S&P Global is one of those businesses that sits behind markets rather than in front of them.

Most investors never think about the company when a corporation issues debt, an exchange-traded fund tracks an index, an analyst opens a data platform, or an energy trader references a benchmark price.

But S&P Global is often somewhere in the workflow.

That is what interests me.

The company has spent decades becoming embedded in decisions that financial institutions, corporations, governments, and investors make every day.

The products are different, but the underlying economics have a lot in common: trusted data, benchmarks, regulatory relevance, recurring workflows, and very little physical capital required to serve the next customer.

S&P Global completed the separation of Mobility Global on July 1, 2026.

The remaining company is now centered on four divisions:

I like the cleaner structure.

Mobility was a good business, but the remaining S&P Global is even more concentrated around financial infrastructure, benchmarks, data, and workflow tools.

In the second quarter of 2026, pro forma revenue excluding Mobility increased 11% to $3.68 billion. Adjusted operating margin reached 54.3%, up 200 basis points from the prior year, and adjusted earnings per share increased 23% to $4.83.

Ratings and Indices both reported record results.

That is a strong quarter, but the more important question is why these businesses are so difficult to displace.

Credit ratings are probably the clearest example of S&P Global’s moat.

When a company issues debt, investors need a way to evaluate credit quality. Regulators, investment mandates, risk systems, and internal processes also rely on ratings.

S&P Global Ratings does not simply sell an opinion.

It sells an opinion that the market already recognizes.

A new competitor could hire talented analysts and build a credit model. What it cannot easily recreate is the history, regulatory recognition, issuer relationships, investor acceptance, surveillance infrastructure, and trust that S&P Global has accumulated over decades.

In the second quarter of 2026, Ratings revenue increased 17%.

For the first six months of the year, Ratings revenue increased 15%.

Some of that business is tied to issuance activity, which makes it cyclical. Strong debt markets help. Weak issuance can hurt transaction revenue.

But Ratings also earns non-transaction revenue from surveillance, entity credit ratings, research, and relationship-based pricing programs.

That recurring piece makes the business more durable than simply betting on the next quarter’s bond issuance.

I do not view Ratings as perfectly predictable. Issuance will move around. What matters to me is that when companies return to the debt market, S&P Global remains one of the firms they already know investors will recognize.

Indices is the part of S&P Global I find easiest to admire.

S&P Dow Jones Indices owns and licenses benchmarks that sit underneath trillions of dollars of investment products.

When assets flow into products tracking the S&P 500 or other licensed indices, S&P Global can earn asset-linked fees.

The company can also earn usage-based royalties and subscription revenue.

The beauty is how little capital the model requires.

S&P Global does not have to manufacture a new physical product when another dollar flows into an index-linked fund.

The benchmark already exists.

That creates unusually scalable economics.

In the second quarter of 2026, Indices revenue increased 20%.

Asset-linked fees increased 22%, and sales usage-based royalties also increased 22%.

Those numbers will move with markets and investment flows, so I would not extrapolate a 20% growth rate forever.

What I like is the toll-like structure: once assets gather around a benchmark, S&P Global can participate without needing to put comparable new capital into the business.

Market Intelligence is less visible but important.

Its products include Capital IQ, data feeds, analytics, valuation services, research, private-markets data, and software integrated into financial workflows.

That makes the moat different from Ratings or Indices.

The strength here is not one famous benchmark.

It is workflow integration.

Once a bank, asset manager, private equity firm, corporation, or regulator builds data and processes around a platform, switching is not frictionless.

The customer has trained employees, connected data feeds, integrated software, built models, and created habits around the product.

That does not make switching impossible.

It makes switching costly enough that the incumbent gets a meaningful advantage.

Market Intelligence adjusted revenue increased 6% in the second quarter and 7% for the first half of 2026.

That is not spectacular growth.

But recurring subscription revenue, pricing, product expansion, private markets, and deeper workflow integration can make mid-single-digit growth valuable when margins are strong and retention is high.

This is also the division where I would watch artificial intelligence most closely.

AI is the risk I take most seriously in Market Intelligence.

I think it attacks the interface before it attacks the underlying asset.

A model can summarize public information quickly. It cannot easily recreate decades of proprietary datasets, ratings histories, benchmark governance, pricing methodologies, permissions, and workflow integrations.

If AI makes those assets easier to search and use, S&P Global could become more valuable to customers rather than less.

What would worry me is a different outcome: customers deciding they can get 80% of the value from commodity data plus a general-purpose model at a fraction of the price.

That is why I would watch renewal rates, pricing, and usage more closely than headlines about the newest model. If those weaken materially, the moat is changing.

S&P Global is investing aggressively in AI products and acquisitions. I think that is the right response, but I do not assume incumbency guarantees the outcome.

The Energy division includes commodity benchmarks, market data, pricing, research, and workflow products.

Platts is the best-known brand.

The economics resemble the rest of S&P Global more than they resemble an energy producer.

S&P Global is not drilling wells or building pipelines.

It provides information and price assessments used by people who do.

Energy adjusted revenue increased 3% in the second quarter and 5% during the first half of 2026.

That is slower than Ratings and Indices.

I am fine with that if the business remains durable, cash generative, and strategically useful.

Not every division needs to grow 15%.

The question is whether each one strengthens the overall infrastructure franchise and earns attractive returns on the capital committed to it.

S&P Global generated $15.3 billion of revenue in 2025 and $5.1 billion of free cash flow.

The business required only about $195 million of capital expenditures that year.

That is an unusual relationship between scale and physical capital needs.

The company can invest heavily in people, data, technology, acquisitions, and product development without needing factories, stores, or enormous inventories.

That leaves substantial cash available for acquisitions, dividends, debt reduction, and share repurchases.

The historical chart includes Mobility, which was spun off in July 2026, so I would not treat it as a clean picture of the current company.

But it illustrates the broader point.

S&P Global has historically converted a meaningful amount of revenue into cash while requiring relatively little physical capital.

The point is not simply that S&P Global produces a lot of cash. It is that growth does not require the company to consume most of that cash just to stand still.

In 2025, S&P Global returned $6.2 billion to shareholders through $5.0 billion of share repurchases and roughly $1.2 billion of dividends.

That exceeded reported free cash flow for the year.

In 2026, the company has become even more aggressive.

Through the second quarter, it had repurchased $1.5 billion of stock, and management said it expected more than $7 billion of total repurchases during the year.

I like a shrinking share count when the price is attractive.

I like it less when a company treats repurchases as something that must happen regardless of valuation.

S&P Global’s diluted share count declined 3% year over year in the second quarter, which helped adjusted earnings per share grow faster than adjusted net income.

That is useful for continuing shareholders.

But the purchase price still matters.

A great business can make a mediocre capital allocation decision if it buys its own shares too aggressively at an expensive valuation.

The separation of Mobility is strategically important.

S&P Global now has four divisions that fit together more naturally around financial and commodity infrastructure.

The remaining company is simpler, more asset-light, and more concentrated in the franchises I find most attractive.

There are costs and risks.

S&P Global loses some diversification and some synergies.

The company also has to execute through organizational changes in Market Intelligence and Energy.

But I generally prefer simpler businesses when simplicity increases the concentration of the best economics.

That appears to be what happened here.

The business is easy to like. The price still matters.

At around $395, S&P Global trades at about 22.5x the midpoint of management’s 2026 adjusted earnings guidance.

I think about the return in a few simple cases.

If earnings grow 10% annually for five years and the stock ends at 22x earnings, the shares would be worth about $624, or roughly a 9.5% annual price return before dividends.

At 12% growth and a 25x ending multiple, the value is about $777, or roughly 14.4% annually.

At 8% growth and a 20x multiple, the value is about $518, or roughly 5.5% annually.

Those are not forecasts. They are simply a way for me to see what today’s price requires to produce an attractive return.

My takeaway is that S&P Global is becoming more interesting, but I would not call it a screaming bargain. I like the quality more than I like the base-case return, which means I still want some patience on price.

I also would not give full credit to adjusted earnings without thinking about the exclusions. Management’s guidance removes acquisition-related amortization and other items. That does not make the number useless, but it is another reason I do not want to stretch the valuation.

The risks I am watching most closely are:

Debt issuance slows. Ratings transaction revenue benefits from healthy issuance markets. A weaker refinancing cycle or recession can reduce activity.

AI weakens data and workflow products. If customers can replace expensive data products with cheaper tools, Market Intelligence could face pressure on pricing or retention.

Regulation changes the economics. Ratings, benchmarks, data, and commodity pricing all operate in regulated environments. More regulation can raise costs or alter how products are used.

Benchmark competition increases. Index fees are attractive precisely because the business is so good. That attracts competition and pressure from large asset managers.

Capital allocation becomes less disciplined. Large acquisitions or repurchases at poor prices can destroy value even when the underlying business remains strong.

The Mobility separation does not deliver the expected focus. The remaining company still has to prove that the cleaner structure produces better growth and returns.

I would become more cautious if:

The stock going down doesn’t worry me by itself. The business getting weaker does.

What I like about S&P Global is that it gets paid for infrastructure the market already depends on.

Companies need ratings.

Investors need benchmarks.

Financial institutions need data and workflow tools.

Commodity markets need trusted pricing and reference points.

S&P Global does not have to predict which bond performs best, which index constituent wins, or which commodity price moves next.

It gets paid for helping the market function.

That is the type of business I want to own.

The company is not perfect. Ratings is cyclical. Market Intelligence faces disruption. Indices is exposed to market levels. And capital allocation still has to be judged against price.

But the combination of trust, embedded workflows, proprietary data, benchmark economics, strong margins, and low capital intensity is difficult to recreate.

At roughly 22.5 times management’s 2026 adjusted earnings guidance, the valuation is getting more interesting.

The quality was already obvious.

Disclosure: I own shares of S&P Global as of the date of publication. This research reflects my personal views and is not investment advice.