5 Great AI Infrastructure Stocks to Buy on a Dip

It's no secret that stocks with exposure to the artificial intelligence (AI) sector have dipped recently. It's hardly surprising, given that no high-growth industries ever grow linearly. Still, the long-term growth outlook remains highly positive, and all five of these companies have reported improving conditions even as their stock prices have declined over the last three months.

All of which suggests now is a good time to start thinking about buying AI infrastructure stocks like GE Vernova (NYSE: GEV), Vertiv (NYSE: VRT), ON Semiconductor (NASDAQ: ON), Comfort Systems USA (NYSE: FIX), and SPX Technologies (NYSE: SPXC).

GE Vernova is building long-term recurring revenue

The former General Electric company's core business is gas power turbines from which it generates long-term, high-margin service revenue under long-term service agreements (LTSAs) attached to equipment sales. These turbines power AI data centers and support the increased electricity demand from data centers.

ON data by YCharts.

The good news from the company's latest earnings report is that its momentum appears to be accelerating rather than slowing. This dynamic is clearly evident in its outlook for backlog through 2026. Having started the year expecting a $200 billion backlog in 2028, management updated the target to 2027 in April.

Fast-forward to an investor conference where GE Vernova CEO Scot Strazik said it would hit the target "very early" in 2027, citing the strength of its orders. Ultimately, the backlog hit $197 billion at the end of the third quarter and is likely to be above $200 billion as I write.

Given that every gas turbine installed, notably its heavy-duty turbines, can generate decades of service revenue, the backlog growth implies a significant ramp in future long-term revenue.

Vertiv is a picks-and-shovels play on AI data centers

Vertiv's power and cooling solutions are an integral part of AI data center spending, and, as with GE Vernova, there's no sign of a slowdown yet. In fact, management has increased its full-year revenue and earnings guidance on every earnings call this year.

CEO Giordano Albertazzi said, "Data center demand is accelerating despite pacing items like power availability," and investors should note that as Vertiv's revenue guidance has increased, so has its margin guidance, implying a leveraging of revenue growth into extra profit growth.

Data source: Vertiv presentations.

ON Semiconductor is an overlooked AI stock

Alongside Vertiv, ON Semiconductor is an Nvidia partner in developing architecture for the next generation of 800 Volt (V) high-voltage direct current (HVDC) data centers. It's a position that emphasizes the growing importance of the company's power and sensing technology to the AI market. It's best known for its power and sensing chips used in electric vehicles (EVs) and industrial applications (EV charging, automation, etc.). Those are attractive end markets in themselves; however, its fastest-growing market is its power chips for AI data centers.

Management recently raised its revenue estimate for the data center end market to $620 million, up from $500 million. For reference, management expects total revenue of $6.6 billion in 2026.

Moreover, management expects its AI data center revenue to double in 2027 (implying at least $1.24 billion in revenue) and to increase by at least $3 billion in 2030, partly driven by higher content value on the new 800V HVDC data centers set to launch in 2027. The stock remains my top pick for 2026.

Comfort Systems USA

This mechanical and electrical systems contractor is another company that's seen a remarkable increase in its backlog in recent years, driven by growth in its technology end market (59% of year-to-date sales) and, in particular, demand for contracting services in data centers.

Data source: Comfort Systems USA presentations.

There are no signs of any slowdown in place. When discussing its data center customers on the recent earnings call, the company's President Trent McKenna outlined that "our customers are committing earlier and for larger scopes because they know that capacity like ours is scarce right now. And we continue to be really selective in what we're looking at."

This outlook mirrors what GE Vernova and Vertiv are saying, and frankly, it's likely to do so as GE Vernova, ON Semiconductor, and Vertiv tend to build their backlogs (and revenue) earlier in the data center spending cycle. As such, investors can expect more growth for Comfort Systems down the line, based on what the others are saying.

Image source: Getty Images.

SPX is an often overlooked AI infrastructure stock

AI data centers need power (GE Vernova), power chips (ON Semiconductor), and power distribution equipment (Vertiv), as well as contractors (Comfort Systems) to install the equipment. They also need cooling solutions provided by heating, ventilation, and air conditioning (HVAC) companies. In addition, data centers use cooling towers to remove heat from the data center and return the cooled water to the system.

That's where SPX Technologies' cooling towers come in. Not only is the company a beneficiary of data center spending, but it's also a beneficiary of increasing scrutiny of how data center water reuse is mandated.

Lee Samaha has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Comfort Systems USA, GE Vernova, Nvidia, and Vertiv. The Motley Fool recommends ON Semiconductor. The Motley Fool has a disclosure policy.