How bipedal machines could reshape one of the world's most dangerous — and most understaffed — industries.
Construction is enormous. Global spending is forecast to surpass $15 trillion in 2025, and the industry must build roughly 13,000 new structures every single day until 2050 to keep pace with urbanisation. Yet construction is also stuck. Productivity growth between 2000 and 2022 averaged just 0.4 percent annually — a fraction of what manufacturing achieved over the same period. The workforce is ageing, recruiting is difficult, and the work remains among the most physically punishing and dangerous of any civilian occupation.
This is the context into which humanoid robots are now being proposed — not as a novelty, but as a potentially serious answer to problems the industry has failed to solve through conventional means.
The Scale of the Problem
Labour Shortages
The construction industry is haemorrhaging workers faster than it can replace them. In the United States, 92 percent of construction firms report difficulty finding qualified workers, according to a 2025 survey by the Associated General Contractors of America. The industry needed an estimated 439,000 additional workers in 2025 alone just to keep pace with demand — and that figure assumed a slowdown in spending growth.
The picture is similar worldwide. In the UK, more than half of firms report difficulty hiring skilled tradespeople. In Japan, 60 percent of construction companies face labour shortages. In Australia, analysis has found that if shortages persist through 2028, over 13,000 new homes simply won't be built, with output declining by billions of dollars.
The root causes are structural. An estimated 41 percent of the current construction workforce is expected to retire by 2031. Younger workers are not entering the trades in sufficient numbers, put off by the physical demands, safety risks, and — until recently — comparatively lower wages. Immigration policy changes in several countries have further reduced the flow of available workers.
Safety
Construction remains one of the deadliest industries in the developed world. In the United States, 1,034 construction workers died on the job in 2024, accounting for roughly one in five of all workplace fatalities nationally. The fatality rate of 9.2 deaths per 100,000 full-time workers remains far above the national average. Falls, slips, and trips are the leading killer, responsible for about 38 percent of all construction fatalities. Transportation incidents — being struck by or caught between vehicles and equipment — account for a further 24 percent.
Beyond fatalities, the non-fatal injury toll is staggering. The industry recorded over 173,000 non-fatal injuries and illnesses in 2023 alone. Construction workers suffer elevated rates of hearing loss, musculoskeletal damage, and respiratory illness from prolonged exposure to noise, vibration, dust, and hazardous materials. The suicide rate among male construction workers is 75 percent higher than the national average.
These numbers point to a clear opportunity: if robots can take on the most dangerous and physically damaging tasks, the human cost of building could be substantially reduced.
Stagnant Productivity
While manufacturing has transformed itself through automation over the past two decades, construction has barely moved. Unlike a factory floor, a construction site is inherently chaotic — layouts change daily, multiple trades work in parallel, terrain is uneven, and conditions vary with the weather. This unpredictability has made the industry resistant to the kind of fixed automation that revolutionised manufacturing.
It is precisely this resistance to fixed automation that makes humanoid robots attractive. A humanoid, designed to operate in spaces built for human workers and to use human-scale tools, could theoretically adapt to the kind of unstructured, ever-changing environments that defeat purpose-built machines.
Robots Already on Site
Before considering where humanoid robots might fit, it's worth understanding what robotic technology is already operating on construction sites today. This is not a standing start.
Task-specific robots are already delivering measurable results. Dusty Robotics produces a machine that performs layout work — marking lines and points on floors — reportedly six times faster than traditional manual methods. Construction Robotics developed SAM (Semi-Automated Mason), a bricklaying robot that works alongside human masons. Canvas has deployed drywall finishing robots on active job sites. Automated plastering systems from companies like OKIBO are scaling across Europe and the United States.
Quadruped robots — most notably Boston Dynamics' Spot — are now a relatively common sight on major construction projects. Spot is used for progress monitoring, laser scanning, and site inspection, traversing terrain that wheeled robots cannot manage. Major firms including Turner Construction, Foster + Partners, Skanska, and BAM Nuttall have deployed Spot on active sites for tasks ranging from 3D scanning and digital twin creation to housekeeping audits and remote inspection of hazardous areas.
These existing deployments are important because they are establishing the workflows, safety protocols, and operational trust that humanoid robots will need to build upon. The industry is not waiting for humanoids to arrive before learning how to work alongside machines.
Where Humanoid Robots Could Fit
The case for humanoid robots in construction rests on a simple insight: construction sites are designed for human-shaped workers. Doorways, staircases, scaffolding, ladders, corridors, and the tools used on site are all scaled for the human body. A robot that shares that form factor can, in theory, navigate these spaces and use these tools without the site being redesigned to accommodate it.
A 2025 McKinsey report on humanoid robots in construction identified several categories of work where humanoids could eventually contribute, roughly ordered by how soon deployment might be feasible.
Near-Term Candidates
Material handling and logistics. Moving heavy or bulky materials around a job site — carrying bags of cement, shifting timber, transporting tools and supplies between floors. This is repetitive, physically demanding work that accounts for a significant share of total labour hours and is a leading cause of musculoskeletal injuries. McKinsey's analysis suggests material handling could represent up to 25 percent of automatable labour hours.
Site preparation and cleanup. Clearing debris, preparing tools, organising work areas, and general housekeeping. These tasks are essential but low-skill, and keeping sites clean directly reduces the risk of trips, falls, and other accidents.
Simple finishing work. Painting, plastering, and other repetitive interior finishing tasks in structured environments — particularly in projects with repeating layouts, such as high-rise residential buildings. McKinsey estimates finishing work could account for about 15 percent of automatable hours.
Medium-Term Possibilities
Inspection and quality assurance. Humanoids equipped with cameras, LiDAR, and other sensors could perform routine quality inspections, comparing as-built conditions against BIM models and flagging defects or deviations in real time. Quadruped robots are already doing versions of this work; humanoids could extend it to spaces that require reaching, climbing, or operating at height.
Demolition and hazardous task substitution. Stripping interiors, removing hazardous materials like asbestos, or working in environments contaminated by chemicals, dust, or noise at levels dangerous to human health. The value proposition here is straightforward: remove humans from harm.
Longer-Term Ambitions
Skilled trade work. Welding, plumbing, electrical installation, and structural work demand a level of dexterity, judgment, and problem-solving that remains well beyond the capabilities of current humanoid robots. These tasks require not just fine motor control but the ability to adapt to constantly changing conditions, improvise solutions to unexpected problems, and exercise the kind of physical intuition that experienced tradespeople develop over years. This is the most ambitious end of the spectrum and the furthest from practical deployment.
The Obstacles
Despite the clear opportunity, significant barriers remain between today's humanoid robot prototypes and useful deployment on construction sites.
Mobility and Durability
A construction site is one of the most challenging environments for any robot. Uneven ground, rubble, mud, standing water, stairs, scaffolding, and confined spaces all demand robust locomotion. Robots will need to navigate these conditions reliably across an entire shift — not just in a controlled demonstration. Current humanoid platforms are improving rapidly in controlled settings, but active construction sites present a different order of difficulty.
Dexterity
Many of the most valuable construction tasks require human-level hand dexterity — gripping irregularly shaped objects, operating hand tools, threading cables through tight spaces, making fine adjustments. This remains one of the most significant technical gaps in humanoid robotics. Current platforms can handle gross manipulation (lifting, carrying, placing) far more capably than the fine motor tasks that define much of skilled construction work.
Battery Life and Uptime
Today's humanoid robots typically offer one to two hours of useful operation on a single charge, depending on the intensity of the task. Construction shifts run eight to twelve hours. For humanoids to be viable, they will need either dramatically improved battery life, rapid battery swapping capabilities, or fast on-site charging — ideally during natural work breaks.
Cost
Current humanoid robots cost between $150,000 and $500,000 per unit. McKinsey estimates that for widespread adoption in construction, unit costs will need to fall to between $20,000 and $50,000 — a price point at which robots become competitive with local labour costs. This is a substantial reduction, though the trajectory of cost declines in robotics and AI hardware gives reason to believe it may be achievable within the next decade.
Site Infrastructure
Robots will require reliable connectivity (5G or equivalent), digital site models to navigate against, and charging infrastructure. Most construction sites lack this digital backbone today. Building it represents an additional investment that contractors will need to plan for.
Regulation and Liability
There are currently no established safety standards or regulatory frameworks specifically governing the use of humanoid robots on construction sites. Questions of liability — who is responsible when a robot causes damage or injury? — remain largely unresolved. Developing certification processes and insurance frameworks will be essential before large-scale deployment can occur.
Heavy Industry Beyond Construction
The potential applications of humanoid robots extend across heavy industry more broadly.
Shipbuilding and offshore. Shipyards and offshore platforms present many of the same challenges as construction — confined spaces, heavy lifting, hazardous conditions, and skilled labour shortages. Welding, inspection, and material handling in these settings are strong candidates for humanoid assistance.
Mining. Underground and open-pit mining operations involve significant safety risks and increasingly struggle to attract workers. Robots that can navigate tunnels, operate equipment, and perform inspections in environments too dangerous for regular human presence could prove valuable.
Infrastructure maintenance. Bridges, tunnels, dams, power stations, and other critical infrastructure require regular inspection and maintenance, often in difficult-to-access locations. Humanoid robots could extend the reach and frequency of inspection programmes while reducing the risk to human workers.
What the Industry Is Doing Now
While full humanoid deployment on construction sites remains years away, the industry is not standing still. McKinsey recommends that construction firms begin preparing now — identifying tasks where humanoids could fill gaps, building digital infrastructure, and developing partnerships with robotics companies.
Several practical steps are already underway across the industry. General contractors are partnering with robotics firms to pilot early-stage deployments. Companies are investing in "innovation gym" environments — controlled test sites where robotic technologies can be evaluated before being brought to active projects. Tele-operation trials, where human operators remotely control robots to perform tasks, are serving as an intermediate step toward greater autonomy.
Contractor attitudes toward robotics have shifted markedly. According to the 2025 Equipment and Robotics Benchmarking report from BuiltWorlds, positive evaluations of robotics technology among construction firms rose from 74 percent in 2024 to over 95 percent in 2025 — a striking change in an industry not known for rapid technology adoption.
A Realistic Timeline
Large-scale humanoid deployment on construction sites is not imminent. Most analysts place meaningful, widespread adoption roughly a decade away. A realistic progression might look something like this:
Now through 2027: Task-specific robots continue to expand. Quadrupeds like Spot become standard on large projects. Early humanoid prototypes begin tele-operated pilot deployments in controlled construction environments, focused on simple material handling and logistics tasks.
2027 to 2030: Semi-autonomous humanoid deployments in structured settings — repetitive interior work in high-rise buildings, warehouse construction, prefabricated modular builds. Costs begin to decline as manufacturing scales.
2030 to 2035: Broader deployment across a wider range of construction tasks. Humanoids handle a growing share of material handling, finishing, inspection, and hazardous work. McKinsey estimates humanoids could handle up to 30 percent of labour tasks by 2035.
Beyond 2035: As dexterity, AI reasoning, and battery technology mature, humanoids begin to take on more complex tasks previously reserved for skilled tradespeople, working alongside human crews as genuinely capable collaborators rather than simple assistants.
The Human Question
Any discussion of robots in construction inevitably raises the question of jobs. Will humanoid robots replace construction workers?
The more accurate framing, at least for the foreseeable future, is augmentation rather than replacement. The industry has a severe labour shortage — it does not have a surplus of workers looking for something to do. Humanoid robots, if they work as intended, would be filling gaps that humans are not filling, taking on tasks that are too dangerous, too repetitive, or too physically punishing for human workers to sustain over a career.
McKinsey and other analysts stress that humanoids will initially function as assistants, with humans continuing to supervise, manage, and perform the complex reasoning and skilled work that robots cannot yet approach. The introduction of robots will also create new roles — robot operators, maintenance technicians, fleet managers, and integration specialists — just as every previous wave of construction technology has done.
That said, the long-term trajectory is worth watching carefully. If costs fall far enough and capabilities improve fast enough, the economics could shift in ways that do affect employment for certain categories of work. This makes it important that the industry, policymakers, and unions engage with these questions early — shaping how the technology is deployed rather than reacting after the fact.
The Bottom Line
Construction and heavy industry sit at the intersection of everything that makes humanoid robots compelling: a severe labour shortage, dangerous working conditions, stagnant productivity, and environments literally built for the human form. No other sector presents quite the same combination of urgent need and natural fit.
The technology is not ready for prime time — not yet. But the direction of travel is clear, and the pace of progress in AI, robotics, and battery technology is accelerating. Construction firms that begin planning for a future alongside humanoid robots now will be better positioned to benefit when that future arrives. Those that wait may find themselves scrambling to catch up.
For an industry that has spent decades struggling with the same fundamental problems, humanoid robots represent something genuinely rare: a new category of solution. Whether they fulfil their promise will depend not just on the machines themselves, but on how thoughtfully the industry prepares to receive them.
This article is part of Droid Brief's Resources section — a comprehensive reference library covering humanoid robotics from foundations to frontiers.
Related reading: