SPGI · Research
S&P Global keeps the ratings gate and weighs a Capital IQ separation
S&P Global is reviewing a separation of Capital IQ and associated data businesses while keeping RatingsDirect, RatingsXpress and Enterprise Solutions. Market Intelligence earned a 24% Q3 operating margin against 66% at Ratings and 70% at Indices.
Scorecard
| Scorecard | Read |
|---|---|
| Verdict | Watch. No base value set: one clean continuing-operations year of owner cash, and the house rule needs three |
| Thesis | S&P is keeping the assets that carry its own ratings record and reviewing the desktop that repackages data others can reach |
| What proves it wrong | The perimeter grows to include RatingsXpress or RatingsDirect, or Kensho Data & Platforms margins rise to Ratings levels before any sale |
| Next catalyst | The Q4 2026 print and 10-K, the first chance at a revenue split for the reviewed businesses (date unconfirmed) |
| Key multiple | 17.25x EV / twelve-month GAAP EBITDA of $8.16B; 27.9x EV / owner cash of $5.05B |
| Debt | $13.06B carrying at Sep 30, 2026 from Note 4: senior notes due 2027 to 2062 plus $700M of commercial paper [1] |
| Stock pay | $127M for nine months from the cash flow line [1] |
What is actually under review
We cover S&P Global for one reason: Capital IQ is one of the notable financial software terminals, a desktop suite that, in the filing's words, provides "data, analytics and third-party research for global finance and corporate professionals" [1]. We have used the product.
That makes the review a software story, not only a ratings story, and it is the lens for this research.
On Oct 27, 2026, S&P Global said it "has initiated a strategic review of alternatives, including potential separation, for Capital IQ and associated data businesses, which constitute the majority of Kensho Data & Platforms within Market Intelligence" [2]. The release adds that "the exact perimeter remains under review" and that the company "expects to retain RatingsDirect, RatingsXpress, and Enterprise Solutions" [2].
CEO Martina Cheung said the businesses in the review "have significant opportunities ahead of them as technology changes how customers access, analyze, and use information" [2].
The story did not start with the release. Bloomberg wrote on Sept 1, 2026 that S&P was in early talks about a Capital IQ Pro spinout, possibly as a listed company [3].
Bloomberg's unnamed sources put a possible valuation in the high single-digit billions of dollars [3]. That figure is press coverage from unnamed sources, not a company number, and we do not build on it.
Own the record, own the connection
Our toll-gate test asks four questions. Does the business own the record customers rely on?
Does it own the connection into their workflow? Does it price on outcomes?
Does it spend on the gate rather than on features?
The record. A credit rating is S&P's own record.
In our reading, issuers pay for it, mandates and contracts often point to it, and nobody else can produce an S&P rating. RatingsDirect and RatingsXpress are the channels that deliver that record as data, and S&P is keeping both [2].
Capital IQ is a different kind of asset. In our reading, most of what it shows (filings, transcripts, prices, estimates, third-party research) is assembled from sources other firms can license or collect.
Its value comes from cleaning, linking and presenting that data, not from owning it. The filing describes Kensho Data & Platforms as covering "both the public and private capital markets" [1], and the November 2025 With Intelligence acquisition added private markets data [1].
Proprietary private markets data is the part of the reviewed perimeter closest to an owned record.
The connection. This is where Capital IQ is strongest.
The desktop, the Office add-in and the Mobile products sit inside analyst and banker workflows, and data feeds and APIs push the same content into client systems [1]. Workflow habits are sticky.
Pricing and spending
Pricing. Market Intelligence is a subscription business.
Q3 subscription revenue was $1,049M, or 83% of the segment, up 6.39% [1]. Recurring variable revenue, which the filing ties to trades processed, assets under management or positions valued, was $176M, or 14%, up 11.39% [1].
Non-subscription revenue was $38M, or 3% [1].
| Market Intelligence ($M) | Q3 2026 | Q3 2025 | Change | 9M 2026 | 9M 2025 | Change |
|---|---|---|---|---|---|---|
| Revenue | 1,263 | 1,183 | 7% | 3,723 | 3,479 | 7% |
| Subscription | 1,049 | 986 | 6% | 3,076 | 2,900 | 6% |
| Recurring variable | 176 | 158 | 11% | 518 | 467 | 11% |
| Non-subscription | 38 | 39 | (2)% | 129 | 112 | 15% |
| Operating profit | 301 | 252 | 19% | 979 | 665 | 47% |
| Operating margin | 24% | 21% | 26% | 19% |
The fastest-growing line is the one priced on usage, and it is the smallest. Nothing in the filing points to outcome pricing in the reviewed businesses [1].
Fees track access, not the result a customer gets. That matters for the superintelligence section below.
Margins. Market Intelligence's 24% Q3 margin is weighed down by $154M of amortization of acquired intangibles in the quarter [1], which in our reading stems largely from the IHS Markit merger.
Excluding severance, deal costs and amortization, the filing says operating profit rose 10% in Q3, held back by With Intelligence costs and "an increase in technology costs"
The superintelligence threat
The case against a desktop is easy to state. A superintelligence agent can read EDGAR, earnings transcripts and press releases directly, build a model and answer the question an analyst used to open Capital IQ to answer.
When the raw inputs are public, the value of aggregating them falls as agents get cheaper. That is our reading, and the filings give no numbers that test it.
The case for the kept assets runs the other way. An agent that needs a credit view still needs the S&P rating, because the rating is the record that contracts, mandates and regulators point to.
RatingsXpress and RatingsDirect deliver that record as data. Every new agent reading credit data is another channel that wants the feed.
On this reading, superintelligence raises the value of an owned record and cuts the value of a screen that displays shared data.
Management's own words fit this split. Cheung tied the review to how "technology changes how customers access, analyze, and use information" [2].
The July reorganization was built around agentic workflows [6]. We read the review as S&P deciding it would rather own the inputs agents need than defend the screen agents replace.
There are two counterpoints. First, Capital IQ's feeds and APIs already serve machines, and well-linked, cleaned data is hard for an agent to rebuild from raw filings at scale.
Second, private markets data has no EDGAR to read, so With Intelligence and similar sets may hold value that public company data does not. Either could make the reviewed business worth more than its 24% segment margin suggests.
What a separation leaves behind
Group results. Q3 2026 revenue was $3,678M, up 7% from $3,443M [1].
Nine-month revenue was $11,074M, up 9% [1]. Q3 operating profit was $1,956M, a 53% margin, against 45% a year earlier, and included a $200M gain on dispositions [1].
Q3 GAAP net income was $1,298M, or 35.29% of revenue [2]. S&P guides 2026 revenue growth of 7.4% to 8.4% [2].
Twelve months, continuing operations. The SEC data tags for 2025 still include Mobility, so we rebuilt the twelve months from the Oct 23, 2026 pro forma filing, which shows 2025 continuing revenue of $13,589M, operating profit of $6,105M, depreciation of $96M and amortization of $766M [5].
Removing Mobility cut 2025 revenue from $15,336M to $13,589M and lifted the operating margin from 42.24% to 44.93% [5]. A Capital IQ separation would restate the group again in the same way.
Adding nine months of 2026 and removing nine months of 2025 [1]
Risks and catalysts
Perimeter risk. If RatingsXpress or RatingsDirect end up in the deal, the thesis that S&P keeps the gate weakens.
Status: unconfirmed, perimeter under review [2].
No deal. The review may end without a transaction [2].
Unconfirmed.
Dis-synergies. Kensho Data & Platforms delivers ratings data alongside Capital IQ content [1].
Splitting shared delivery could raise costs at both companies. Unquantified.
Credit cycle. Ratings is the largest segment at $1,271M of Q3 revenue, growing 2% [1].
Issuance-driven revenue can fall quickly.
Q4 2026 print and 10-K. Expected early 2027, date unconfirmed.
This is the first chance for a revenue split.
2027 maturities. $1,748M of notes come due in 2027 [1].
Confirmed in Note 4.
Leadership and control
Martina Cheung is President and CEO [2]. S&P has one class of common stock, with 289.9M shares outstanding as of Oct 23, 2026, excluding 7.2M held by the Markit employee benefit trust [1].
We have not verified 5% holders this pass, and they are not cited here.
Stock Profiles & Comparable Companies
SPGI Stock Profile →More Published Research
View All →Oracle's $20B cap leaves out the $288B it signed in data center leases
At its Investor Day on Oct 28, Oracle capped new money at $20B through fiscal 2030 and said new contracts will not need incremental capital from Oracle. Both promises leave out leases.
Thoma Bravo paid 8.4 times revenue for the map of who can open what
Varonis is being taken private. On Oct 15, 2026 it signed an all-cash merger with Proofpoint, a Thoma Bravo portfolio company, valued at about $6.3B and expected to close in the first half of 2027.
