$GE sells the engine once. The economics can last for decades. GE Aerospace has roughly 50,000 commercial engines installed globally. Those engines can stay in service for decades, generating parts, maintenance, repairs, and shop-visit revenue. That installed base is what makes the business interesting to me. But a great business isn’t automatically a great investment. Price still matters. I wrote up the moat, economics, risks, and valuation here: therobbenletter.com/p/ge-aer… This is how I think about business quality, valuation, and risk.

This week's Robben Letter: Intercontinental Exchange. Everyone knows it owns the NYSE. That is not why I own the stock. The networks behind it are the business: exchanges, clearing, energy benchmarks, fixed-income data, mortgage software. therobbenletter.com/p/ice-ex… Disclosure: long $ICE.

Public markets are strange. Everyone says they want to own great businesses for a long time. Then the stock falls 20%, the headlines get ugly, and suddenly their time horizon becomes 20 minutes. The edge is often simpler: own a small number of exceptional businesses, know them deeply, and be willing to look wrong for a while.

My rule: business quality is not a substitute for price discipline. S&P Global is the rare one where I get both. That is why it is a core position, not a starter. (I own $SPGI)

Most people think of the NYSE when they think of Intercontinental Exchange. That’s not why I own the stock. $ICE has built a collection of exchanges, clearing networks, data businesses, and software that are difficult to displace once customers build workflows around them. The bigger question now is capital allocation, especially MarketAxess and the added debt. I wrote about the business, valuation, and what I’m watching in the latest Robben Letter: therobbenletter.com/p/ice-ex…

One thing I find attractive about certain infrastructure assets is how difficult they are to recreate. A competitor can’t simply buy land next door and build a competing asset. There are land constraints, permits, regulation, infrastructure, government and business relationships, and decades of traffic patterns involved. The scarcity creates durability. One downside here is that infrastructure is capital intensive and heavily regulated, so valuation still matters. But truly scarce assets will get a closer look from me, especially in an AI world.

S&P Global expects $7 billion in buybacks this year. When a great business retires stock at low-20s earnings, every dollar works harder. Capital allocation matters as much as the business itself. (I own $SPGI)

I believe investors spend too much time talking about revenue growth without going deeper. A dollar of revenue from a customer who has to reconsider the purchase every year is different from a dollar from an essential service. A dollar in a highly competitive market is different from one where the provider has pricing power. Recurring revenue is useful, but I want to know why it recurs... habit, contract, switching costs, regulation, network effects? Essential revenue > discretionary

Pricing power test: a company issuing $2B in debt does not pick its ratings agency because one is 5% cheaper. Credibility decides. Small fee for an enormous value of the decision. That is S&P Global. (I own $SPGI)

Here's the bear case on my S&P Global position. If debt issuance freezes for years, ratings revenue drops with it. If passive investing ever loses share, index fees slow down. I bought it anyway. At 23x forward earnings, the price was the margin of safety. (I own $SPGI)

Stock prices are as volatile (or more) as they ever have been. Glance at the gaps between 52 week high and low and ask yourself whether the underlying business value changed that much. There is always an edge in being willing to think 3-5 years out because most people aren’t doing that. AI can do whatever we tell it to, but as long as humans own the money invested, there will be emotions, impatience and irrational behavior at times. This causes mispricing

The moat: you can build software. You cannot recreate decades of market acceptance for the S&P 500 or an official credit rating. Trust is the moat, and trust compounds. (I own $SPGI)

Tuesday's Robben Letter is the full S&P Global deep dive: the four divisions, the Moody's and MSCI comparison, the bear case, and my valuation math. Free at therobbenletter.com. (I own $SPGI)

I bought more S&P Global this week. It’s now a core position. SPGI owns two of the strongest franchises in finance: credit ratings and the S&P indices. But it trades around 23x forward earnings vs. 27x for Moody’s and 29x for MSCI. I wrote up the moat, four divisions, bear case, and valuation here: therobbenletter.com/p/sp-glo… Disclosure: Long $SPGI and $MSCI.

How to get “lucky” in public markets: Study companies before anyone cares. Keep a list of prices that would make you interested. Do the work before the opportunity shows up. Wait. Then act when something finally gets mispriced. People see the returns. They usually don’t see the six months you spent getting ready before buying.

One thing I have become much more focused on is the difference between historical returns on capital and future returns on incremental capital. A mature business can report fantastic returns because of investments made years ago. This does not always mean it has attractive places to put the next billion dollars. The really interesting/rare companies have both. They already earn attractive returns and they still have a long runway to deploy additional capital at good rates. Compounding picks up pace here.

Revenue growth gets attention. Reinvestment economics create wealth. Company A grows 20% but requires heavy reinvestment at mediocre returns. Company B grows 12% with little incremental capital and converts most earnings to free cash. I’d rather own the second business if the valuation is sensible. Growth matters. But the economics of that growth matter more.

A great business can still produce mediocre investment returns. If a company compounds earnings at 15% for five years but its valuation falls from 30X earnings to 20X, your return is closer to 6% annually. That’s why I separate two questions: How good is the business? and What return am I likely to earn from today’s price? I want both. Exceptional businesses, purchased at prices that can still produce strong long-term returns.

Everyone argues about whether AI demand slows down. I own TSMC and ASML. They get paid to build the capacity either way. I'd rather own the toll road than predict the traffic. Disclosure: long $TSM, $ASML.