The MoatNarrow moat

The quickest way to understand Vertiv is this: when someone builds a data center full of expensive computers, two things have to work perfectly and forever — the electricity coming in and the heat going out. Vertiv sells both. It makes the uninterruptible power supplies, switchgear, and busway that guarantee the current never stops, and the cooling — increasingly liquid cooling, piped straight to the chips — that carries away the furnace of heat a rack of AI accelerators throws off. That has made it one of the purest arms dealers of the artificial-intelligence build-out and one of the stock market's great winners of the decade. My argument is that the business is genuinely excellent and the moat is genuinely real — but it is a narrow moat wearing a wide moat's price tag, and telling those two things apart is the whole job of an investor here.

Start with what keeps the moat narrow, because it is the thing the price ignores. Vertiv is not the biggest or most powerful company in its own markets. In power it goes head-to-head with Schneider Electric and Eaton, electrical giants several times its size⁵ for whom this is the core business, not a sideline. In cooling it leads a fragmented field where "leadership" means a low-double-digit share, not dominance. It owns no category outright. Much of what it sells, however cleverly engineered, is ultimately hardware, and hardware drifts toward commoditization and price competition — especially when your biggest customers order in bulk and design their own standards. A specialist can beat generalists by being faster and more focused, and Vertiv has, taking share and out-growing its rivals in data centers. But that is an advantage of execution, re-earned every quarter, not a structural wall.

What transformed the business is thermal, and specifically liquid cooling. A rack of AI chips runs so hot that moving air can no longer save it; the heat has to be carried off by liquid brought to the silicon itself. This is hard, high-stakes engineering — get it wrong and you cook millions of dollars of chips — and it is exactly the discipline Vertiv is best at. Its cooling revenue has at times more than doubled year over year, and it has fought its way into the reference designs the chipmakers publish, so that when a customer builds to the standard blueprint, Vertiv's cooling is already in it. This is the sharpest growth driver in the whole company. It is also the most contested — which is the recurring pattern here: every strength has a capable crowd chasing it.

The single most moat-like thing Vertiv owns is also its least glamorous, and I think the least understood. Once its power and cooling are bolted into a live, mission-critical data center, they do not come out — you cannot rip the cooling out of a running facility — and for years afterward they need servicing, spare parts, monitoring, and upgrades. That recurring, high-margin service revenue, defended by switching costs and by a global field-service network that is slow and expensive to copy, behaves like an annuity, not like cyclical hardware. And here is the part the market underplays: every box sold in today's boom seeds years of future service revenue, so the record equipment sales are quietly building a far larger annuity underneath. The hardware is what everyone is paying for; the annuity is what may actually endure.

Around that core sit the advantages of scale — global manufacturing capacity that is costly and slow to build, deep co-development relationships with the hyperscalers designing the AI cloud, and a startling order book: a backlog near $15 billion¹ at the end of 2025, more than double a year earlier, after a quarter in which orders grew over 250% organically. But each of these strengths is the mirror of a weakness. Those deep relationships are a concentration — a handful of giant buyers with immense bargaining power drive the demand — and that backlog was built at the high tide of a capital-spending cycle Vertiv does not control.

The headline numbers are extraordinary: revenue up 28% in 2025 to over $10 billion², adjusted operating margin climbing past 20%, guidance repeatedly raised toward $14 billion for 2026. And the stock trades at roughly 56 times trailing earnings³ — a price that assumes Vertiv is a core, assured, permanent beneficiary of the build-out. The market showed how little room that leaves in the second quarter of 2026: Vertiv beat on earnings and raised guidance, but revenue came in a touch light on supply-chain timing⁴ — not a demand problem at all — and the stock fell about 14% in a single day.

So the verdict, plainly. Vertiv is a high-quality, specialized, cash-generative leader at the dead center of the AI era, with a real but narrow moat, larger rivals on every side, a concentrated and cyclical customer base, and a valuation that has already priced in success. Own it knowing exactly which of those things is true. The number to watch is not revenue growth — the boom guarantees that for now — but the recurring services mix: if that annuity keeps compounding faster than the hardware, the narrow moat is quietly widening into something more durable; if it doesn't, this is a cyclical priced as a compounder, and the cycle always eventually turns.