Analysis
Vertiv, a company with roughly eighty years of history rooted in industrial air-conditioning and power equipment, has become one of Wall Street's favored AI infrastructure stocks, Fortune reported this week, under veteran Fortune 500 CEO Dave Cote.
Why an air-conditioning company became an AI story
Vertiv's core products -- data-center cooling systems and power management equipment -- sit at exactly the physical bottleneck the AI infrastructure buildout has hit repeatedly this year. Dense racks of AI training and inference chips generate far more heat and draw far more power per square foot than the data-center equipment built for a prior generation of computing was ever designed to handle. As hyperscalers and AI labs race to build out capacity, the cooling and power infrastructure required to actually operate that capacity safely has become as much a gating factor as chip supply itself -- a dynamic Pulse has covered from multiple angles this year, from Starcloud's pitch that orbital data centers sidestep terrestrial cooling and power constraints entirely to Apollo Atomics' bet on compact nuclear reactors built specifically to power data centers.
Dave Cote's operational playbook
Cote is best known for his long tenure turning around Honeywell, where he built a reputation for operational discipline and margin improvement at an industrial conglomerate that predated him by decades. Bringing that same playbook to Vertiv -- a company most generalist investors hadn't previously associated with the AI boom at all -- positioned Vertiv as a company benefiting directly from AI-driven capital expenditure without carrying the same risk profile as a chipmaker exposed to model-architecture shifts or a memory producer exposed to commodity pricing cycles.
The "picks and shovels" investment thesis
Vertiv's stock performance reflects a broader rotation among investors toward AI infrastructure companies supplying the physical building blocks -- cooling, power distribution, networking equipment -- that benefit from AI capital spending regardless of which specific AI lab, model architecture, or chip vendor ultimately wins the underlying competition. That's a structurally different risk profile than betting directly on Nvidia, Anthropic or OpenAI: a cooling-equipment company gets paid whether a hyperscaler is building capacity for training runs or inference workloads, and whether the chips inside that data center come from Nvidia, AMD, or a custom ASIC.
The counterweight
Vertiv's re-rating as an AI infrastructure darling still depends entirely on the underlying AI capex cycle continuing at its current pace -- a company selling cooling and power equipment for data centers is exposed to a slowdown in AI infrastructure spending just as directly as a chipmaker would be, even if it's diversified across which specific AI company is doing the spending. An eighty-year operating history and a well-regarded CEO are real assets, but they don't insulate Vertiv from a broader pullback in AI data-center construction if capital expenditure growth decelerates faster than the market currently expects, and "picks and shovels" investing has historically still meant riding the same boom-and-bust cycle as the underlying resource rush, just with somewhat less company-specific risk.
