The AI Boom’s Next Big Winners May Wear Hard Hats

Key Points

• EMCOR Group has grown its frontline workforce from about 34,000 to 44,000 as data center electrical and mechanical demand lifted its backlog toward $17 billion
• Dycom Industries lays the fiber that connects data centers and has moved into power work, yet its stock has sold off with other AI infrastructure names
• United Rentals supplies the equipment behind every buildout, and Altimetry argues the market is pricing its earnings as roughly flat

The scarcest input in the AI buildout may not be a chip or a power contract. It may be a licensed electrician.

Rob Spivey, director of research at Altimetry, calls this group of stocks the blue-collar millionaire plays. As hyperscaler spending climbs, the companies that employ skilled trades, and supply the equipment those workers use, could capture more of the AI payoff than the market is currently pricing in.

Why Skilled Trades Are Becoming the AI Bottleneck

Spivey sees two education trends moving in opposite directions. The number of college-age Americans has peaked, and he argues the wage premium for many degrees has shrunk just as tuition reached record highs. Entry-level white-collar roles in fields like computer science are also among the most exposed to AI.

Vocational training is heading the other way. Spivey points to roughly 50% enrollment growth since 2022 in programs for welders, electricians, HVAC technicians, and plumbers.

Demand explains the shift. Microsoft NASDAQ: MSFT President Brad Smith wrote in 2025 that the United States may need half a million new electricians over the next decade. Spivey argues that the estimate came before the latest jump in spending. He cites hyperscaler capital expenditures (CapEx) of about $800 billion this year, with forecasts of $1.1 trillion and then $1.2 trillion in the years ahead. At that pace, he believes the need could be closer to a million workers.

When something gets scarce, it gets expensive. That's good news for tradespeople and for the companies that can recruit, train, and keep them.

EMCOR Turns Workforce Investment Into Backlog Growth

EMCOR Group NYSE: EME handles the electrical and mechanical work inside complex projects, typically working under a general contractor. Spivey estimates about half its business is now tied to data centers, with the rest spread across high-tech manufacturing, industrial and healthcare.

The demand shows up clearly. Remaining performance obligations hit a record $17.14 billion at the end of June, up nearly 44% from a year earlier, with data center contracts a major driver.

What stands out to Spivey is how management talks about the business. Workforce comes first on earnings calls. He notes frontline headcount has grown from about 34,000 to 44,000 in recent years, supported by recruiting and apprenticeship programs. In his view, that labor pool is the moat. More trained workers means more projects EMCOR can take on.

The market's biggest question is durability once data centers are built. Spivey's answer centers on hyperscaler economics. He projects their earnings could reach roughly $1 trillion a year by 2030 on his uniform accounting basis, which would make today's spending pay back quickly and give companies a reason to keep investing.

Dycom Rides Fiber, Broadband, and Data Center Power

Dycom Industries NYSE: DY is the company telecom carriers call to lay fiber. Spivey sees three tailwinds: government-funded rural broadband, the need to replace aging fiber that can't handle today's speeds, and the acquisition of Power Solutions, a roughly $1.95 billion deal that moved Dycom into electrical work for data centers in the Washington, D.C., region.

Dycom also isn't exposed to product obsolescence the way a manufacturer like Corning NYSE: GLW can be. New technology still has to go into the ground.

The chart has been rough. Shares have pulled back sharply in recent months, and Spivey ties much of that to broader worry across AI infrastructure as local opposition to data centers becomes an election issue. Company results played a role too: the stock fell after its latest quarterly report on concerns about communications margins and a shift in the timing of wireless revenue.

Political pushback may change where data centers get built, not whether they get built. Spivey expects more projects in power-rich, land-rich areas like Texas, which would mean longer fiber runs and more work for Dycom.

United Rentals Supplies the Equipment Behind the Buildout

Every trench, foundation and power line needs lifts, forklifts, backhoes, and earth-moving equipment. United Rentals NYSE: URI rents all of it.

The company also creates blue-collar millionaires directly. It's an active acquirer of family-owned rental businesses, turning decades of hard work into a payout for the owners.

Spivey argues the market is pricing United Rentals' earnings as roughly flat. His uniform accounting points to the 2012 acquisition of RSC Holdings, which brought in a higher-return specialty rental model. On his adjusted basis, return on assets climbed from about 6% to 7% to around 12%, an improvement that the reported figures obscured. That mix of specialty services and scale sets it apart from rivals like Herc Holdings NYSE: HRI.

The demand isn't only about AI, either. Reshoring has fueled steady construction of factories, warehouses, and logistics hubs since the pandemic.

Weighing the Upside Against the Risk for AI Infrastructure Stocks

The upside is a multiyear spending cycle meeting a real labor shortage. Companies that have already built their workforce could have pricing power that competitors can't easily match.

The risk is that sentiment around AI CapEx can shift fast, and political or permitting delays could push out project timelines. Execution on acquisitions and margins still matters, as Dycom's recent quarter showed.

Stay focused on backlog and headcount—because in this buildout, skilled hands are what turn CapEx into revenue.

Altimetry has also been digging into SpaceX NASDAQ: SPCX as an AI company, and the public stocks are positioned around that story. Get Altimetry's special report on the SpaceX AI opportunity to see the names on the research team's radar.