Our targets, not measurements. Set by automation progress; move any dial to try your own.
Useful output while running, including waits for work, compared with a trained worker sewing at a steady pace.
Hardware with financing, repairs, and power. Our estimate for two YAM arms.
A technician is on duty whenever robots run, at least one shift a day. Assumes factories big enough to keep each technician busy. Rises as robots learn to swap their own broken modules.
Office, building, and support costs, compared with a traditional factory. The building and power always remain.
Measured in days of the robot's own output. A sewing line loses about one day. Matters most for small orders.
Order
Order size
Delivery
Fabric for US and Mexican factories
Trade policy
Added US tariff on Asia
Garment
Supply
People and overhead
Cost per garment sold
Everything it takes to put one garment in a US customer's hands, including stock that never sells. The dashed line is the benchmark: today's Asian factories, or US factories when Asia can't take the order.
Everything it takes to put one garment in a US customer's hands, including stock that never sells. Hover over or tap a bar to see what it's made of.
Breakdown
At a glance
How robot factories compare with today's factories at each stage. Ranges cover T-shirts, scrubs, and fast fashion, at default settings.
Takeaways
Each takeaway is computed from the model and tested on random mixes of every input's plausible range.
What matters
Tested and left out
Political risk beyond tariffs. Port closures or sanctions have no reliable price. The tariff setting and sea freight cover their measurable effects.
Factory margins and sourcing costs. Left out everywhere. Including them would slightly favor robots: Asian imports also pay duty on the factory's margin, plus agent, inspection, and port fees, about 2% to 8% more.
Missed sales from slow supply. Ordering months ahead means running out of what sells and marking down more of what doesn't. That would favor US and Mexican factories, mostly for fashion, but it depends on how a brand plans, so we leave it out.
Slow seasons and ramp-up. Left out everywhere. They would hurt robots, which cost the same with or without work, while staffing can flex.
Made-to-measure. Covered by rush orders.
Check it yourself: which assumptions move the result most
Each bar moves one assumption across its plausible range and shows how much that changes the gap between robots in the US and today's Asian factories (or today's US factories, when Asia can't take the order: military uniforms or rush orders). Longer bars matter more.
How the model works
What one garment costs to make and deliver, priced the same way for six factories. Left out: building factories, ramp-up, profit margins, and trucking from US factories to buyers, which adds a few cents. Minutes of work come from time studies in real factories, so they include the machines those factories already use.
Traditional US
Workers at US wages, close to customers.
Traditional Asia
Workers at Asian wages, months from customers.
Traditional Mexico
Workers at Mexican wages, days from customers by truck.
Automated US
Our robots in the US, close to customers.
Automated Asia
Our robots in Asia, months from customers.
Automated Mexico
Our robots in Mexico, days from customers by truck.
Price a minute of work
A worker's wage, or a robot's yearly cost plus its share of a technician, divided by the minutes actually worked.
wage ÷ efficiency robot cost ÷ minutes worked
Make the garment
Minutes of work times that price, plus materials and overhead. US and Mexican factories buy North American or imported fabric, whichever is cheaper. Failed robot garments add waste.
minutes × price + materials + overhead
Ship it
Garments from Asia pay a US tariff and freight. Mexican garments are trucked, duty-free when made from North American fabric.
+ tariff + freight
Hold stock
Ordering from Asia means holding stock for months longer.
+ holding cost
Count only what sells
Ordering further ahead leaves more unsold. Markdowns recover part of that cost; the rest is spread over what sells.
Vietnam Briefing: overtime rules. 48 regular hours a week; garment overtime is capped at 300 hours a year. Total employer cost of about 14.6M VND a month, over these hours, is about $2.50 an hour.
What does a Mexican sewing operator cost? $4.75 an hour, all in. 4 sources
How much of their paid time do workers spend producing? 50% in Asia, 55% in Mexico, and 65% in the US. Mexico and the US have little data, so they get the benefit of the doubt. 4 sources
How many hours do workers and technicians work? Workers: 2,080 a year in the US, 2,496 elsewhere. Technicians: 1,880 in the US, 2,300 elsewhere. 3 sources
How many minutes of work does each garment take? T-shirt 8, scrubs 24, fast-fashion top 15, and uniform 80. Only the T-shirt has a published study; the rest are estimated from similar garments. 4 sources
TIS cargo handbook: garments. Garments ship at about 105 to 435 kg a cubic meter, about 10 tonnes a container. With port fees and trucking, about $0.40 a kg.
What does air freight cost? $5 a kg in a normal year; about $7 in 2026. 3 sources
What does factory overhead cost per minute of work? $0.28 in the US, $0.09 in Mexico, and $0.045 in Vietnam, built from staffing data and plant cost cases. 3 sources
BLS: staffing in cut-and-sew apparel factories, May 2023. Supervisors, inspectors, mechanics, and handlers earn 31% as much as sewing operators in total: about $0.19 per minute of work in the US. Building, office, and utilities add about $0.09 to $0.13.
Adecco: Vietnam Salary Guide 2024. Automation engineers earn 15M to 35M VND a month in Ho Chi Minh City; technicians earn about half.
Robots
Where do the robot numbers come from? They're our targets, not measurements. Robot cost is our estimate for two YAM arms, which sell for about $3,000 each at retail and less at scale, plus financing, repairs, and power. 2 sources