The AI build-out is missing a third metric

Every dashboard in the AI build-out tracks two families of numbers. Capacity asks: can we build it? Efficiency asks: what does it cost to run? Both are indispensable. Both are measured entirely inside the fence line.

Neither can see a slipped interconnection, a water moratorium, a denied permit, or a community that organizes against the hum — the failures that, in the first quarter of 2026 alone, blocked or delayed at least 75 data-center projects worth roughly $130 billion.¹ Those failures do not happen inside the machine. They happen at the boundary between the machine and the systems that host it. Over the same quarter, the number of active opposition groups more than doubled, from 396 to 833, spanning 49 states.

Infrastructure Health is the degree to which a facility and its host systems — grid, watershed, airshed, land, and community — sustain one another over the asset's operating life. It is a boundary property, it is bidirectional, and it is an engineering objective rather than a virtue: something to be specified, measured, designed for, and traded off like any other requirement.