Pieter Slegers — 10 Stocks for the next 20 years (Arka's selection)

Visa does not lend money or issue cards. It owns the wires between the shop, the shopper's bank and the card issuer, and takes a very small fee every time money moves along them. Because the network already exists, each extra transaction costs almost nothing to carry, so growth in payment volume drops almost entirely into profit.

The twenty-year argument is that cash keeps disappearing, that a rival would have to sign up every shop and every cardholder simultaneously before its network was worth anything to either, and that the tollbooth position is therefore effectively permanent. It is ranked first here and is also one of the largest positions in the real portfolio, held at the firm's highest conviction level — the strongest agreement between a thought-experiment list and actual money in this batch.

Alphabet owns Google search, YouTube, Android, Maps and a large cloud-computing business, and makes most of its money selling advertising against what people are already looking for. That is the most valuable kind of advertising there is, because intent is already present.

Three reasons are given for the next twenty years: people will always need a way to find things; the accumulated record of what billions of people search for is a training advantage for AI that cannot be bought; and the products are woven into daily habit. The obvious counter — that AI assistants replace the search box entirely — is not addressed here, though two days earlier in this archive a growth investor described buying the shares precisely because the market had over-believed that story.

Microsoft sells the software companies run on — Windows, Office, Teams — and rents them computing power through Azure. Once an organisation's documents, email, identity system and applications all live in one vendor's ecosystem, moving is a multi-year project nobody wants to lead.

It is also the pick that proves the article's own point: Microsoft was the second-largest company in America in 2005 and the fourth-largest in 2026, one of only a handful to hold that position across two decades of technological upheaval. It did so by moving from selling packaged software to renting cloud capacity — exactly the "continuous adaptation" the piece names as the common factor.

ASML builds the machines that print the circuit patterns onto silicon wafers. For the most advanced chips this requires extreme ultraviolet light, and ASML is the only company on earth that can make a machine capable of it — each one costing hundreds of millions and taking years to build and install.

That is as close to a genuine monopoly as public markets contain: every advanced chip in every phone and AI system starts inside one of its machines. The risk the piece does not mention is the one that follows from the same fact. A single company controlling the chokepoint of the most strategically contested technology on earth is a political asset as much as a commercial one, and over twenty years that cuts both ways — export controls, forced technology transfer and state-funded attempts at replication are all live.

Xylem makes the equipment that moves, treats, tests and meters water — pumps, filtration systems, smart meters and the software that runs them. Its customers are mostly water utilities and municipalities, plus industrial users.

The case is straightforward and long-dated: clean water is not optional, the pipe networks of the developed world were largely laid a century ago and are failing, and replacing them is a multi-decade programme that has barely started. Adding sensors and software to those networks — detecting leaks, managing pressure — is the growth layer on top.

The consideration to hold alongside it is that much of the revenue comes from public budgets rather than commercial customers. That makes demand very durable and also very slow, and means the price is set through procurement rather than through pricing power. It is the least conventional name on the list and the one most dependent on politics.

Schneider Electric makes the equipment that manages electricity inside a building or a factory — switchgear, circuit protection, uninterruptible power supplies, building-management software. It operates in more than a hundred countries.

Three forces are named, and they compound: the general shift from burning fuel to using electricity; data centres, whose demand for power the piece calls bottomless; and the pressure on every building and plant to use less energy for the same output. Each of those turns into orders for the same catalogue of products.

This is the AI-power constraint tracked elsewhere in this research hub, expressed as a supplier rather than as a utility — which means it captures the build-out without owning the generating assets or carrying the regulated returns. Note it trades in Paris; the ticker "$SU" in the post is the French listing, not Suncor.

Applied Materials makes the machines that deposit, etch and shape the layers of material on a silicon wafer — everything, roughly speaking, other than the lithography step that ASML dominates. Almost every chip made anywhere passes through its equipment.

The attraction is that it does not have to pick a winner. Whichever chipmaker gains share, the equipment gets bought, so the position is a claim on total chip manufacturing rather than on any one manufacturer. The engineering is at atomic scale and takes decades to replicate, which is the barrier.

Worth noting the cyclicality this argument glosses over: semiconductor equipment orders swing violently with the capital-spending cycle, and a twenty-year hold means living through several of those swings without selling.

Thermo Fisher supplies laboratories with almost everything they use — instruments, reagents, consumables, diagnostic equipment, and outsourced drug-manufacturing services. It sells to essentially every pharmaceutical company, biotech and research institution in the world.

Two things make it durable. First, it is a picks-and-shovels business: it is paid whether or not any given drug works, which is the same reason Medpace is held in the real portfolio. Second, and more powerfully, once a specific instrument or reagent is written into a regulator-approved testing protocol, switching means revalidating the whole process. That is a switching cost enforced by law rather than by preference, and it is the strongest form there is.

Old Dominion moves freight that is too big for a parcel carrier and too small to fill a lorry, by consolidating loads from many customers onto shared trucks routed through a network of more than 260 terminals across North America.

The terminals are the business. A competitor cannot serve a customer well until it has enough of them in the right places, and buying that much industrial land near cities today is close to impossible — which is why the network is described as costing a billion dollars and being nearly uncopiable. The company also runs at the best margins in its industry, which is what allows it to keep funding the network while rivals cut back.

Ferrari sells about ten thousand cars a year and could sell considerably more. It chooses not to. Waiting lists are years long, existing owners get priority on new models, and the scarcity is manufactured and defended deliberately.

That is why the argument here is that it is a luxury company rather than a car company: it is not competing on performance or price against other manufacturers, it is selling access to something deliberately withheld. As long as the number of very wealthy people grows faster than Ferrari's production, the price can rise indefinitely and demand will not fall.

It is the mirror image of the pricing power described elsewhere in this batch. Diploma and PPG raise prices because the customer does not notice; Ferrari raises prices because the customer cannot get the product any other way. The risk, unstated, is that desirability is cultural and cultures change — and that racing, the brand's original justification, is a shrinking part of the story.