AI’s data center buildout is turning into a power and plumbing trade, not just a chip story. McKinsey projects global data center investment could reach $7 trillion through 2030, while a separate S&P Global estimate says combined capital expenditures by six major hyperscalers could top $1.3 trillion by 2027. That spending is already rippling through the companies that make switchgear, cooling systems and electrical gear, with Schneider Electric, Vertiv and Eaton all posting sharp growth in the quarter ended June 30, 2026.
The scale is now so large that electricity itself is becoming the choke point. NEMA says data centers are expected to account for 38% of net U.S. electricity consumption growth through 2037. That matters because the AI boom is not just adding more buildings. It is forcing denser racks, heavier loads and new power architectures that require a different class of infrastructure. In this market, the winners are less likely to be the companies grabbing headlines and more likely to be the vendors quietly keeping the racks alive.
Schneider Electric’s North American sales rose 23% in the second quarter, helping drive record global revenue, according to the Yahoo Finance report. Vertiv lifted its full-year 2026 net sales growth outlook to 31%, with operating profit up 44% and adjusted operating profit up 51%. Eaton reported Electrical Americas orders up 41%, while its total electrical backlog rose 43%. Those are not soft signals. They point to a supply chain that is still running hot as hyperscalers and enterprise customers keep ordering the hardware needed to bring more AI capacity online.
The story’s central message is that the physical buildout has become urgent. McKinsey’s underlying report, The cost of compute: A $7 trillion race to scale data centers, says data centers will require $6.7 trillion worldwide by 2030, including $5.2 trillion for AI workloads and $1.5 trillion for traditional IT. The firm’s scenarios range from constrained momentum at $3.7 trillion, to a central case of about $5.2 trillion in AI capex, to an accelerated case of $7.9 trillion. However investors parse the math, the direction is clear: AI infrastructure spending is moving far beyond an ordinary equipment cycle.
Schneider sits in the middle of the bottleneck because it supplies the power distribution, switchgear and cooling systems that make high-density computing possible. Steve Carlini, chief advocate of data centers and AI at Schneider Electric, said, “It’s a great time to be in the data center business.” He added, “Ten years ago, data centers were much smaller. But now they all need—the big ones need—their own medium-voltage switchgear.” That is the kind of detail that shows how AI changes an industry from the ground up. More compute means more electricity. More electricity means more gear before a single model is trained.
Carlini said Schneider expects installed data center capacity to reach about 200 gigawatts globally by 2030, roughly double today’s approximately 100 gigawatts. McKinsey’s broader forecast is even more aggressive, saying global data center capacity could nearly triple by 2030 as hyperscalers and other technology companies race to add AI infrastructure. For suppliers, the implication is that the market is not just expanding; it is becoming more technically demanding. Dense AI systems need more than racks and network cables. They need power delivery that can keep up with the load.
The biggest change is happening at the rack level. Carlini said, “We’re going to see it almost doubling every year,” referring to expected rack density. He pointed to Nvidia’s infrastructure roadmap, where rack power is moving from roughly 227 kilowatts toward 400 kilowatts as newer AI systems arrive. That is an enormous jump in a short period, and it helps explain why traditional data center designs are under strain. Higher density means more heat, more electrical complexity and more pressure on the equipment that feeds each rack.
Carlini also said, “Once you get to the Rubin Ultra, the industry has to go into what’s called high voltage DC power distribution, which is 800 volts.” That matters because a move to 800 volts is not a minor tweak. It signals a structural shift in how data centers are powered. Schneider has already deployed more than four gigawatts of liquid-cooling capacity, but Carlini said that is set to increase rapidly now that Blackwell Ultras and Vera Rubins are all liquid-cooled. In other words, the cooling market is following the compute market upward almost in lockstep.
The opportunity is not limited to the manufacturers. The story says the surge is flowing through the reseller channel as well. Insight Enterprises reported a 65% increase in consolidated net earnings, and Connection posted a 33.8% increase in net income. That is important because it shows how the AI buildout is spreading beyond the obvious names. When demand moves from chips to power to cooling to deployment, distributors and integrators can benefit alongside the large industrial suppliers.
Carlini said the big global internet companies are not just offering long-range forecasts. “They give us not only a forecast, but they give us purchase orders,” he said. “The demand for these data centers is actually there and there is no speculation about who is going to go into these facilities.” That is the kind of comment investors like because it suggests the demand is already committed, not hypothetical. It also reinforces why supply access may matter as much as access to Nvidia chips.
The S&P Global figure that combined capex by Amazon, Microsoft, Alphabet, Meta, Oracle and SpaceX could exceed $1.3 trillion by 2027 underscores how much capital is still moving into the buildout. The inclusion of SpaceX is unusual, but the larger point remains that a handful of deep-pocketed companies are driving a huge share of demand. As long as those budgets stay elevated, suppliers of electrical gear, cooling and integration services should keep seeing strong order flow.
For now, the market is telling a consistent story: AI infrastructure is not slowing, and the limiting factor is increasingly power, not ambition. The next phase of the trade may belong to the firms that can deliver higher-voltage distribution, liquid cooling and enough electrical capacity to keep denser racks running. That is why the most important earnings in this boom may continue to come from the companies nobody outside the industry used to watch closely.
