digit EPS, ~85% FCF return); >95% differentiated-revenue framing; MI TAM ~$80B+; AI strategy.
• Mobility Global Investor Day (May 12, 2026) — spin-co financial profile (~$1.75B revenue, ~40% adj EBITDA margin), standalone cost detail (~$20–25M stranded), spin timeline.
• Special call (Apr 7, 2026, “Capital IQ: Reinvented”) — MI/GenAI product showcase (Visible Alpha, With Intelligence, Kensho, ChatIQ, Document Intelligence).

Quantitative data feeds

• SEC EDGAR XBRL — revenue, net income, operating income, diluted shares, equity, goodwill, long-term debt, OCF, dividends, buybacks (multi-year series, reconciled).
• Own-history valuation percentiles — SPGI’s current multiples sit at roughly the 28th percentile of its own ten-year range (P/E ~29th, P/B ~24th, P/S ~31st, as of June 10, 2026).
• Market data — price $426.38, 52-wk $381.61–$579.05, EV/market-cap cross-check.

Secondary — industry & competitive sources

• SEC NRSRO Annual Staff Report and NRSRO Statistics (2025) — ratings market shares (S&P ~48.9%, Moody’s ~34.2%, Fitch ~13–15%).
• Public compilations of credit-rating-agency market structure and the issuer-pays model; Dodd-Frank Rule 17g commentary.
• S&P Dow Jones Indices / CME joint-venture ownership (CME 27%).
• Index-industry competitive set (MSCI, FTSE Russell/LSEG, Bloomberg, Nasdaq, Solactive); passive-investing AUM/flow data.
• Financial-data competitive set (Bloomberg ~33–36%, LSEG/Refinitiv ~20%, FactSet, Moody’s Analytics, Morningstar).
• Platts/Dated Brent benchmark coverage; competitors Argus, OPIS, ICE.
• 2026 maturity-wall (>$8T through 2028), hyperscaler issuance (~$600B capex), and private-credit (~$1.7T→~$5T by 2029) data; FSB/Bloomberg private-credit “ratings shopping” scrutiny (Dec 2025–May 2026).
• BlackRock–Preqin acquisition (~$3.2B, closed Mar 3, 2025).

Peer context

• Public filings and disclosures of peer index, ratings, exchange, and financial-data companies (MSCI, Moody’s, FactSet, Intercontinental Exchange, CME Group, Verisk) were used for industry-structure framing and cross-sectional valuation context. All SPGI-specific conclusions rest on the primary sources above.