Dear Friends,
Every time a company issues a bond, someone has to grade how safe it is. Every time you buy an ETF, someone owns the index it tracks — and collects a fee. Two companies sit at the center of all of this: SPGI(SPGI) and MSCI Inc (MSCI). They are the toll booths of global finance: money flows through the markets, and they take a small cut of nearly everything that passes.
If you have never heard of them, here is the one-minute version.
What do these companies actually do
SPGI has three big businesses. First, it grades bonds when a company or government borrows money, S&P gives it a credit score, the way a credit bureau scores you and me. Borrowers pay for the grade, investors trust it, and a 160-year-old name is very hard to compete with. Second, it owns the S&P 500 and Dow Jones index families - every fund tracking those indexes pays S&P a licensing fee. Third, it sells financial data and analytics to banks and investors. (Until this July it also owned Carfax and other car-data businesses, but it split those off into a separate company to focus on finance.)
MSCI is the index company behind most of the world’s international stock ETFs - if you own an emerging-markets or Europe fund, it probably tracks an MSCI index, and a slice of the fees flows to MSCI. It also sells risk-management software and sustainability data to big investors. Its customers renew at about 95% a year: once an index is wired into trillions of dollars of investments, nobody rips it out.
Both are classic long-term compounders: subscription-like revenue, profit margins above 30%, and businesses that get stronger the more the market uses them.
Where we left off
Last November, in Byte #30, I flagged both as long-term holds after they had slipped from their highs. Both traded at around 35 times earnings and I leaned toward SPGI because its business was more diversified.
Eleven months later, the two have gone in very different directions. Time for an honest revisit.
The scoreboard: then vs. now (approx. numbers)
Same toll-booth model but SPGI fell ~20% while MSCI held flat: a 19-point gap in price return since November. The difference was not the businesses; both performed. It was the story: one company spoke the market and is still paying for it; the other took the same February hit but bounced back on AI momentum.
What happened since: SPGI’s rough year
February was the turning point. SPGI reported a decent quarter but told Wall Street that 2026 profits would come in a touch below expectations and that it needed to spend heavily rebuilding its platforms around AI. The stock crashed 16 –18% in a single day, its worst in years, and dragged the whole financial-data sector down with it. The market’s fear: if AI can do what data companies do, are these toll booths still safe?
Three things happened next:
The CEO put her own money in. In late April, CEO Martina Cheung bought about $1M of S&P stock on the open market at ~$430. Two other senior leaders bought about $1.5M combined. Insider buying is never a guarantee, but executives rarely commit millions of dollars of their own capital if they believe the stock is materially overvalued.
The company got leaner. On July 1, S&P split its car-data business (Carfax and friends) into a separately listed company. What remains is pure financial infrastructure: bond ratings, indexes, market data, and energy data.
The business roared back. The next quarter set a revenue record for the bond-ratings business (+17%) and a 13th straight record for indexes (+20%), with earnings up 23%. Management raised its plan to buy back its own shares to over $7 billion -buybacks shrink the share count, so each remaining share owns a bigger slice of the company.
Institutional buying. Pershing Square opened a new ~$1.06B position in Q2 and also a few others.
And yet the stock sits near $389, down about 21% this year. The market still has not forgiven February.
What happened since: MSCI’s bumpier year than it looks
MSCI’s year looks calm on the surface - roughly flat at ~$552 but the chart tells a rougher story, with two sharp selloffs.
The first came in February: the stock fell from ~$630 to ~$502 in about two weeks. MSCI hadn’t reported anything bad - it had just posted a solid quarter. It was collateral damage from SPGI’s February 10 warning, which set off a sector-wide rout in data stocks on AI disruption fears. The same fear that crushed S&P hit MSCI too; it just recovered faster.
The second came in July: second-quarter revenue beat expectations (+12.2%), but the company raised its 2026 expense forecast - acquisition costs, employee incentives, investment spending and earnings missed by a whisker. The stock fell ~11% in a day. (Back in January, the CFO had already flagged heavier 2026 spending - a new London office, more software investment so this wasn’t entirely out of the blue. Not a margin collapse, but expenses are the line item to watch.)
In between the scares, the business kept executing: revenue growth accelerated past 13% in the first quarter, customer retention held near 95%, the dividend went up ~14%, and billions went to share buybacks. And MSCI leaned into AI instead of fearing it: a tool letting clients interrogate index data through ChatGPT-style assistants, new AI supply-chain indexes, and daily price benchmarks for private assets. The market now treats MSCI as an AI winner but at ~29x earnings with expenses creeping up, it gets no benefit of the doubt either.
MSCI in 2026: ~$630 → ~$502 in February (sector-wide AI rout), a recovery toward ~$650, then ~$640 → ~$558 on July’s expense-forecast raise.
What I think now
- The AI fear cut both ways. A year ago, I listed AI as a shared risk. Instead, the market decided MSCI is using AI to sell more products - twice as many firmwide launches in 2026 as in all of 2024 while S&P is spending on AI to defend itself. Offense versus defense, and stock prices followed the narrative.
- SPGI’s Market Intelligence (~31% of revenue, weakest margins) is the genuinely exposed piece - Capital IQ faces AI pressure; Analysts calls the fear overdone, but it's the reigning narrative and it caps the multiple.
- SPGI now trades at ~24x earnings instead of 35x. MSCI barely got cheaper. Profit margins above 30% put it in the top quartile of the S&P 500 - and that high-margin profile makes it "low duration," meaning it is less jerked around by the economy than most stocks.
- Don’t confuse stocks with the company. S&P’s bond-ratings business just had a record quarter; MSCI keeps ~95% of its customers. Both businesses are working fine. Revenue grew. Customers stayed. Moats held. Only the market's story changed.
Food for Thought
- When a great business guides conservatively and spends to defend its moat, is a 20% drawdown a broken thesis - or the entry point the November byte was waiting for?
- MSCI has executed better and the market now pays it an AI-beneficiary premium. At ~29x with consensus targets near $700, is the easy money already made?
- SPGI’s leaner post-spin portfolio is ~65% benchmark businesses at >80% of operating profit. If you believed the moat story at $489, what exactly changed at $389?
- Analysts expect earnings to grow ~11–12% a year at SPGI and ~13% at MSCI over the next five years - steady compounders, not 20% growers, which is exactly why it matters whether you pay 24x or 29x earnings for that growth.
My lean hasn’t changed - it’s sharpened. I favored SPGI in November for diversification; today I favor it for valuation: a record-setting Ratings and Indices franchise, a $7B+ buyback, insider buying from the C-suite, and institutional buying. MSCI remains the better-executed company with more momentum, but I’d rather buy excellence on sale than at full price.
Let’s call it what it is: this is partly a multiple play. You’re buying SPGI’s ~12% annual earnings growth at 24x - a multiple the market charged 35x for a year ago. If the AI fear fades and the multiple drifts back toward 30x, that’s roughly a 25% bonus on top of the compounding. And if it never re-rates? The earnings growth alone still carries the thesis.
Both report Q3 later this month (MSCI Oct 20, SPGI Oct 27) - I’ll be watching whether the AI-spend narrative at SPGI starts converting into the AI-beneficiary narrative MSCI already enjoys.
As always, I’d love your thoughts - your pushback is what makes these Bytes better.
Just a reminder: the thoughts shared in this Byte are meant to help you think more clearly about investing, not to tell you what to buy or sell. Every investor's situation is unique, so please do your own due diligence before making any investment decisions.
Cheers,
Pooja