Designed Into the Reference ArchitecturesNarrow moat

One of the most valuable positions in the AI build-out is to be designed into the reference architectures — the standard blueprints that the chipmakers and system builders publish for how to construct an AI cluster. When Nvidia or another platform vendor specifies a recommended rack-scale design, the cooling, power, and infrastructure that go into that blueprint become the default choice for the many customers who build to the standard rather than engineering from scratch. Vertiv has worked hard to earn a place in these reference designs, co-developing with the chip and system vendors so that its liquid-cooling and power solutions are part of the recommended architecture.

Being in the blueprint is a powerful advantage: it puts Vertiv in front of customers at the moment of design, lends the credibility of the platform vendor's endorsement, and channels a flow of standardized demand toward the company's products. It is a form of designed-in preference that is hard for a competitor outside the reference design to displace. But — as the accompanying threat develops — it is a double-edged position: the same chipmakers whose reference designs elevate Vertiv also hold the power to standardize the cooling specification in ways that commoditize it, to include multiple suppliers, or to shift the design in a competitor's favor. Being designed in is a genuine advantage, but it rents credibility from the chip vendors rather than owning it, and that dependence is part of why the moat is narrow — however much revenue the +28% ride delivers¹.