■ HIGH RISK ■ Manufacturing & Production
Yes, for the dangerous repetitive core — pouring, molding, and grinding are exactly what foundries are automating first, with safety regulators cheering them on. The remaining jobs are technicians and metallurgical problem-solvers, not laborers at the lip of a ladle.
“Pouring molten metal is best done by something that can't feel pain.”
Our AI replacement risk score — how we score jobs
Foundry work is metal casting at industrial scale: preparing sand or investment molds, melting metal in induction or arc furnaces, transferring and pouring it at temperatures north of 2,000°F, then shakeout, and 'fettling' — cutting off gates and risers, grinding flash, and finishing castings. It is loud, hot, silica-dusted work with one of manufacturing's uglier injury and occupational-disease profiles, which is precisely why nobody sentimental is defending the status quo.
Automation here is driven as much by OSHA and insurance as by payroll math. Automated pouring systems meter molten metal into molds with a consistency no tired human matches; robotic mold-handling lines and core-setting arms run the sand side; and grinding-and-cutting robots handle fettling, the job segment with the worst vibration and dust exposure. Additive manufacturing is a second front: 3D-printed sand molds eliminate pattern-making steps, and some small metal parts skip casting for direct printing entirely. Machine-vision inspection catches casting defects faster than an experienced eye. Our 71 reflects a sector where every dirty, repetitive station has a commercially available robotic replacement — and where domestic foundries automate or lose bids to those that did.
What holds out is the craft-and-chaos layer. Foundries are old, cluttered, high-mix operations; a jobbing shop pouring fifty different castings a week can't tool a robot cell for each one, so low-volume and large-casting work stays manual longer. Melters and metallurgical techs who judge alloy chemistry, furnace operators managing temperamental equipment, maintenance crews keeping robots alive in an environment that eats machinery — those roles grow relatively. The laborer positions at pour, shakeout, and the grinding wheel are the ones being engineered out, deliberately and without much mourning.
Automatability: our editorial assessment of current and near-term AI capability
Actively underway and accelerating through the late 2020s: automated pouring and robotic grinding are standard in high-volume foundries now, and labor shortages plus safety liability push even mid-size shops to automate station by station. 3D-printed molds compress the sand side further. Low-volume jobbing shops and giant one-off castings stay manual longest, but the general laborer role in casting is on a firm countdown this decade.
The laborer positions are, and mostly on purpose — pouring, shakeout, and grinding are dangerous enough that regulators, insurers, and workers themselves welcome the robots. Total foundry employment keeps declining, but the shops that survive are hiring different people: robot techs, furnace operators, metallurgy and quality staff. It's less job destruction than a forced upgrade of the job description.
The judgment and maintenance layers. Melters who manage alloy chemistry, furnace operators handling temperamental equipment, metallurgical and quality technicians, and the maintenance crews keeping robotic cells alive in an environment full of heat, dust, and vibration. Low-volume jobbing work — custom castings in small runs — also resists, because you can't economically tool a robot for fifty one-off parts a week.
As a technician, yes; as a laborer, no. Foundries face genuine skilled-labor shortages and pay real money for people who combine casting knowledge with electrical, robotics, or metallurgical skills — and the industry isn't going anywhere while engines, pumps, and infrastructure need cast parts. Enter through an apprenticeship or technical program aimed at the melt deck or maintenance, not the grinding room.
It's replacing pieces of the process more than the process. Printed sand molds are already standard for prototypes and complex cores, killing traditional pattern-making, and direct metal printing takes some small, high-value parts. But casting remains unbeatable economics for large parts and high volumes — engine blocks aren't getting printed. For workers, the takeaway is that mold-making skill is migrating from wood patterns to software.