Not Chips. Not Memory. Data Center Cooling Is the Unsung AI Infrastructure Play Nobody's Watching

Vertiv reported Q2 revenue growth of 24% year over year to $3.27 billion and adjusted EPS growth of 60% to $1.52, while raising full-year targets to $14 billion in revenue (+37%) and $6.70 in EPS (+60%). Its $15 billion backlog and 2.9x Q4 2025 book-to-bill ratio support continued demand for AI data-center cooling infrastructure. Although shares are up 50% in 2026, they have fallen 35% from their May 14 high; the article argues the resulting 28x forward P/E offers an attractive long-term entry point, projecting potential 120% upside to $537 by 2030.

Analysis

VRT’s investment case is increasingly an execution-and-capacity story rather than a simple AI-demand proxy. The key earnings sensitivity is conversion of high-density cooling orders into revenue without expedited freight, component shortages, or field-installation bottlenecks eroding gross margin; this is where VRT’s integrated power-plus-thermal offering can take share from narrower cooling vendors such as JCI and private Munters. A sustained shift toward liquid cooling also raises customer switching costs after design-in, supporting aftermarket service revenue and potentially a higher through-cycle multiple.

The near-term setup is less clean than the growth narrative implies. A large backlog provides visibility but also embeds hyperscaler capex concentration, cancellation/deferment risk, and multi-quarter delivery exposure; investors should distinguish bookings from funded, construction-ready projects. Over the next 1-3 months, order growth, backlog mix, lead-time normalization, and gross-margin guidance matter more than another headline AI demand statistic. A miss on conversion or evidence that customers are standardizing lower-cost white-space cooling would likely compress VRT faster than revenue estimates decline.

Contrarian view: the relevant constraint may migrate from accelerator availability to utility interconnects and data-center commissioning. That delays VRT revenue recognition but does not necessarily destroy demand, creating volatility around quarterly results and an opportunity if bookings remain intact while shipments slip. Over 6-18 months, grid-power scarcity favors suppliers with power-distribution systems alongside thermal equipment; VRT is better positioned than pure-play cooling exposure, but valuation support requires continued margin expansion rather than just top-line growth.

Key Decisions for Investors

• Accumulate VRT in tranches over the next 1-3 months only after confirming book-to-bill remains above 1.2x and management maintains gross-margin/FCF conversion guidance; target a 15-20% position upside on earnings revision and multiple stabilization, with thesis invalidated by two consecutive quarters of backlog decline or a material margin-guide cut.
• Prefer a relative-value long VRT / short JCI basket over an outright AI-infrastructure long for 6-12 months. VRT has greater exposure to high-density, liquid-cooled deployments, while JCI retains more conventional commercial-building exposure; exit if liquid-cooling attach rates fail to rise or JCI demonstrates comparable data-center order acceleration.
• Use post-earnings weakness rather than pre-earnings calls: implied volatility can be justified by backlog-conversion uncertainty, and the missing data are customer concentration, cancellation provisions, and liquid-cooling mix. Establish an alert for a guidance reduction tied to project commissioning delays rather than demand cancellation; the former is more likely a buyable dislocation.
• Avoid treating NVDA strength as a sufficient confirmation signal for VRT. Monitor hyperscaler capex guidance, utility-interconnect timelines, and data-center construction starts; a widening gap between accelerator procurement and facility commissioning is a 2027 revenue-timing risk for VRT despite intact long-term demand.