The US is building solar farms, data centres, and battery storage at a historic pace. There aren’t enough workers. Xpanner thinks the answer is turning the equipment contractors already own into robots.
Drive past any utility-scale solar farm breaking ground in America right now, and you’ll see the same scene: a pile driver working the field, not enough people around it to hit the production schedule. The job that once required a crew of five is getting done with two. Sometimes one.
That gap isn’t a planning failure. It’s the new normal.
The US construction industry needs an estimated 349,000 net new workers in 2026 just to keep pace with demand, according to Associated Builders and Contractors, down from 439,000 the year before, but still a structural shortfall that no recruiter can paper over. Meanwhile, the five largest US technology companies have committed between $660 billion and $700 billion in capital expenditure this year, largely for AI-driven infrastructure. Solar farms, data centres, and battery storage facilities are all breaking ground simultaneously, competing for the same shrinking pool of skilled workers. The scale of data centre construction alone has ballooned: sites that once peaked at a few hundred workers are now reported by contractors to require crews numbering in the thousands.
“You cannot recruit your way out of a structural shortfall of hundreds of thousands of people,” says Henri Lee, CEO and co-founder of Xpanner. “You have to change what each person on a jobsite can actually do.”
Xpanner, a California-headquartered startup founded in South Korea in 2020, just raised $18 million in Series B bridge funding to do exactly that. The round was led by existing backers Korea Investment Partners and KB Investment, bringing total funding to $38 million. The capital will accelerate product development and expand Xpanner’s subscription automation platform across the US and into international markets.
Same machine, smarter jobsite
The construction robotics space has attracted serious capital. Built Robotics, SafeAI, and autonomous divisions at Caterpillar and Komatsu have all pursued the same thesis: build a new autonomous machine, run pilots, convince contractors to replace their fleets. Commercial scale has remained elusive.
Xpanner took a different route. Its flagship X1 Kit is not a robot; it is a retrofit. A hardware control unit and an AI processor that attaches to a pile driver a contractor already owns, turning it into what Xpanner calls Software-Defined Machinery. No new equipment purchase. No rip-and-replace. Contractors pay a subscription for a specific automation licence (currently piling and material handling), and their existing machine learns to position itself, drive piles to depth, and log quality data with minimal supervision. The company says deployments have significantly cut both labour requirements and operation time, though independent verification of specific performance figures is not yet publicly available.
The model produces software economics from hardware. Gross margins sit above 80%. “Once hardware is deployed, incremental subscription revenue flows at near-zero marginal costs,” says co-founder, CFO and CSO Ryan Park.
The traction
Xpanner entered the US in 2023, timing its arrival to the infrastructure wave triggered by the Inflation Reduction Act. Solar EPCs were suddenly breaking ground on projects they had planned for years, all at once, without the workers to staff them.
The company found customers fast. It has completed transactions with or entered active discussions with 19 of the top 20 US solar farm EPCs, including Mortenson, Black & Veatch, and QCells. Revenue grew from $3 million in 2023 to $7 million in 2024 to $21 million in 2025. In Q1 2026 alone, it generated $8 million in revenue and $1 million in EBIT, its first operating profit. It is targeting $60 million in ARR by year-end, with zero customer churn since the US launch.
“Few Physical AI companies in construction reach commercial traction and profitability this quickly,” says Sangjoon Park, Managing Director at Korea Investment Partners. “The shift to subscriptions is what turns a strong product into a truly scalable business.”
What comes next
Solar piling proved the model. Battery energy storage and AI data centre construction are the next targets, and the opportunity is expanding fast. The data centre construction pipeline across the US is substantial; ConstructConnect has tracked dozens of projects set to start in the coming months, valued at tens of billions of dollars, though specific figures shift as projects are announced and revised.
The construction industry has long been called the last major sector to be automated. The labour crisis is rewriting that story. When you cannot hire the workers you need, and the projects cannot wait, the question stops being whether to automate and becomes how fast.
Xpanner’s answer is straightforward: attach the software to the machine you already own, pay per task, and skip the large capital commitment. Unglamorous and effective. In a market facing a structural shortfall of hundreds of thousands of workers, that is exactly what contractors are looking for.