Why Vertiv's Bigger Growth Target Points to the Power Path, Not the Chip

When Vertiv stepped onto Goldman Sachs' technology conference stage in early September, the company behind the cooling and power-delivery gear inside AI data centers told investors something it had not said a year earlier: it now expects its own revenue to compound at 20%–22% a year through 2030, up from a prior target of 12%–14%. It is a striking upgrade, and it reads as management saying the demand wave it sees is larger — and more global, not just American — than it assumed in 2024.

To understand why that single number matters, it helps to locate Vertiv in the AI buildout. The chips get the headlines, but an AI data center is, physically, an enormous electrical problem. Conventional racks might draw on the order of 10 kilowatts; high-density AI pods run past 140 kilowatts, and a single large training facility can approach a gigawatt of demand140 kilowatts. Power has to be transformed from the grid, distributed, converted, and protected before it ever reaches a GPU — and then the waste heat has to be pulled away. The GPU is the obvious bottleneck. The power-and-cooling path beneath it is where the build actually stalls. Industry reporting this year found more than half of planned U.S. data centers may slip because transformers, switchgear, and related electrical equipment are scarcemore than half of planned U.S. data centers.

Vertiv is among the purest public ways to own that layer: data centers are essentially the whole business. The numbers support the framing. Backlog stood near $15 billion as of late 2025$15 billion backlog. In the second quarter of 2026, net sales rose 24% to $3.27 billion, adjusted operating margin expanded to 22.6%, and adjusted earnings per share grew 60%net sales rose 24% to $3.27 billion. Return on invested capital is a healthy ~26%, which is what a provider with pricing power and a long, qualification-laden customer relationship typically earns.

The more interesting part is where Vertiv is aiming next. At the same conference, management laid out a roadmap to move up the power stack: an 800-volt powertrain is on track, with accelerating "sidecar" products in late 2027 and full powertrain volumes in 2028. And just days before the conference, on September 2, Vertiv agreed to buy Utility Innovation Group for roughly $1.45 billion in cash (up to about $2.6 billion with performance targets)$1.45 billion in cash — a maker of microgrid controls, onsite-generation orchestration, and behind-the-meter power architecture. That is Vertiv deliberately inserting itself at the tightest point of the constraint: getting electricity from the grid (or an onsite plant) behind the meter and into the rack, when grid capacity and grid-connected equipment are the industry's scarcest input.

That is a structural point worth separating from a hopeful one. A chokepoint is a dependency that is concentrated, hard to replace quickly, and economically decisive. Vertiv is not a monopoly — it competes with Eaton, Schneider Electric, ABB, and others across power and cooling, and its own complex "converged" products (OneCore, SmartRun) have their own supply-chain congestion. What Vertiv does have is qualification and integration: critical-infrastructure buyers cannot tolerate failure in this equipment, so switching is slow even when a cheaper rival appears. The move toward powertrain architecture and behind-the-meter power is best read as Vertiv trying to move itself closer to the genuinely scarce node rather than resting on components a competitor could copy.

Which brings up the harder question, the one the conference does not settle: price. Vertiv's shares have roughly doubled over the past year and are up about 80% year to date. At a market value near $112 billion, the stock trades at roughly 44 times trailing EV/EBITDA and about 65 times trailing earnings. A genuine 20%–22% compounder can justify a premium — but a multiple that high already discounts a great deal of flawless execution over several years. The raised growth target is better evidence about the strength of the demand map than it is about the cheapness of the entry point.

For an investor framing the decision, the separation matters. The relationship is real: AI computing cannot scale without the power-and-cooling layer, and Vertiv is one of the cleanest, best-qualified providers of it, now pushing into the exact node where the industry is bottlenecked. What remains genuinely open is whether the market has already paid for the plan. Watch whether Vertiv converts its backlog into cash and margins at the pace the stock implies, whether competitor capacity arrives faster than expected, and whether ownership has crowded in after this run. A real structural position and a demanding valuation can be true at the same time — and they often are.