Why molds and cores stay on the shop floor
Foundry mold and coremakers build the sand molds and cores that give a casting its shape. The work is packing and ramming sand around a pattern, setting cores by hand, cutting vents and gates, and patching a mold that cracks before metal goes in. Almost none of it is reading, writing or data entry. It happens next to hot metal, in a space where a soft corner or a shifted core shows up as scrap an hour later.
The material changes from shift to shift. Sand moisture, binder mix and pattern wear all drift. A coremaker adjusts by feel and by eye, then fixes the mold rather than scrapping it. Language models have no route into that loop. They can summarize a defect report; they cannot press a core into place or judge when a mold face is hard enough.
There is real pressure on this job, but it comes from machines and from consolidation, not from chatbots. The US Bureau of Labor Statistics counts about 12,790 people in the occupation, with median pay around $48,110, and projects employment to fall roughly 23% between 2025 and 2035 (BLS). Work shifts to automated molding lines and to molding machine operators, while hand molding holds on in job shops and repair work.
What AI does, what it helps with, and what it leaves to people
No part of this job has moved into the AI-does-it group. The share of task time software handles end to end is 0%. That is the figure behind the coverage question, which asks only how much of the work AI can carry today. The method behind it is set out on the coverage scoring page.
The support side is where software shows up in foundries. Casting simulation predicts fill and solidification before a pattern is cut. Scrap logs get sorted, scheduling gets tighter, and plant sensor data gets watched for drift. Our split puts 0% of task time in that assisted group. The output is better information for the person at the bench, not a mold that makes itself.
Everything physical stays with people: setting cores to tolerance in complex shapes, repairing and finishing mold faces, lifting and aligning pattern plates, and the last look before a pour. That share reads 100% of task time. If you want to see how a different job compares on the same split, put two of them side by side in the job comparison tool.
How strong the evidence is
Weak, and we say so. The evidence grade here is D, which means no study has tested an AI system against a working coremaker on this job’s tasks. Because of that, we publish no parity number for this occupation. A grade with no measurement is not the same as a result showing people win; it is an open question.
What would settle it is narrow and testable: a timed trial on the same pattern and the same sand, comparing scrap rate, cycle time and rework between a robotic molding cell and a trained coremaker; or plant-level defect data before and after a line is converted. Until something like that exists, the parity question stays unanswered. How we grade that question is explained on the quality parity page.
When this could change
Most likely after 2046 (8 in 10 of our scenarios). What that window means, and how we build it, is described on the replacement year method page.
Two things could pull it earlier. First, 3D sand printing moving from prototype shops into production, which removes pattern and core assembly steps rather than automating them. Second, continued consolidation into high-volume plants, where automated molding lines already do the shaping and the hand work shrinks around them. The BLS projection for 2025 to 2035 points the same way.
Two things hold it back. The work is physical from start to finish, and our robotics read puts the automation route in the fixed automation tier: dedicated machines bought for one product line, not flexible robots that walk into a job shop. That math only pays at volume. Small foundries run short batches with changing patterns, tight capital and old buildings, which is exactly where a general-purpose machine struggles. The guide to robots and physical work covers why that gap is wide.
How to stay needed in the foundry
Lean into the parts of the job that resist both software and fixed machines. Core setting and assembly on complex, cored castings is the first: tolerances there are judgment calls. Mold repair and surface finishing is the second, because fixing beats scrapping on every cost sheet. The third is the pre-pour check, where catching a shifted core or a blocked vent saves a whole heat.
Two skills widen your options. One is reading casting simulation output and acting on it, so you are the person who connects the model to the mold. The other is molding machine setup and tending, which keeps you employable when a plant converts a line. Pattern and tooling work is a third route worth a look.
What to do: look up the machine-operator and casting roles next to this one and compare how their task splits differ.
Close neighbors on the same career path include pourers and casters, patternmakers in metal and plastic, and model makers in metal and plastic. For the wider picture, there is the metal and plastic workers family, the manufacturing sector page, and our list of jobs that mostly need a person. Every figure on this page comes from open data under our published scoring method.