Editorials

For nearly a century, S&P Global has thrived as a gatekeeper. The company sits at the crossing points of global capital markets, energy trading, automotive supply chains, and indexing. Its databases and benchmarks have accumulated switching costs and network effects that seemed almost unassailable. Market Intelligence generates $4.2 billion in annual revenue, Ratings contributes $1.9 billion, and the combined enterprise commands a market cap near $121 billion.

But the rapid deployment of large language models and AI-driven analytics is rewriting the calculus. For SPGI, this shift presents both existential threats and unexploited opportunities that investors have yet to fully price in.

S&P Global's core franchise has long rested on a simple truth: proprietary data and expert synthesis are scarce. A portfolio manager pays thousands in annual subscriptions for Capital IQ and RatingsDirect because extracting intelligence from raw market data and credit spreads requires human judgment, patterns recognized over decades, and access to hard-to-find information.

Genative AI breaks this model in three ways.

First, large public datasets are becoming self-sufficiently analytical. A competent LLM can now synthesize earnings transcripts, regulatory filings, and news feeds at a fraction of SPGI's cost. Why pay for curated research when a trained model can generate bespoke analysis in seconds?

Second, AI lowers the barrier to entry for competitors. A well-capitalized hedge fund or fintech startup can now train proprietary models on historical financial data and deploy analysis engines that match or exceed SPGI's advisory and workflow products. The company's intangible moat shrinks.

Third, price discovery itself may become more efficient and transparent. If energy traders can use AI to mine real-time supply-chain signals, weather data, and logistics networks as effectively as S&P Global's Platts price assessments, why pay the subscription? The market may demand transparency over authority.

But SPGI is not powerless. In fact, the company is in a position to own the AI layer if it executes with conviction.

S&P Global already controls the inputs: 44,500 employees globally, 150+ years of credit ratings data, proprietary price benchmarks across energy and commodities, and relationships embedded across institutional capital markets. No startup owns that.

The opportunity is to stop selling data and human analysis separately, and instead sell the reasoning engine. Imagine Capital IQ not as a database you query, but as an AI agent that knows your portfolio, your risk tolerances, and your regulatory constraints, and acts as a true analyst-on-demand. Imagine Platts price assessments not as published benchmarks, but as a real-time predictive model that incorporates satellite imagery, shipping manifests, and geopolitical signals before humans even see them.

This requires SPGI to invert its business model from subscription-to-access toward subscription-to-outcome. It's a harder sell and a longer sales cycle, but it's also far more defensible against commoditization.

S&P Global trades at a P/E of 28.1x and an EV/EBITDA of 17.2x. Those multiples price in sustainable competitive advantage. The question is whether AI erodes that moat or widens it.

Management's planned separation of Mobility in mid-2026 suggests confidence in the core business. But the real test will come in the next 18 months: Can SPGI demonstrate that its AI-powered analytics command pricing power equivalent to its legacy products? Or will the market see through the rebranding and demand a lower multiple?

The answer will determine whether S&P Global remains a moat-protected compounder or becomes a slow-motion value trap.