Xpanner Raises $18M Series B to Bring Physical AI to Construction Sites

Construction has spent the last 20 years pretending software alone would save it. The industry bought tablets, dashboards, drones, cloud platforms, scheduling tools, and enough SaaS subscriptions to make an accountant develop a twitch. Meanwhile, job sites still run on labor shortages, machine downtime, delays, and the eternal ritual of somebody yelling over diesel engines while another guy squints at a grading stake like it’s an ancient prophecy. That disconnect is why Xpanner matters.

The Santa Fe Springs-based construction automation company just raised $18M in Series B bridge funding led by Korea Investment Partners, with participation from KB Investment Co. The round brings Xpanner’s total funding to $38M and signals growing investor appetite for companies attacking the physical layer of AI adoption, not just the software layer. Xpanner is not building another AI assistant for office workers pretending to be “10x productive” while answering Slack messages at midnight. The company retrofits existing construction equipment with robotics and Physical AI systems designed to automate real-world construction workflows involving dirt-moving, piling, and site operations in environments where weather, labor shortages, and equipment costs make PowerPoint optimism disappear fast.

More importantly, Xpanner is expanding at a moment when energy infrastructure, AI data center construction, and labor scarcity are all colliding at once.

What Happened

Xpanner announced $18M in Series B bridge financing led by Korea Investment Partners, with returning participation from KB Investment Co. The funding increases the company’s total capital raised to $38M since its founding in 2020. The company was originally established in South Korea before expanding into the U.S. market in 2023. Today, Xpanner operates out of Santa Fe Springs, California, positioning itself inside one of the largest infrastructure and construction markets in the world while maintaining deep ties to South Korea’s industrial and robotics ecosystem.

Xpanner’s leadership team reads less like a startup roster and more like a collision between heavy equipment manufacturing, industrial automation, and strategic finance. Henri Lee, CEO and co-founder, previously spent 14+ years at Bobcat and Hyundai Infracore driving global strategy and after-sales operations. CTO and co-founder David Shin spent 15 years at Volvo Construction Equipment leading automation and digitization development. Ryan Park, CFO/CSO and co-founder, brings 10+ years of strategic finance and M&A experience from Doosan Group and Shinhan.

That matters because construction technology has a long history of attracting software founders who have never stood on an active job site longer than it takes to shoot a product demo video. Xpanner’s founders came from the machinery side first, and that changes the psychology of the product.

Why Xpanner Matters

Most AI startups today live entirely inside screens, while Xpanner lives inside physics. The company’s flagship X1 Kit retrofits existing construction machines with Physical AI capabilities, allowing contractors to automate workflows without replacing entire fleets. That retrofit-first strategy is a major distinction in the construction automation market because fleet replacement is brutally expensive and operationally disruptive.

Construction companies do not swap heavy equipment the way startups swap project management software. A contractor might run the same machine for years because the economics demand it. That creates a giant opening for retrofit automation instead of full equipment replacement, and Xpanner recognized that reality early. The company positions its platform as “Automation-as-a-Service,” essentially bringing recurring software-style economics into one of the oldest industrial sectors on earth. Investors like that model because recurring revenue is predictable. Contractors like it because capital expenditures are painful enough already.

There is also a broader macro story underneath this funding round. The U.S. is entering an infrastructure cycle where solar projects, battery storage facilities, logistics expansion, semiconductor manufacturing, and AI data centers are all competing for labor simultaneously. Everybody wants skilled workers. Nobody can find enough of them. Construction automation stops sounding futuristic once labor shortages start delaying billion-dollar projects. Then it starts sounding inevitable.

Market Context

The construction industry sits in a strange place culturally. It is simultaneously one of the most economically important sectors in the world and one of the slowest to digitize at scale. Part of that hesitation is rational because construction sites are chaotic environments where dust, vibration, weather, shifting terrain, safety concerns, and fragmented subcontracting chains make software implementation much harder than Silicon Valley likes to admit.

It is easy to automate spreadsheets. It is much harder to automate a machine operating in uneven terrain at a live solar construction site in August heat. That is why robotics and Physical AI companies entering construction face a credibility test that SaaS companies rarely encounter. The technology either works in real-world conditions or it becomes another expensive science experiment hiding behind investor decks and conference panels. Xpanner’s positioning around field-proven automation is strategically important because the market has become increasingly skeptical of industrial AI claims without operational deployment evidence.

This is where the broader AI market is heading in general. The first wave of AI created digital productivity tools. The next wave is trying to automate physical infrastructure, industrial workflows, logistics, manufacturing, and construction. That transition is much harder technically, but also potentially far larger economically. Software changed office work. Physical AI changes labor economics.

Competitive Landscape

Xpanner enters a construction automation market that includes players across robotics, machine control, autonomy, and heavy-equipment software infrastructure. Companies like Built Robotics helped establish early market awareness around autonomous construction equipment. Traditional industrial giants including Caterpillar, Komatsu, and Volvo Construction Equipment continue investing heavily in automation and connected machinery capabilities, while enterprise infrastructure players such as Trimble dominate machine guidance and site intelligence systems.

Xpanner’s differentiation sits in its retrofit-first strategy combined with its focus on scalable automation across existing fleets and mixed equipment environments. That approach matters because construction sites rarely operate with standardized equipment fleets from a single manufacturer. Reality is messier than investor pitch decks. Contractors mix brands, machine vintages, attachments, and operational systems constantly.

Construction is not clean-room software infrastructure. It is operational improvisation with deadlines, and the companies that understand that tend to survive longer.

What This Signals

The deeper signal behind Xpanner’s funding round is that venture capital is increasingly shifting toward industrial AI infrastructure rather than pure software abstraction. For years, AI investment heavily favored consumer interfaces, SaaS copilots, content generation, and workflow optimization platforms. Those markets became crowded quickly because software distribution scales fast and barriers to entry collapsed.

Physical AI is different. Deploying robotics into active construction environments requires operational knowledge, deployment infrastructure, hardware integration, safety management, and industrial trust. Those barriers are painful, expensive, and slow, but they are also defensible.

That is exactly why investors continue backing companies capable of bridging AI software with physical infrastructure systems. Everybody wants AI. Very few companies can make AI survive contact with mud, steel, deadlines, and human operators who absolutely do not care about your product roadmap.