■ MODERATE RISK ■ Agriculture & Environment
Not fully — the shoe is getting automated, the horse is not. AI can design a plate from a hoof scan, but someone still has to trim a live animal that outweighs them by a thousand pounds and has opinions about it.
“3D-printed titanium horseshoes are custom-fit by AI. Your forge is now a museum.”
Our AI replacement risk score — how we score jobs
A racing farrier's week runs on a cycle most people never see: trimming hooves every four to six weeks, hot- or cold-fitting shoes, balancing feet so a half-ton animal can hit 40 mph without tearing a tendon, and reading subtle lameness signs the trainer missed. The craft half is shoeing; the diagnostic half is knowing that a slightly worn lateral branch means the horse is landing wrong, and that the fix is in the trim, not the shoe. Racing work adds stakes — aluminum racing plates, glue-on shoes for cracked hooves, and trainers who want everything done between workouts.
The design side is where automation has genuinely arrived. Hoof scanning apps generate 3D models, and custom plates — titanium, polymer, 3D-printed — can be fabricated off-site to tolerances no anvil matches. Gait-analysis software flags asymmetries earlier than the human eye. That erodes the forge work: if the shoe arrives finished, the farrier's fabrication hours shrink and the job tilts toward fitting and application. Stables already using printed plates treat the farrier more like an installer with veterinary instincts than a blacksmith.
What resists is everything involving the actual horse. Trimming is irreversible carpentry on a living structure that varies day to day with ground conditions, diet, and training load. No robot currently holds a nervous thoroughbred's leg, judges sole depth by feel, and adjusts mid-trim when the animal flinches. The farrier also functions as a translator between vet, trainer, and owner — a role that's mostly trust. Our risk score of 47 reflects a trade whose product is being automated while its practice largely isn't: fewer forging hours, same number of hooves.
Automatability: our editorial assessment of current and near-term AI capability
Printed and precision-machined plates are already in elite racing stables, and hoof-scanning tools will be standard kit by the early 2030s. Expect the fabrication half of the trade to keep hollowing out through the decade while the hands-on trimming and fitting work stays stubbornly human. By ~2040 the role looks like a mobile hoof-care clinician who rarely lights a forge — transformed, not eliminated.
Yes, with a caveat. Horses will need hoof care as long as racing exists, and there's a well-documented shortage of skilled farriers in most regions. What's shrinking is the blacksmithing component — custom fabrication is moving to scanners and printers. The farriers who thrive will be the ones selling judgment and hands-on skill rather than metalwork.
For the shoe itself, often yes — a plate machined from a 3D hoof scan can hit tolerances hand-forging can't, and materials like titanium and composites outperform steel for racing. But a perfect shoe on a badly trimmed hoof is still a lame horse. The design is automatable; deciding what that individual foot needs this month isn't.
Digital hoof mapping, gait-analysis software, and glue-on and therapeutic shoeing techniques. The trade is drifting toward hoof-care clinician and away from blacksmith, so anatomy and lameness knowledge pays better than anvil time. Farriers who can read a vet's radiographs and translate them into a trim plan will be booked solid.
Not any time soon. Shoeing means restraining and working under a large, unpredictable prey animal — a robotics problem far harder than welding car panels. Automated trimming chutes exist for cattle hoof care, but racehorses are a different insurance conversation entirely. The physical act of shoeing is among the most automation-resistant parts of the job.