Vertiv raised full-year guidance across the board and the stock still dropped double digits. The miss was the top line, not profits. Blip or the first crack?

Vertiv (VRT) is the unglamorous side of the AI buildout. They make the backup power, the power distribution, and the liquid and air cooling that every data center needs to actually run the GPUs. So the thesis kind of writes itself: if hyperscaler capex keeps climbing, Vertiv collects a slice of every new facility.

Q2 came out July 29 and it was a strange one. On profit they beat their own guidance, EPS $1.52 and adjusted operating profit $738M, both above the top of the range they'd set. Then they raised the full-year outlook again, revenue to at least $13.8B, operating profit to around $3.37B, free cash flow to $2.5B. Normally that's a green day.

The stock fell double digits anyway (somewhere around 10-17% depending which writeup you read). The reason was the one line that's supposed to be bulletproof here: revenue. Q2 sales landed at $3.27B, the bottom of their own guide and short of the \~$3.39B the street wanted. Management pinned it on "supply chain congestion and multi-phased project execution as deployments scale," basically timing on big lumpy builds.

Here's what I keep chewing on. The business quality isn't really in question. ROIC is around 30%, net debt is tiny (under half a turn of EBITDA), the free cash flow is real. The question is what you pay for it. Trailing P/E is up near 67x. Forward is lower, roughly 34x, because earnings are growing fast, but either way you're paying a premium for a company that just wobbled on the exact metric the premium is built on.

Two readings and I honestly go back and forth:

Take management at their word and the miss is timing. Demand is fine (they raised the year), the revenue just slid a quarter to the right because these AI builds keep getting bigger and harder to schedule. In that world a double-digit drop on a raised guide is the market handing you a discount on a picks-and-shovels compounder.

Don't take them at their word and the miss is the first sign that "scaling deployments" is harder to execute than a 34x forward multiple assumes. At that price revenue has to keep compounding 25-30% for years. A supply-chain excuse this quarter can quietly become a demand or margin excuse two quarters out, and a stock priced this richly doesn't forgive the second miss.

One more wrinkle. Eaton does a chunk of the same power-management work and trades cheaper, so part of the VRT premium is just the pure-play data-center exposure. That works both ways, more torque if AI capex holds, more pain if it cools.

No verdict from me, I'm trying to frame the bet. For those who follow it: do you take the "timing" explanation at face value, or does a top-line miss matter more than a profit beat when you're paying 34x forward? And what would actually change your mind, one more revenue miss, or the backlog number next quarter?

(Numbers from Vertiv's Q2 CY2026 release, July 29 2026, plus GuruFocus for the forward multiple. Information, not advice, I'm just pressure-testing the thesis so tell me what I'm getting wrong.)