S&P Global at Bernstein: "Cap IQ Is Less Than 6% of Revenue" — CEO Reframes the Bull Case Around Benchmarks and AI Data Monetization
Bernstein 42nd Annual Strategic Decisions Conference, May 27, 2026 — S&P Global CEO Martina Cheung makes the case that investors are looking at the wrong businesses
The Benchmark Reality Check
Martina Cheung opened with a framing that deserves to sit at the top of every S&P Global model. Speaking at the Bernstein 42nd Annual Strategic Decisions Conference, the President and CEO was direct: "It's a benchmarks business predominantly. It's two-thirds of our revenue, three-quarters of our profit." The corollary, which she delivered pointedly, is that Cap IQ Pro — the product that attracts the most investor anxiety around AI disruption — represents less than 6% of revenue and an even smaller share of operating margin. For investors who have been repricing the stock around competitive threats to the desktop intelligence product, Cheung's message was essentially that they are focused on the wrong line item.
The benchmarks pillar spans Ratings, the S&P Index franchise, and the Platts price assessments within Commodity Insights. Cheung described these three as deeply moated and structurally entrenched: S&P Global is the sole provider of the S&P 500 index, the sole provider of the S&P Global rating, and effectively the sole provider of the Platts Brent crude benchmark. "That is something that is unique to S&P Global and will continue, is deeply mooted, deeply entrenched in the global macro environment," she said.
AI Data Monetization Is Already Happening — 35% to 45% Pricing Uplifts at Renewal
The most substantive new disclosure from the conference was around early evidence of AI-driven pricing power in the data business. Cheung confirmed that S&P Global is already seeing clients willing to pay 35% to 45% more at contract renewal to receive the AI-ready version of data sets. This is not a theoretical future revenue stream — it is showing up in current renewal conversations, though Cheung acknowledged it remains early days.
She offered a specific example from the first quarter: a large global bank that renewed Cap IQ, expanded its use cases within the platform for the native AI capabilities, and simultaneously subscribed to AI-ready data — making S&P Global's AI-ready data the internal data standard for the bank's own AI platform. "These are the types of things that we're seeing with even the most sophisticated of our clients, which I think is a very important signal around the value that they get from S&P Global," she said.
Cheung also noted that API call volumes have increased fivefold, a metric she had flagged on the prior earnings call. Under the enterprise pricing model, the financial benefit of that usage increase flows through at renewal rather than in real time, meaning the monetization of AI agent deployments against S&P data is largely a 2026 and 2027 renewal-cycle story. That lag is intentional and structural, not a pricing failure.
The AI Threat Argument Has a Factual Problem
On the recurring concern that increasingly capable LLMs will erode the value of S&P Global's data processing and cleaning advantages, Cheung offered a direct rebuttal: the vast majority of S&P Global's IP is not publicly available. "For there to be a thesis that the value that we provide through the IP is lowered because an LLM can sift through data more quickly, you'd have to assume that the LLM has open and free access to all of that data, and that's just simply not the case," she said.
She reinforced this with an anecdote about a large investment bank that deployed a frontier model in a sandbox, believed it was functioning well, then pushed it into production and had to shut it down immediately because the outputs could not be trusted. The bank came back to S&P Global requesting an extension of the existing relationship and continued investment in S&P's own products. Trust, she argued, is not a commodity feature that can be replicated from publicly available data.
Market Intelligence Leadership Change: Reorganization, Not Instability
The day prior to the conference, S&P Global announced a leadership transition within Market Intelligence, with division head Saugata Saha departing at the end of July. Cheung framed this as an opportunity rather than a setback, and used it to announce what is arguably a more consequential structural change: the Enterprise Data Office has been shifted to report under Firdaus Bhathena, the newly appointed Chief Technology and Transformation Officer, consolidating data and technology under a single executive for the first time.
"This is really a great opportunity for us to marry the data organization with the technology organization in the sense that it's the technology that is going to help us unlock the additional value that we see across our very vast data estate across the enterprise," Cheung explained. She described the existing AI-ready data products on the market as "a handful of data sets" relative to the full breadth of S&P Global's proprietary data estate, suggesting significant runway as more data sets are brought into the data fabric and made AI-ready.
Cheung was candid that the technology landscape has shifted materially even in the past six months and that the MI leadership transition creates a useful moment to reassess investment priorities and accelerate AI integration. She committed to moving quickly on the remaining structure decisions, characterizing this as weeks rather than months.
Ratings: Conservative Guide, But Upside Is Real
On Ratings, Cheung acknowledged that the guidance for the year embeds caution around two variables: the rate environment and geopolitical risk. The structural backdrop remains compelling — approximately $8 trillion in maturities through 2028, which she described as historically elevated. Q1 billed issuance came in double-digit, driven by strong investment-grade supply and some M&A activity, and Cheung indicated the company assumed a portion of the hyperscaler-driven pull-forward occurred in Q1 rather than being spread across the year.
The nuanced point she made is that guidance was built assuming roughly half of announced hyperscaler CapEx would be debt-financed, and the company did not bake in aggressive assumptions around pull-forward from 2027 maturities. "If we didn't see a major deterioration in the rate environment, the geopolitical environment, there is a possibility that we could see some outperformance on the outlook for this year in Ratings," she said — a carefully worded but meaningful signal that the bar for a guidance raise is lower than the current numbers imply.
Private Credit: Scrutiny Is a Tailwind, Not a Headwind
Cheung pushed back on any interpretation that regulatory and LP scrutiny of private credit creates a headwind for S&P Global's ratings business in that market. The opposite is true. "Additional scrutiny is really increasing the demand for high-quality independent opinions, whether it be ratings, assessments, valuations or otherwise," she said. The private credit ratings revenue base is now in the hundreds of millions of dollars, growing over four years from a much smaller base, with demand increasingly international — she specifically cited the EU and Asia as geographies where LP interest in independent ratings for private credit has accelerated over the past twelve to eighteen months.
On competitive differentiation, Cheung made a methodological argument that is worth understanding. S&P Global applies a consistent credit methodology across public and private markets rather than a separate private credit framework. This matters practically because issuers increasingly refinance between the two markets, and because LPs managing cross-portfolio exposures need comparable analytics. "LPs need to be able to actually track their exposures using a consistent methodology," she said, adding that this point has become more appreciated — not less — over the past year as market complexity has increased.
Index: New Products in Fixed Income and Private Markets
On the Index business, which continues to run at high growth and high margins, Cheung flagged two notable first-quarter product launches: the first digital-native U.S. Treasuries index, and — in partnership with Lincoln — a first-of-its-kind private loan index series covering the U.S. and Europe. Both represent the business extending its franchise beyond its core equity index products into fixed income and private markets, which Cheung described as faster-growing areas off smaller bases. She expressed confidence in Cathy Clay's leadership of the division and highlighted the iBoxx franchise and the liquid derivatives ecosystem as areas of ongoing traction.
Mobility Spin and the RemainCo Thesis
The Mobility spin is on track for a July 1 effective date. Cheung declined to rehash the SpinCo story given its own recent investor day but used the question to articulate what RemainCo is: four divisions — Ratings, Market Intelligence, Commodity Insights, and Index — built around the benchmarks franchise, private markets adjacencies, and AI-enabled enterprise capabilities. She made one disclosure that is new relative to prior communications: the medium-term guidance issued at Investor Day did not assume meaningful AI-driven transformation. With Bhathena now in post as Chief Technology and Transformation Officer, that transformation is explicitly on the agenda, and Cheung described potential upside to both the revenue growth and margin trajectories over the three-to-five year planning horizon as transformation accelerates.
Capital Allocation: The Bar for M&A Is Very High
On capital allocation, Cheung was unusually direct about the current valuation constraint on M&A. "At this point, with the valuation the way it is, any deal, even a tuck-in size deal, quite frankly, would have to hit a really high bar for it to be a better outcome for shareholders than returning to shareholders." She ruled out transformational deals entirely, and on tuck-ins, signaled a strong preference to avoid areas with undifferentiated content — naming Market Intelligence explicitly as a segment where she would be cautious about adding inorganic exposure. The default capital return mechanism remains distributing 85% or more of adjusted free cash flow through dividends and buybacks.