$SPGI at ~21x forward earnings. What you own at that price: Roughly half the global credit ratings market. Almost every bond issued on earth is rated by them or Moody's. The S&P 500 itself. Every fund tracking it pays a license fee. Platts, which sets the benchmark most of the world's oil is priced against. The data terminal a large share of Wall Street runs on. Four businesses nobody can build a competitor to, at a multiple the market hands to average companies.

$GRAB in 2026: Profitable. Buying back stock. EBITDA up over 50%. Guidance raised to 30% CAGR through 2028. Expanding into Taiwan, drone delivery and autonomous vehicles. The CEO bought at $2.89. The COO bought too. Stock is down 38% this year at $3.15. Make it make sense.

Everyone thinks $CPRT is an online auction company. It owns more than 20,000 acres of land. That's the actual business. Salvage vehicles have to physically go somewhere, and nobody is getting permission to build new yards near cities. The stock is at 20.9x free cash flow against a 43.3x historical average. Free cash flow compounded at 23% a year since 2013. Growth is slowing to 10-12%, and there's a proposed ACV acquisition to look at. Still cheap for what it is.

$CPRT has compounded revenue per share at roughly 14.2% annually for 21 YEARS. From $0.30 in 2005 to $4.88 in 2026. More than 16x. Only ONE annual decline in that stretch: 2015. Through the 2008–2009 financial crisis, revenue per share held around $0.55. Even during COVID, it climbed from $2.12 in 2019 to $2.31 in 2020 and $2.80 in 2021. Copart built this business auctioning vehicles, with its platform now reaching roughly 1 million members across 185+ countries. What interests me is how difficult that combination of yards, insurer relationships and global buyers would be to replicate… The latest year was much slower, though: FY2026 revenue grew just 0.4%, while buybacks helped revenue per share increase about 2.6%.

Visa vs Mastercard looks like a coin toss. Until you zoom out. Over the 5 years through September 18, 2026, with dividends reinvested, $MA returned 11.16% annually. $V returned 11.52%. Pretty close. Now extend that same comparison to TEN years. Mastercard: 19.69% annually. Visa: 17.02%. That turned the same $10,000 investment into roughly $60,300 with Mastercard versus $48,200 with Visa. Both made patient shareholders serious money. But calling them interchangeable misses how much a few percentage points can add up. I like businesses like these. I’m still going to care what I pay for them, because those historical returns aren’t a promise about the next decade.

A missing bolt can stop a production line. Keeping that bolt available is part of the $FAST business. Fastenal places inventory, smart bins and vending machines inside customer facilities, then helps manage replenishment. Q2 2026 sales grew 14.7% to $2.39 billion. Sales to contract customers increased 17.6%. The interesting part is how closely the supplier becomes involved in everyday operations. Once Fastenal manages part of a factory’s inventory, it has an opportunity to earn more of that factory’s spending.

Airlines have a strong incentive to find cheaper replacement parts. Getting those parts approved is the hard part. $HEI makes FAA-approved aircraft replacement parts and provides repair services. Fiscal Q3 2026 sales grew 23% to $1.41 billion, including 14% organic growth. Net income increased 33%. The attraction is what happens after approval: parts wear out, maintenance continues, and customers reorder. A growing catalog gives HEICO more ways to earn money from aircraft already flying.

A surgical robot can work for years. Its sterile drapes get thrown away after the procedure. $UFPT makes those drapes, equipment covers and other disposable medical components. Q2 revenue grew 15.1% to $174 million, including 12.4% organic growth. Net income grew 21.4%. For its single-use products, the customer has to keep ordering as procedures continue. The disposable part of an expensive machine can be a pretty good business.