Pieter Slegers — Best Buys: July 2026

S&P Global sits in two places where money has to pass through and takes a cut at each. When a company borrows by issuing a bond, it needs a credit rating, and there are only three firms that can give one — S&P, Moody's and Fitch. When an investor buys an S&P 500 tracker, the fund pays to use the index name and the list of companies in it, and S&P owns that index.

Slegers' point is that neither business is really about being clever. They are about being one of three permitted operators in a market everyone else must use. That is what he means by a toll bridge: the traffic decides the volume, but nobody gets to build a competing bridge. Oligopolies of that kind, he says, "enjoy high barriers to entry, pricing power, and high margins."

He ranks it the single most likely purchase of the month, and a week later he buys it — this is the shortlist the portfolio actually draws from. The company also recently separated its mobility division into its own listed business, which the market took well.

An insurer collects your premium today and pays your claim, if you ever make one, years later. In between it is holding a large pile of other people's money — the industry calls it float — and whatever it earns investing that pile belongs to shareholders. That is the whole Berkshire idea, and Fairfax runs it deliberately.

The person doing the investing is Prem Watsa, whom Slegers calls "The Warren Buffett of Canada." The name Fairfax is the method: fair prices, friendly deals, no hostile takeovers — sellers approach them because they are not going to be dismantled.

There is one number that tells you whether the insurance half is working: the combined ratio, which compares what the insurer pays out in claims and expenses against the premiums it takes in. Under 100% means the underwriting itself makes money, so the float is genuinely free — the investing returns are pure profit on top. Fairfax's is below 100 and the float keeps growing, which is exactly the compounding engine he is buying.

Mastercard does not lend anyone money and does not carry the risk if a cardholder defaults. It owns the wires — the network that tells a shop's terminal whether your bank will pay — and it takes a fraction of a cent every time that happens, billions of times a day.

The defence is a loop nobody can enter from either side. Shops accept the card because everyone carries one; everyone carries one because every shop accepts it. A new entrant would have to sign up both halves simultaneously and, as Slegers puts it, "building a competing network from scratch is practically impossible."

The second attraction is that the network is already built and paid for. Each extra transaction costs almost nothing to run, so revenue growth drops almost straight through to profit. On top of that Mastercard now sells fraud detection, cybersecurity and analytics to the same banks and merchants — extra revenue that rides on infrastructure it already owns.

When a car is written off after an accident, the insurer does not want to keep it — it wants cash. Copart runs the online auction where those wrecks are sold to dealers and dismantlers, and takes a fee on every sale. It never owns the risk of the car; it owns the marketplace.

Marketplaces defend themselves the same way everywhere: insurers list where the most buyers are, because more bidders means a higher price, and buyers shop where the most cars are, because more cars means more choice. Each side makes the other side stronger, and a rival starting today has neither.

The less obvious moat is dirt. Copart owns its salvage yards outright while competitors rent theirs. Nobody wants a noisy yard full of leaking fluids near their house, so zoning and environmental rules make new yards near big cities almost impossible to permit — the land Copart already holds cannot be replicated at any price. Owning it also means no rent increases, and each additional car processed on land you already own costs almost nothing, so volume turns straight into margin.

The timing argument is that the shares fell 12.8% in the month, growth has slowed, and the board has brought back Jay Adair — the CEO under whom the stock rose more than 2,000% between 2010 and 2024. Meanwhile the company holds net cash worth about 15% of its entire market value, which is expected to go into buying back shares.

Adyen handles the plumbing when you pay a large online business — moving the money between the shopper's bank, the card network and the merchant, and taking a small fee. Its competitors mostly grew by acquisition and run a patchwork of incompatible systems; Adyen built one platform for the whole world, which is why big multinational retailers prefer it.

One platform also means one pool of data. Every payment Adyen sees teaches the system a little more about which transactions are genuine, so it approves more real purchases and blocks more fraud than a rival looking at a fraction of the traffic. More merchants means more data means better results means more merchants.

Two things are new in this issue. Adyen, which had famously built everything itself, has made its first two acquisitions — Talon.One for customer loyalty and promotions, Orb for AI-driven enterprise billing — with the stated aim of becoming a full commerce platform rather than just a payments processor. And the alignment is unusual: founder and co-CEO Pieter van der Does still owns close to 3% of the company, worth over €750 million, with other insiders holding nine-figure stakes of their own.

Paychex runs payroll. When a small or mid-sized American business pays its staff, works out the tax withholding, files it correctly and administers the health plan and the retirement plan, there is a decent chance Paychex is doing it — 800,000 clients, and one in every eleven American workers.

The market's fear is simple: it is software, and AI eats software. Slegers' answer is that the product being sold is not really software, it is the absence of a catastrophe. Getting payroll or tax filing wrong means angry employees, regulators and lawsuits, so almost nobody switches provider to save a few percent — and the more services a client buys (HR outsourcing, retirement, insurance), the more unthinkable moving becomes.

The numbers he quotes back that up: a 25-30% net margin, a 15-20% return on invested capital, and free cash flow consistently above reported net income, which means the accounting profit is real cash rather than a bookkeeping figure. The growth lever is the 2025 purchase of Paycor for $4.1 billion, which pushes Paychex up-market into companies with more than a hundred employees — bigger clients, stickier relationships and more room to raise prices.

Note the status: this is the monthly spotlight, not one of the five ranked buys — a business being made ready for the list rather than one on it.