Every Division Beat and the Stock Still Faded: S&P Global's AI Evidence Arrives Before the Market Will Pay for It

Results vs. Consensus

Q1 2026 Scorecard

Year-over-Year Comparison

Sequential Comparison

S&P Global's expense base is back-weighted, so the fourth quarter carries incentive-compensation true-ups that the first quarter does not. The sequential margin step is therefore seasonal rather than operational, and is shown here for calibration rather than as evidence of a trend.

Revenue. Ten percent reported growth on a $15B-plus revenue base, delivered in a quarter that contained an energy shock, a private-credit scare and a broad AI-driven repricing of the entire information-services sector, is a demonstration of what the business actually is. The revenue is contractual, the pricing is annual, and the parts most exposed to the macro (Ratings transactions, Indices asset-linked fees) are the parts that grew fastest. Subscription revenue, the piece that matters most to the disruption debate, grew 6% and was described as carrying a 50bp drag from revenue-recognition timing that reverses later in the year. That is the single most contested number in the print, and it decelerated slightly from the 6.5% to 7% band the company had run at over recent quarters. Management's answer was that annual contract value growth tracked subscription growth at around 6%, that net renewal rates improved roughly 100bps, and that pipeline built through the quarter. The evidence supports the framing, but 6% subscription growth is not a number that wins an argument on its own.

Margins. All five divisions expanded adjusted operating margin, which is the fact the company led with and the one worth interrogating. The pattern is consistent with pricing power (Ratings +150bps on 13% revenue growth, Indices +100bps on 17%) rather than pure cost action, and the enterprise result of 51.8% is an unusually high level of profitability to still be expanding. Set against that, the CFO was unusually candid that the first quarter's spending pace was a deliberate response to macro risk rather than a structural change, and the second-quarter phasing guidance confirms it. The honest read is that roughly half of the 100bps looks durable and roughly half is a timing shift into later quarters. Nothing about the year's 50 to 75bps ex-OSTTRA target changed.

EPS. The 3.3% reduction in diluted shares is doing meaningful work here, and it is going to do more. The company repurchased $1.0B in the quarter against $650M a year ago, and raised its full-year capital-return commitment from 85% of adjusted free cash flow to 100% or more, implying roughly $4.5B of repurchases. At the print-day close that is around 3.5% of market capitalisation retired in a single year, funded entirely from free cash flow and executed into a de-rated multiple. For a business that converted $5,481M of adjusted free cash flow against $5,441M of adjusted net income across full-year 2025, buying back stock at 22x forward earnings is the highest-return use of that cash available, and management said as much.

Segment Performance

Total adjusted operating profit of $2,159M is stated after adjusted corporate unallocated expense of $46M, which is why it does not equal the sum of the five divisional figures.

Revenue by Type

S&P Global's revenue mix is the crux of the disruption debate, because the subscription line is the one a large language model is theoretically capable of substituting and the other five lines are not. The split below is where the argument should be conducted.

Assessment: subscription is 48% of revenue and growing 6%. The other 52% grew 15% in aggregate and is tied to issuance volumes, index assets, derivative trading volumes and physical commodity price assessments, none of which a chatbot can produce. The bear case that treats S&P Global as a data vendor facing commoditisation is arguing about slightly under half the company.

Ratings (31% of revenue, +13%)

The quarter's engine, and the source of both the beat and the forward risk. Transaction revenue rose 15% to $712M on investment-grade strength, with management pointing specifically to hyperscaler debt financing of AI infrastructure and to several large M&A-driven issuances. Non-transaction revenue rose 11% to $590M on higher annual fees and strong CRISIL performance. Billed issuance grew 14% year over year even as bank loan volumes fell by a high-teens percentage against a difficult prior-year comparison. Private-markets ratings revenue grew more than 25%.

"Despite the turmoil in the macro environment, issuance was resilient. Billed issuance increased 14% year-over-year in the first quarter, primarily driven by strength in investment grade. Investment grade benefited from hyperscaler investments in AI infrastructure."
— Martina Cheung, President and CEO

The important qualifier came immediately after: management believes the strength was timing rather than upside, and left full-year billed-issuance expectations unchanged.

"Our full year expectations for the debt markets are largely unchanged."
— Martina Cheung, President and CEO

Assessment: a 67.7% adjusted operating margin on 13% growth is as good as this business gets, and the AI capital-expenditure cycle is turning into a multi-year issuance tailwind for the ratings duopoly in a way the market has not fully connected. But the company told investors exactly how to model it: no acceleration in Q2, moderation in Q3, negative growth in Q4. Anyone extrapolating Q1's 13% into the year is modelling against explicit guidance.

Market Intelligence (31% of revenue, +8%)

The division at the centre of the AI-disruption thesis, and the one carrying the most complicated optics. Reported growth of 8% flatters organic constant-currency growth of 6% because of the With Intelligence acquisition offset by the January divestitures of EDM and thinkFolio. Within that, Data Analytics & Insights grew 11% (six points from With Intelligence, five points organic), Enterprise Solutions grew 3% on a reported basis but 14% organically with the divestitures stripped out, and Credit & Risk Solutions grew 6%. One-time and volume-driven revenue grew 18% in aggregate. Margin expanded 70bps to 33.6%.

"Subscription growth included a 50 basis point headwind from the timing of revenue recognition that we expect to reverse in the back half of the year."
— Eric Aboaf, CFO

The Enterprise Solutions figure deserves more attention than it received. A 14% organic growth rate in the workflow-software franchise, at a moment when the market has decided workflow software is what AI eats first, is a direct rebuttal delivered without emphasis.

Assessment: Market Intelligence is not accelerating, and pretending otherwise would be dishonest. It is also not decelerating in the way a genuinely disrupted franchise would. Annual contract value growth of roughly 6%, net renewal rates up about 100bps and a building pipeline describe a business under pricing pressure at the margin but not losing its position. The reversal of the 50bp recognition headwind is a checkable commitment for the back half.

Indices (12% of revenue, +17%)

The fastest-growing division and the highest-margin asset in the portfolio at 73.8%. Asset-linked fees grew 18% on equity market appreciation and net inflows; exchange-traded derivatives revenue grew 18% on strong volumes, particularly in SPX options; data and custom subscriptions grew 12% in a third consecutive quarter of double-digit growth. Management flagged an important mix effect: in volatile periods, flows rotate from higher-priced sector, factor and thematic products into lower-priced core exposures such as the S&P 500, which reduced average realised price year over year even as asset-linked revenue grew 18%.

Assessment: the derivatives franchise is the underappreciated part of this division. It is a genuine volatility hedge sitting inside a business whose asset-linked fees are levered to equity market levels, which means Indices revenue is far more stable across market regimes than a pure beta read would suggest. The realised-price decline is worth monitoring, because a persistent rotation into core beta is a slow structural drag on a business that has been priced as a compounder of both assets and rate.

Energy (16% of revenue, +7%)

The only division guided down, and the clearest transmission of the Iran conflict into the P&L. Advisory & Transactional Services grew 15% on record CERAWeek attendance and revenue, and Global Trading Services grew close to 30% on elevated commodity volatility. Energy & Resources Data & Insights grew 7% and Price Assessments grew 6%, each carrying a sanctions-related drag of 100bps and 140bps respectively. Upstream Data & Insights declined 5% against a prior-year one-time fee. Margin still expanded 140bps to 49.4% on 4% adjusted expense growth.

"Some of the Energy customers in the Middle East have experienced a direct impact to their facilities and many are facing supply chain and/or distribution disruptions."
— Eric Aboaf, CFO

Assessment: the division is running a barbell. Volatility is a revenue tailwind for trading services and events while being a demand headwind for subscriptions, and in Q1 the tailwind won. Management does not expect that to hold, and guided second-quarter Energy revenue growth below the full-year range. The Upstream reset is a separate and longer story: the software portfolio is being divested, the remaining proprietary data business is being rebuilt around the new CERA Titan platform, and management explicitly warned it could take several more quarters before this shows up as growth.

Mobility (11% of revenue, +8%)

With the Form 10 due to file this quarter and separation targeted for mid-2026, Mobility posted a clean quarter on its way out. Dealer revenue grew 9% on CARFAX and automotiveMastermind customer additions, Manufacturing grew 5%, and Financials & Other grew 8% on underwriting volumes. Margin expanded 150bps to 40.1% on 5% adjusted expense growth.

Assessment: the operating result matters less than the separation mechanics at this point. What Q1 establishes is that the asset being spun is growing high single digits at a 40% margin in a difficult automotive OEM environment, which is a defensible standalone profile. For the remaining S&P Global, the more relevant number is the roughly $2B of debt Mobility is expected to issue in conjunction with the spin, with proceeds funding a cash payment to the parent.

Key Topics & Management Commentary

Overall Management Tone: confident on the operating results and unusually forthcoming about the parts of the beat that will not repeat, volunteering both the issuance pull-forward and the deliberate expense restraint before anyone asked about either. The bulk of the prepared remarks went to artificial intelligence, not because the numbers required it but because the share price does, and that section carried the most specific disclosure the company has yet given on how its data is being consumed and paid for. Where the tone thinned was the macro: the Iran conflict was described precisely in its Energy effects and only in scenario language everywhere else.

1. The De-Rating Is the Context for Everything Else on This Call

S&P Global entered this print down 16.3% year to date against an S&P 500 up 4.8%, a 21-point relative drawdown accumulated in under four months. The proximate cause was the February quarter, when initial 2026 adjusted EPS guidance of $19.40 to $19.65 landed below a consensus near $19.95 and the shares fell 9.7% in a single session on nearly 11 million shares. The broader cause is a sector-wide repricing of financial information businesses on the thesis that large language models erode the value of proprietary datasets and the workflow tools built on them.

Neither the CEO nor the CFO named the de-rating. Both spent most of the call answering it. The structure of the prepared remarks made the intent obvious: a detailed breakdown of Market Intelligence revenue by data category, an explicit statement that undifferentiated data is less than 5% of enterprise revenue and only 12% of Market Intelligence revenue, and an extended argument about why the curated datasets are hard to replicate.

"For our curated data, perhaps the biggest challenge in replicating some of these data sets like Compustat and SNL is the means by which we aggregated these data sets to begin with. Often, employees would have to physically scan and paper documents in local offices. While some of that data may be publicly available, many of these types of data sets are only available in digital formats from S&P Global."
— Martina Cheung, President and CEO

Assessment: the argument is correct and it is also the argument every incumbent makes. What separates this call from a defensive one is that the company brought customer behaviour rather than assertion, which is the subject of the next section. The disclosure choice itself is the tell: a management team that believed the market had simply misread the quarter would have led with the quarter.

2. The Monetization Evidence: Renewal Uplifts, Price Premiums and Retention

This is the most important new information in the print, and it arrived in fragments across the prepared remarks and Q&A rather than as a single disclosure. Assembled, it describes customers paying measurably more, and leaving measurably less, because of AI access rather than despite it.

"For instance, in the first quarter, the volume of API calls made by our customers was more than 5x the volume that we saw just 1 quarter ago. Volumes doubled month-over-month just from February to March."
— Martina Cheung, President and CEO

Volume alone proves consumption, not economics. The economics came in three specific data points. Two financial clients renewing subscriptions accepted increases in the range of 35% to 45% purely to receive the same data in an AI-ready format. A buy-side client working with Kensho cancelled an incumbent provider and moved to S&P Global at roughly a 20% higher price. And usage is showing up in retention.

"In the clients that have been using our AI tools and availing themselves of those in MI, we're seeing a couple of hundred basis points higher retention rates. In Energy, over 500 basis points of higher retention rates because, again, usage is value for clients."
— Eric Aboaf, CFO

On top of that, contract value growth among AI-using customers runs 30% above the rest of the Market Intelligence base and roughly double the rest of the Energy base, and the Kensho LLM-ready API programme has gone from about 150 customers interacting with the data in March to more than 300 under contract or in trial.

Assessment: the bear thesis holds that AI commoditises the data and compresses the price. The observed behaviour in this quarter is the opposite in every measured dimension: higher price, higher retention, higher contract value growth. That is one quarter and a small sample of named deals, and management was careful to say the aggregate financial effect will take time to appear. But it is the first hard evidence either side has produced, and it points the other way from the share price.

3. Distribution Strategy: The Desktop, the Plug-In and the Frontier Model

The strategic question underneath the disruption debate is whether S&P Global gets disintermediated when its customers stop opening Capital IQ Pro and start asking a model instead. Management's answer is that it intends to be paid in all three channels, and it is building for each: the integrated desktop, an S&P Global plug-in launched alongside the Claude for Financial Services announcement in the quarter, and open access to data through Model Context Protocol and agent-to-agent protocols.

"Our customers are on their own AI journeys, and adopting these new platforms in different ways, depending on urgency, comfort level and regulatory sensitivity. We will continue to invest in new ways to create value for our customers, including delivery through MCP and agent agent protocol, to ensure that customers can access our data and tools where they need it."
— Martina Cheung, President and CEO

The most useful illustration was a large global bank that in the quarter simultaneously extended its Capital IQ Pro desktop seats and expanded AI licensing of several datasets, then made S&P Global data the standard on its own internal model. That is a customer buying both channels rather than substituting one for the other, which is the pattern management says dominates its conversations today.

Assessment: the strategy is right and the risk is real anyway. Being present in the frontier-model channel protects the data but not necessarily the workflow layer, and the workflow layer is where a meaningful share of Market Intelligence pricing sits. Management's own framing acknowledges a scenario where clients move entirely to in-house models; the claim is that S&P Global's data becomes more valuable in that world, not less. That claim is untested and it is the crux of the multi-year debate.

4. Pricing Philosophy: Enterprise Value, Not Seats and Not Usage

Asked directly about the monetization model across distribution channels, the CEO drew a line the market has largely ignored. S&P Global does not license by seat and does not price on consumption alone, which structurally decouples its revenue from the question of whether a human or an agent is doing the reading.

"The point I would make on monetization is that we are really thinking about monetization through the lens of enterprise value. So as you know, we don't do seat-based licensing. We don't do usage only. We track usage, channels, the value we create and a number of other metrics as part of the discussions that we have with our clients on value and price accordingly."
— Martina Cheung, President and CEO

Assessment: this deserves more weight than it got. A meaningful part of the bear case for information services rests on seat compression as agents replace analysts. A firm that never sold seats does not face that mechanism directly, and an enterprise-value contract structure gives it a path to capture the productivity gain rather than lose to it. The counter is that enterprise-value pricing still requires the customer to agree the value went up, and it is slower to reprice than a usage meter would be.

5. The Ratings Pull-Forward and the Fourth-Quarter Cliff

Ratings produced the quarter's headline growth and the year's clearest identified risk in the same breath. Billed issuance grew 14%, driven by investment-grade volume that management attributes substantially to hyperscalers financing AI infrastructure, plus several large M&A transactions. Rather than raise the full-year issuance outlook, management said the strength was calendar timing against a guidance assumption that had spread hyperscaler issuance more evenly.

"Given both the outperformance in the first quarter and the more modest expectations for Q2, we do not expect to see acceleration in Ratings revenue growth in the second quarter. We continue to expect Ratings growth to moderate in the third quarter before turning negative in the fourth quarter as we lap prior year highs."
— Eric Aboaf, CFO

Assessment: volunteering a negative fourth quarter in April is credibility-building and it is also a warning to anyone building a model off the first-quarter run rate. The deeper point for the multi-year case is that the AI capital-expenditure cycle is arriving in the P&L as debt-financed issuance, which the ratings duopoly monetises with no incremental capital. That is an AI beneficiary channel the market is not crediting to this stock while it debates whether AI destroys the other half of it.

6. Expense Restraint as a Deliberate Response to Macro Risk

Every division expanded margin, and the CFO attributed a material share of that to a decision taken in late February rather than to structural improvement. When the Iran conflict began, the company slowed discretionary spending across the enterprise, which produced better first-quarter margins in every division than the initial guidance had assumed.

"On second quarter margins, we expect margin expansion to be above the enterprise full year range for Ratings and Indices, slightly below the range for Mobility and Energy and within the range for Market Intelligence. This is largely due to the timing and quarterly phasing of expense recognition as we were very disciplined in our approach in the first quarter."
— Eric Aboaf, CFO

Assessment: this is the single most important caveat on the beat, and management put it on the record without being pushed. The first quarter borrowed margin from the second. Full-year divisional expectations are unchanged, which means the 100bps of enterprise expansion in Q1 should not be annualised. What it does demonstrate is a management team with fast levers, which is worth something in a year whose macro assumptions include a conflict resolving by the end of the second quarter.

7. Energy Under the Iran Shock, and the Upstream Reset

The Energy division is where the geopolitical shock is landing, and it is the only place guidance moved down. Full-year organic constant-currency growth for the division was cut to 4.5% to 6.0% from 5.5% to 7.0%, with the guidance assuming elevated disruption persists through the second quarter and supply chains take until later in the year to normalise.

"The conflict in Iran has led to the largest energy shock since the 1970s and counterbalance what was previously a broadly favorable economic environment for business."
— Eric Aboaf, CFO

Running alongside the macro problem is a self-inflicted restructuring that is finally producing artefacts. The Upstream software portfolio is being sold, sized by management at roughly 25% of Upstream revenue, with a distribution partnership attached to the buyer. The remaining 75% is proprietary subsurface, geoscience and benchmarking content, now fronted by a new AI-native platform called CERA Titan that was soft-launched to about 70 customers at CERAWeek.

Assessment: the guidance cut is small in enterprise terms, roughly 15bps on consolidated organic growth given Energy is 16% of revenue, and it is the correct response to a real shock. The Upstream story is the one worth tracking, because it is the only division-level growth problem S&P Global owns rather than inherits from the macro. Management guided to several more quarters before it turns, which is honest and also the third consecutive framing of Upstream as a future fix.

8. The Mobility Separation and the Debt Attached to It

The spin is on schedule for mid-2026. The Form 10 files this quarter, the Mobility Global investor day is set for 12 May in New York, an equity roadshow follows, and a public debt offering targeting an investment-grade rating is planned within the quarter. S&P Global will continue to fully consolidate Mobility in reported financials and in 2026 guidance until separation completes, then recast the four quarters of 2025 and any 2026 periods and reissue guidance excluding Mobility.

"We expect to issue approximately $2 billion in debt at Mobility in conjunction with the spin. Proceeds are expected to fund a cash payment to S&P Global, which we would expect to use for a combination of incremental share repurchases and some debt reduction."
— Eric Aboaf, CFO

The leverage arithmetic was given explicitly: gross leverage sits at 2.3x trailing-12-month EBITDA against a 2.0x to 2.5x target, and the loss of Mobility EBITDA alone would push that to 2.4x by year end. The $2B dividend from Mobility is what keeps the balance sheet inside the range while still funding buyback.

Assessment: the mechanics are clean and the cash payment is the piece investors should be modelling, because it is incremental repurchase capacity landing into a de-rated multiple. The cost is six months of estimate noise: recast financials, stranded costs the company has explicitly excluded from current guidance, and a guidance reissue mid-year. Stranded costs are the number nobody has yet quantified.

9. Capital Allocation: The Buyback Goes to 100% of Free Cash Flow

The capital-return commitment was raised from 85% of adjusted free cash flow to 100% or more, implying roughly $4.5 billion of repurchases for the year, on top of a quarterly dividend of $0.97 that continues a 53-year run of increases. The company bought $1.0 billion of stock in the first quarter against $650 million a year earlier.

"Given the strength and resilience of our business and our confidence in its long-term profitable growth, we believe the current share price reflects an attractive opportunity to increase our repurchases from the expected 85% of adjusted free cash flow to at least 100% or to roughly $4.5 billion for the year."
— Eric Aboaf, CFO

Assessment: an explicit statement that the board considers its own stock cheap, delivered with a raised number attached rather than as a sentiment. At the print-day close, $4.5 billion retires roughly 3.5% of the shares outstanding in a single year, and it compounds against a share count already down 3.3% year over year. This is the clearest signal in the quarter of what management thinks the de-rating is worth.

10. Private Credit: A Growth Engine Meeting a Stressed Cycle

Private markets revenue crossed $600 million at the enterprise level in 2025 and Ratings private-credit revenue grew more than 25% in the quarter. The CEO also opened the call by naming private credit as a sector under pressure, citing increased scrutiny, wider spreads and elevated redemptions.

"So it's an area that we are, I would say, cautiously optimistic about over the very immediate time frame just given some of the stresses on the sector that we mentioned. But we started this year with those potential stresses in mind. We didn't necessarily assume there was going to be huge growth in middle market CLOs, for example, we assumed that there would be some softness in BDCs."
— Martina Cheung, President and CEO

Assessment: the guidance was built with the stress assumed, which is the right answer and one that can be checked against results as the year progresses. The more interesting strategic point is that a stressed private-credit cycle increases demand for exactly what S&P Global sells into it: transparency, benchmarks and third-party performance data. The Cambridge Associates and Mercer partnership released its first tranche of private credit and infrastructure data in the quarter, and With Intelligence documents were integrated into Capital IQ Pro. A disorderly private-credit cycle would hurt the ratings revenue and help the data franchise.

Guidance & Outlook

Full-Year 2026 (adjusted unless noted as GAAP)

The four moving pieces are all small and all explained. Reported revenue growth came down 30bps on reduced currency tailwind with organic constant-currency growth untouched, which is a translation effect and not a demand signal. Net interest expense rose $10 million on the additional repurchase activity. The capital-return commitment went from 85% or more of adjusted free cash flow to 100% or more. Energy is the only division whose underlying outlook changed, and the enterprise arithmetic on a 100bp cut to a division representing 16% of revenue is roughly 15bps at the consolidated line. Every line that sets the earnings algorithm itself, organic growth, EPS, margin, tax and amortisation, was left alone.

Guidance is still constructed on the assumption that Mobility contributes for the full year and excludes any stranded costs, with a full reset promised on completion of the spin. The macro assumptions behind it were disclosed in unusual detail: global GDP growth of 3.2% including 2.2% in the United States, US CPI of 3.2%, one Federal Reserve rate cut, and stabilisation of the Iran conflict by the end of the second quarter.

Implied ramp. Delivering the guidance midpoint of $19.53 after a $4.97 first quarter requires roughly $14.56 across the remaining three quarters against $13.46 in the comparable 2025 period, or about 8% growth. The company grew adjusted EPS 14% in Q1. The guide therefore embeds a six-point deceleration for the balance of the year, which is exactly what the Ratings cadence commentary describes: no acceleration in Q2, moderation in Q3, negative growth in Q4. On the revenue line the same pattern holds. The 6.3% to 8.3% reported range implies 4.9% to 7.6% growth over the final three quarters against 10% delivered in the first.

Street at. Consensus for full-year adjusted EPS sat at $19.51 going into the print, effectively on top of the $19.53 midpoint. The Street had already capitulated to the February guide rather than modelling above it, which means the unchanged guidance produces no negative revision but also gives estimate momentum nothing to work with.

Guidance style. Conservative, and consistently so. This management team set its initial 2026 outlook in February below consensus and absorbed a 9.7% single-day decline for it. Three months later it beat its own first quarter by roughly $0.15, cut one division, and left the full-year EPS range untouched rather than flowing the beat through. The pattern says the guide is a floor built to be cleared rather than a target built to be hit, which is the right posture for a business whose second-half comparisons are the hardest of the year. It also means the first raise, if it comes, arrives with the post-spin reset in the third quarter.

Analyst Q&A Highlights

Sixteen questions, and the distribution of them was the story. Six went directly to artificial intelligence: the partnership strategy, the value of proprietary data inside workflow tools, the efficiency contribution to margin, the go-to-market model for the new Upstream platform, the size of the AI benefit inside the margin target, and the split between direct and third-party consumption. Three went to the credit franchise and two to Market Intelligence contract value. Nobody asked about the Mobility separation mechanics, and nobody pressed the Energy guidance reduction directly. The Street is no longer arguing about whether the quarter was good. It is arguing about whether the franchise survives the decade.

Partnering With the Frontier-Model Vendors, and What the Data Is Worth in That Channel

The opening question set the agenda for the entire session, asking whether the company intends to build applications on the large model platforms or simply pipe data into them, and what the economics look like across channels. The answer confirmed that purpose-built applications are the intent, described the S&P Global plug-in launched during the quarter, and then delivered the quarter's single most concrete monetization datapoint without being asked for one.

Q: "I was hoping that you could expand on how you're thinking about the partnership strategy with the large AI players? Are you building SMP, MCP apps on the platforms? Or you just plan to continue to provide the data through the MCP integrations and the APIs? And maybe if you could just talk about the monetization model and directional economics between the different distribution channels."
— Toni Kaplan, Morgan Stanley

A: "Perhaps maybe one other example I would provide is in the quarter two financial clients who are just subscribing to our data at renewal, were opting to get that data available in an AI-ready format. And were willing to pay in the range of 35% to 45% on the renewal increase to get the AI access. So again, early days, but some very strong signal here around the monetization from an enterprise value standpoint."
— Martina Cheung, President and CEO

Assessment: a 35% to 45% renewal uplift for a format change on data the customer already licensed is not a data point a disrupted franchise produces. Two contracts is a sample, not a trend, and management flagged it as early. But the direction is unambiguous and it is the number to track quarter by quarter, because the entire re-rating case rests on whether it generalises.

Whether the Margin Expansion Is AI Efficiency or Expense Restraint

A direct attempt to separate structural productivity from cyclical cost control in the Market Intelligence margin. The answer led with the cyclical explanation rather than the structural one, which was the honest ordering and also the less flattering one.

Q: "On the Market Intelligence margin, just wondering if you can maybe help contextualize how much of the margin expansion that you're seeing is being driven by efficiency gains associated with AI."
— Scott Wurtzel, Wolfe Research

A: "Margin expansion has come in nicely in MI in particular, in first quarter. We were careful with the external environment. Starting late February, we -- the Iran conflict started. We're careful about our discretionary spending. And so you saw particularly strong performance in MI as well as our other 4 divisions as we just carefully thought about pacing expenses through the year."
— Eric Aboaf, CFO

Assessment: the question was about AI and the first sentence of the answer was about discretionary spending. That ordering is the most reliable signal in the exchange. AI-driven data operations and software development were both named as contributors afterwards, but the quarter's margin beat was principally a decision about pacing, and the phasing guidance for the second quarter confirms it.

How Much of the Ratings Quarter Was Borrowed

The only sustained line of questioning on the credit franchise, prompted by the unusual strength in non-transaction revenue. Management answered the mechanical question on annual fees and CRISIL, then volunteered the more consequential point about the transaction line unprompted.

Q: "Just if we go through, I guess, how to think about the balance of transaction, non-transaction growth within the Ratings business for the rest of the year? And I guess just for the first quarter, what drove a pretty notable spike in the non-transaction numbers."
— Curtis Nagle, Bank of America

A: "As we looked at the amount of hyperscale issuance in Q1, we believe that there was some pull forward there relative to our expectations for hyperscale issuance. And this is one of the reasons why we are continuing to maintain our expectations for billed issuance for the full year."
— Martina Cheung, President and CEO

Assessment: a management team optimising for the next print takes the issuance beat and raises the issuance outlook. This one took the beat and explicitly told investors not to extrapolate it. The credibility that buys is worth more over a full cycle than the quarter it costs, and it is a useful data point on how the February guide should be read.

Sizing the Energy Software Divestiture

The most practically useful modelling question of the call, and it produced a number the press release did not contain. The Upstream software portfolio being sold represents about a quarter of Upstream revenue, and the divestiture comes with a distribution agreement attached to the acquirer rather than a clean exit.

Q: "I think there were like 7 or 8 different brands, I think you're selling in Energy. I was just hoping you could help us size that for our models? Like how much are you getting selling to SLB."
— Manav Patnaik, Barclays

A: "Maybe to start, the size of that is about 25% of Upstream revenues. And that software portfolio, as you mentioned, is actually quite varied and quite distinct. And so one of the reasons that really informed our decision there is that we think SLB is a very good partner on that. And as part of that decision to divest, we also have a new distribution partnership with SLB that we are quite excited about as we close that."
— Martina Cheung, President and CEO

Assessment: the number was given, the sale price was not, and neither was the revenue dollar figure for Upstream overall. The distribution partnership is the part that changes the character of the transaction: this is a channel restructuring rather than a retreat, and it leaves S&P Global with the proprietary content and someone else carrying the software cost base.

When AI Efficiency Actually Reaches the Margin Target

A pointed attempt to quantify how much of the 50 to 75bps ex-OSTTRA margin target for 2026 is attributable to AI. The company declined to give a figure and instead pushed the material benefit out by two years, which is a more useful answer than a number would have been.

Q: "And also, Eric, I wanted to ask your 50 to 75 basis points expected improvement, excluding OSTTRA, how much ballpark do you think AI efficiencies is actually helping that number?"
— Craig Huber, Huber Research Partners

A: "But I think the upside from the broad adoption of Frontier models is just beginning. And really will have an impact in '27, '28 and in the future years as they get expanded into a wide range of these strategic and important processes that we operate and will be helpful in that regard."
— Eric Aboaf, CFO

Assessment: management characterised AI's current margin contribution as "just beginning" and placed the real benefit in 2027 and 2028. That is a refusal to take credit today for something that has not yet arrived, and it means the 2026 margin guidance stands on conventional productivity rather than on a technology narrative. It also implies a second, later leg of margin expansion that no current estimate contains.

Direct Platform Consumption Versus the Frontier-Model Channel

The most forward-looking question of the call asked where the balance between S&P Global's own delivery and third-party model consumption ultimately settles. The answer described both channels growing at once, with quantified usage on each side.

Q: "Are you seeing any meaningful differences in usage patterns or engagement with your data across those two broad channels today? And I guess I'm wondering, as adoption scales, where do you see the balance between direct delivery through your own solutions and third-party large language models ultimately settling out?"
— David Motemaden, Evercore

A: "On the direct usage side, right, where clients are using our platforms and within our platforms, usage continues to build very substantially. I described in our Energy core platform, AI queries are up 2x in iLEVEL, the automated data ingestion through AI is up 2x. And so seeing very significant increases, which we're monitoring in our minds, that's the way clients are gaining value."
— Eric Aboaf, CFO

Assessment: the case was made through a large global bank that in the same quarter expanded desktop seats and licensed datasets for its internal model, which is the complement pattern rather than the substitution pattern. Management did not claim the substitution scenario is impossible, only that it would make the data more valuable rather than less. That distinction is the whole multi-year debate, and it will not be settled by a quarter of usage statistics.

Whether Annual Contract Value Growth Decelerated

The sharpest pushback of the session, and the only exchange where an analyst pressed a specific number against the company's own recent disclosure history. Annual contract value growth of roughly 6% sits at the bottom of the range the company has been quoting.

Q: "Can you just clarify on the ACV growth? I think you said that it was 6% in the quarter. I believe the past couple of quarters is 6.5% to 7%. So did it decelerate? And if so, what drove that? Because the commentary on revenue side optimistic for the rest of the year."
— Jason Haas, Wells Fargo

A: "I said the ACV growth was in line with subscription revenue growth, which was around 6%. I think we've quoted over the last 5 quarters, 6% to 6.5%, 6.5% to upper 6s percent. So it's in the range. There's always going to be a little bit of volatility. But what we see is that the underlying drivers are moving in the right direction. We're feeling good about net sales, net renewals and so forth across MI."
— Eric Aboaf, CFO

Assessment: "it's in the range" is doing a lot of work when the range as described runs from 6% to the upper 6s and the print is 6%. The forward indicators cited (net renewal rates up about 100bps, larger average deal size, a building pipeline) are real and checkable, and the 50bp revenue-recognition headwind reverses in the second half. But this is the metric the disruption thesis will be argued over, and a company answering "in the range" rather than "accelerating" has conceded the near-term point.

What They're NOT Saying

1. The size of the AI revenue, in dollars. The call produced retention deltas, contract-value growth differentials, renewal-uplift percentages, customer counts and API call multiples. It produced no revenue figure for AI-linked products, no attach rate expressed in dollars, and no disclosure of what share of the installed base has repriced. Management explicitly said the financial manifestation will take time, which is fair. It also means the strongest counter-argument to the de-rating cannot yet be checked against the income statement.
2. Stranded costs from the Mobility separation. Guidance assumes Mobility contributes for the full year and explicitly excludes any impact from anticipated stranded costs. That is a disclosed exclusion rather than a hidden one, but the number has not been sized on any call, and it lands in the post-spin guidance reset alongside recast financials. It is the largest unquantified item in the 2026 P&L.
3. What the Upstream software portfolio sold for. Management sized the divested business at roughly 25% of Upstream revenue and named the counterparty. It did not give a purchase price, a dollar revenue figure for Upstream, or the margin profile of what is being sold versus what is being kept. Without those, the dilution from the transaction cannot be modelled.
4. Whether the Ratings fourth-quarter decline is a decline in revenue or in growth. The phrasing was that Ratings growth turns negative in the fourth quarter as prior-year highs are lapped. That reads as a year-over-year revenue decline, but the magnitude was never framed and it was not pressed in Q&A. On a division representing 31% of revenue at a 67.7% adjusted margin, the difference between a low-single-digit decline and something larger is the difference between the guidance holding and not.
5. Any quantification of what a protracted Iran conflict does to the guide. The macro assumptions were laid out precisely, including stabilisation by the end of the second quarter. The downside scenario was described qualitatively as "more significant direct headwinds" in Energy and "significant indirect headwinds" in the market-sensitive businesses, with no sensitivity attached to either. Given the guide is built on a resolution that has not happened, the absence of a bear-case range is the most conspicuous omission in the outlook.
6. Average realised price in Indices, as a number. Management disclosed that mix rotation into lower-priced core products drove a modest year-over-year decline in average realised price on asset-linked fees, and quantified none of it. In a division growing 18% on asset-linked fees at a 73.8% margin, the price component of that growth is the part with structural risk attached, and it is the part that was described adjectivally.
7. Whether the buyback increase reflects conviction or a leverage constraint. The stated logic is that the share price is attractive. The same paragraph disclosed that Mobility's departure pushes gross leverage from 2.3x to 2.4x and that roughly $2 billion of Mobility debt proceeds are coming back to the parent. Both explanations fit the facts, and management only offered the flattering one.

Market Reaction

• Pre-print setup: the shares closed at $437.22 on 27 April, down 16.3% year to date against an S&P 500 up 4.8%, down 8.8% over the trailing twelve months and up 7.6% over the trailing thirty days. The 52-week closing range entering the print was $390.76 to $564.15, so the stock came in roughly 12% above its low and 23% below its high. The low was set on 11 February, the day after the fourth-quarter print.
• Print-day session (28 April, before the open): opened at $442.15, up 1.1%, traded as high as $453.95 (+3.8%) and as low as $433.09 (-0.9%), and closed at $433.47, down 0.9% or $3.75. Premarket indications had the stock up roughly 1.1% to 1.3%.
• Volume: 2.5 million shares against a 30-day average of 1.9 million, or 1.3 times normal. Elevated, but not the kind of volume that accompanies a positioning reset.
• Context: the S&P 500 fell 0.5% on the session. Among the closest information-services comparables, Moody's fell 0.6%, MSCI rose 0.6% and FactSet rose 1.9%.

The full round trip is the whole story. The stock opened up, rallied nearly 4% into the morning, and closed at the low end of its range for a 0.9% loss. That is not a market rejecting the numbers, which beat on both lines. It is a market that bought the print, listened to the call, and sold it back.

Three things on the call plausibly caused the fade, and they arrived in sequence. The Ratings cadence commentary told anyone modelling off the first quarter to stop, and it did so while confirming a negative fourth quarter. The second-quarter margin phasing commentary retracted a meaningful share of the margin beat before the market had finished pricing it. And the Energy guidance cut, small as it is in enterprise terms, was the first divisional reduction of the year and arrived with a macro assumption attached that the company cannot control. Taken together, the call converted a two-line beat into a quarter with no incremental full-year earnings power, and the shares repriced accordingly.

The peer comparison is the more revealing detail. Moody's, MSCI and FactSet did not move materially on the day, which means this was not a sector event and not a read-through. It was specific to what S&P Global said about its own second half. The contrast with 10 February is instructive: that session took the whole complex down together on a guidance reset, and this one did not travel at all.

Street Perspective

Debate: Does AI erode the data moat or widen it?

Bull view: the bull case being made on the Street is that proprietary and contributory datasets become more valuable, not less, when every institution is building agents that need trustworthy inputs, and that S&P Global is uniquely positioned because it owns benchmarks and standards rather than merely aggregating public information. This quarter's renewal uplifts and retention deltas are cited as the first tangible confirmation.

Bear view: the bear camp contends that the desktop terminal is the product actually at risk, that a customer able to query its own model against licensed data will eventually stop paying for the interface layer, and that a 6% subscription growth rate at a company charging premium prices is already the early evidence of that compression. On this view the price uplifts are a transitional artefact of scarcity in AI-ready formats that competitors will erase.

Our take: the bulls have the better of it for now, on evidence rather than argument. The measured behaviour in this quarter runs in the opposite direction to the bear thesis on every observable dimension: price, retention and contract-value growth all higher among AI-using customers. The bears are right that the sample is small and that the interface layer is genuinely exposed. But roughly half of S&P Global's revenue has no interface layer at all, and that half grew 15%.

Debate: Is the Ratings cycle peaking just as the market needs it to carry the year?

Bull view: a growing consensus view holds that hyperscaler and utility debt financing of AI infrastructure represents a multi-year issuance supercycle that the ratings duopoly monetises with essentially no incremental capital, and that the maturity wall entering 2026 plus reviving M&A provide a second and third leg beneath it.

Bear view: some sell-side desks argue the reverse, that the first quarter borrowed from the rest of the year, that the fourth quarter is guided to an outright revenue decline, and that credit spreads widening on private-credit stress and geopolitical risk is the classic precursor to an issuance air pocket. On this view Ratings has already delivered its best quarter of 2026.

Our take: both are right on different horizons, and the company effectively conceded the bear case for this calendar year by refusing to raise the issuance outlook. The fourth quarter will be ugly on a reported basis and the market knows it. What the bears are underweighting is that the AI capital-expenditure cycle is a genuinely new and durable source of investment-grade supply, and that it reaches this business as high-margin transaction revenue rather than as a cost.

Debate: Does the Mobility separation create value or just complexity?

Bull view: the argument for the spin is that a high-single-digit-growth automotive data business at a 40% margin has been valued inside a conglomerate at the conglomerate's discount, that separation forces the market to price the remaining benchmark and ratings franchise on its own merits, and that roughly $2 billion of debt proceeds returning to the parent funds incremental repurchase at a de-rated share price.

Bear view: the sceptical case is that spin-offs of good assets remove growth and diversification from the remaining entity, that stranded costs have not been quantified, and that six months of recast financials and a mid-year guidance reset destroy estimate visibility precisely when the stock needs a clean narrative.

Our take: the bull argument is stronger on the arithmetic and the bear argument is stronger on the timing. The separation is the right structural decision and the cash payment is real, incremental and arrives at an attractive repurchase price. But the stranded-cost figure is the largest unquantified item in the 2026 outlook, and a company already fighting a narrative problem is about to spend two quarters explaining what its financials mean.

Model Update Needed

This is an initiation, so the table below sets our starting assumptions against the company's guidance and where consensus sits rather than revising a prior model.

Valuation. At the 28 April close of $433.47 the shares trade at 22.2x the midpoint of unchanged full-year guidance and 24.3x trailing 2025 adjusted EPS of $17.83. Entering the year, on the then-prevailing consensus, the multiple was closer to 26x. Nothing in the operating results explains the compression; the February guidance reset and the sector-wide AI de-rating explain all of it. Applying a 26x multiple to our $19.60 estimate produces a 12-month target of $510, which implies +17.7% from the 28 April close, before a dividend yielding roughly 0.9%. That target requires no acceleration, no estimate revision and no resolution of the AI debate. It requires only that the multiple stop contracting.

Thesis Scorecard Post-Earnings

This is our first published quarter on S&P Global, so the table below establishes the thesis pillars rather than grading a standing one. Each will be carried forward and scored against subsequent quarters.

Overall: thesis established at initiation, and the quarter supports the constructive side of it. Every pillar of the bull case found evidence in the print or on the call, and the two bear points that carry real weight are a growth rate that did not improve and a calendar that gets harder. Neither is a challenge to the franchise. Both are challenges to the next two prints.

Action: buy. The case does not require the company to beat, and this quarter demonstrated that beating does not currently help. It requires the market to stop applying a disruption discount to a business whose own customers are paying more for AI access than they were paying for human access. At 22.2x unchanged guidance, with 3.5% of the shares being retired annually and a separation catalyst arriving mid-year, the risk of being early is a full multiple and the risk of being absent is the entire re-rating.