Up 250%, Still Underowned: Why Vertiv Remains AI's Best-Kept Data Center Bet

Heat, not chips, is becoming the limiting factor in AI builds

The real AI bottleneck is increasingly about getting heat out of the rack fast enough to keep compute running. That is why Vertiv matters now. The market has been chasing AI winners at the silicon layer, but extreme densification and gigawatt scaling are making power and cooling more central to how much AI work a data center can actually use.

That tension is visible in the stock. Shares have climbed over 250%, and the stock now trades at roughly 47 times forward earnings, which is rich for an industrial business. Bulls, however, see a real operating proof point: Q4 organic orders growth of 252%, a $15 billion backlog, and 2026 revenue guidance of $13.25 billion to $13.75 billion. In plain English, demand is already showing up.

Skeptics also have a fair point. A company priced for growth has less room for error. But this does not look like a hollow narrative trade. Order growth has been broad-based, and the product mix is shifting toward advanced liquid-cooling systems that can increase content per rack. That is the kind of change that can support a higher multiple if it holds.

The decision is simple: if hyperscaler spending stays narrowly focused on chips, VRT may look stretched. If spending keeps hitting thermal, power, and layout constraints, today's valuation may still prove justified.

Higher rack densities are making cooling a system-level purchase

Why the thermal mix is changing

Data centers are crossing a physical threshold. Once server racks run well above 20 kW, traditional air cooling became inadequate, and liquid cooling is being adopted as the primary thermal solution for the hottest rows. That changes the economics.

Early on, operators could solve problems with simpler fixes. As densities rise, the requirement becomes a full thermal chain: coolant distribution, pumps, heat rejection, and service. Vertiv is positioned across that chain, while advanced direct-to-chip liquid cooling is becoming the dominant architecture for the most power-hungry AI rows. That matters because system-level solutions can be stickier than bolt-on parts and can deepen customer dependence through design, commissioning, and maintenance.

Why Vertiv had time to prepare

This was not a last-minute pivot. Between 2021 and 2024, Vertiv built out partnerships and acquired coolant-distribution technology ahead of the sharper demand turn. That preparation matters because it lets the company sell more of the thermal stack, not just individual components.

If that mix shift continues, it helps explain why investors are willing to pay up. The opportunity is not just more demand; it is a better position in a higher-value part of the buildout.

Where a second rerating could come from

The underowned argument is now more about framing than discovery. Many investors still think of VRT as an industrial supplier rather than a core part of AI thermal infrastructure. If valuation starts to reflect that broader role, the multiple could expand again.

The execution watchpoint is straightforward: designs need to keep moving toward liquid cooling, and Vertiv needs to keep winning a large share of the system bill of materials.

Capacity is expanding, but expectations are rising faster

Capacity is expanding for a reason. Vertiv's Italy campus is set to double chiller production capacity in the region by the end of 2026, and a new large-scale testing laboratory is planned for early 2027. That suggests the bottleneck is shifting from design work to physical buildout.

The bull case: integrated cooling and uptime matter more

Bulls are betting that hotter racks make the full cooling system more valuable. As deployments move toward advanced liquid-cooling systems, Vertiv can sell integrated solutions rather than isolated components.

That is where reliability becomes part of the product. Vertiv says digital twin technology and adaptive liquid cooling can help advance AI deployment at scale. In practical terms, that means better forecasting, fewer surprises, and more confidence that critical workloads stay online.

The bear case: a backlog can slip

Bears make a practical argument. A backlog can become revenue, but it can also slip. Large data-center projects can be delayed by site validation, financing, power approvals, or slower AI demand downstream. That risk matters more because the stock has already rerated sharply and still trades at about 47 times forward earnings.

So the debate is no longer whether demand exists. It is whether demand will convert into revenue quickly enough to justify the current multiple.

What would confirm or challenge the thesis

Confirmation signals - Italy capacity comes online on schedule: double chiller production capacity by end-2026 and the large-scale testing laboratory by early-2027. - Liquid cooling continues to become the primary thermal solution for dense AI rows. - Vertiv's digital twin technology and adaptive liquid cooling become more visible in customer deployments.

Challenge signals - Project conversions lag even as capacity expands. - Power, financing, or approval delays push deployment timelines out. - The stock remains around 47 times forward earnings without another operating beat to support it.

Vertiv is not a speculative side bet. It is a critical supplier in a market that may still be underestimating how much AI value is shifting into power, cooling, and system integration.