■ HIGH RISK ■ Construction & Trades
The shop half of structural steel — cutting, drilling, welding — is heavily robotic already. The field half, connecting iron in the wind, resists hard; the squeeze is that ever more of the work is being moved from the field into the robot-friendly shop.
“Welding robots don't get dizzy walking steel beams 40 stories up.”
Our AI replacement risk score — how we score jobs
Ironworkers split into overlapping specialties: structural workers who connect columns and beams as buildings rise, rodbusters who tie rebar in mats and cages, riggers who move absurd loads with cranes, and welders who join it all. Field days mean climbing steel, walking beams, setting deck, bolting connections, and guiding crane picks by radio — dangerous, weather-exposed, physically brutal work with a strong union culture and apprenticeship system.
The automation story here is geographic: it happens indoors first. Fabrication shops that supply job sites now run robotic beam lines that cut, drill, and cope steel from the 3D model, and robotic welding cells handle repetitive assemblies. Rebar tying robots exist and work — semi-autonomous machines crawl across mats tying intersections — and rebar cages increasingly arrive prefabricated. The deeper trend is design-for-manufacture: BIM-driven projects shift labor from field connections to shop assemblies, modular construction moves whole steel modules indoors, and every hour moved into the shop is an hour a robot can bid on. Field robotics is harder — demos of autonomous welding crawlers and layout robots are real but marginal — yet drones already do the inspection climbs, and crane automation is inching forward. Our 65 score prices the shop displacement plus the shrinking field share, not robots walking beams.
What resists: the connection gang. Setting iron in wind, plumbing a structure, improvising when steel arrives an inch wrong, rigging one-off picks — this is high-consequence adaptive work in an unstructured environment, which is robotics' worst case. The apprenticeship pipeline, union density, and a national welder shortage all buffer wages. And infrastructure spending keeps demand strong. The trade's future is fewer total hours per ton of steel, with the remaining hours concentrated in the field skills machines can't touch.
Automatability: our editorial assessment of current and near-term AI capability
Shop-side robotics is mainstream now and deepening with each fabricator's capital cycle; rebar automation and prefab are expanding this decade. Field ironwork faces no near-term robot replacement, but its share of total hours shrinks as design pushes work into shops and modules. Expect serious pressure on shop and rodbuster roles by 2030, while structural connection and rigging skills stay scarce and well-paid past that horizon.
In the shop, extensively — robotic beam lines and welding cells already fabricate much of the steel that reaches a job site, and rebar-tying robots handle flat mats. In the field, barely: connecting iron at height in wind, rigging odd loads, and fixing misfit steel remain firmly human. The trade's exposure comes from work migrating shopward, where the robots are.
No — but it's a consolidating one. Total labor hours per ton of erected steel keep falling as fabrication automates and prefabrication grows, even while infrastructure spending keeps crews busy. Field ironworkers with connection, rigging, and certified-welding skills face strong demand and no plausible robot substitute this decade. The vulnerable positions are repetitive shop welding and flat-deck rebar tying.
The scary ones. Structural connecting at height, crane rigging and signaling, plumbing and aligning steel, and field welds in awkward positions all involve unstructured environments and improvisation — robotics' hardest problems. Add digital layout skills and blueprint/BIM fluency and you're the person coordinating between the model and the iron, which is the last role automated on any site.
More than their field counterparts, yes. Robotic welding cells thrive on exactly what shops produce — repetitive joints, controlled positioning, high volume — and every fabricator's next capital purchase includes more of them. Shop welders should chase certifications robots don't hold: field welding, critical and inspected connections, exotic materials. Or follow the work outdoors, where the wind keeps the robots away.