Textile manufacturing is making a comeback in the US, and small businesses play a role in making it happen. But it’s going to look different this time around, with a greater emphasis on robotics and tech.

At a factory in Burlington, Vermont, a laser traces the outlines of a T-shirt across a sheet of fabric. Nearby, a machine turns a knitted tube into a beanie in less than a minute. A few steps away, a young woman works on the finishing details of a men’s jacket by hand. The machines, and the people moving between them, are part of Carey Strobeck’s answer to a question that has haunted American apparel for decades: What can still be made here at a price customers will pay?

Strobeck’s Fourbital Factory makes beanies for its own 4T2D brand and for other companies, nonprofits and institutions. It also produces small runs of apparel, which are sold at the local store in downtown Burlington. During a visit to its roughly 12,000-square-foot Burlington facility, Strobeck said it could knit about 2,500 beanies a week and employed about 30 people across the factory and retail store. The beanies are the scalable product; the clothes allow her team to design, sew, and test new ideas close to home. The beanies, though, are their bread and butter, at least for now.

“Time is our enemy,” Strobeck says. Every extra step raises a garment’s price. Automation helps her compete on beanies, while a small production run lets the company see whether a dress or T-shirt sells before manufacturing hundreds of them. “I don’t have to think five years out,” she argues, referring to mass market brands who are often styling seasons ahead to be able to order, manufacture, and ship from overseas on time.

One of the company’s technical designers, Jenna Baginski, points to a shirt the team made last Thursday; they are currently wearing it in the office. If it resonates with the team, they’ll make more units. “It’s that fast. We literally wear the test pieces and see how they feel.”

Fourbital is one of several attempts to rebuild parts of a domestic textile industry that once supported far more jobs. The scale of the loss is hard to miss: U.S. textile mills employed 503,300 people in January 1990 and 78,300 in August 2026 , according to federal data. Recent small-factory investments do not reverse the national employment trend. But in Vermont, New Hampshire and North Carolina, manufacturers and researchers are trying to find the products, technology and skills that could make more production viable here.

Start with what a machine can do

In Ashland, New Hampshire, Lawson Glidden came to domestic production through a family history that explains why it is so difficult. His family’s L.W. Packard mill began making wool cloth in 1916. It closed in 2002 as customers and suppliers shifted production overseas.

Minus33 , the merino wool business his father developed, still sources most of its products abroad. Glidden, now its president, wanted to bring at least some manufacturing back to Ashland.

He started with socks. After failing to find a supplier that could consistently make the sock he wanted, he bought knitting machines from Italy and worked through hundreds of iterations himself, changing the fit and testing the results. A sock emerges from the machine largely formed, avoiding much of the skilled sewing that makes a shirt or base layer more expensive to produce in the United States, he explains. Minus33 now makes its Mountain Heritage socks in Ashland and is developing locally made hats and scarves.

Four or five of its roughly 25 employees are dedicated to the sock, hat and scarf program, he reports. Glidden sees room to add manufacturing jobs, but says the economics depend on “finding the right machine and the right market.” He has also purchased another building so the historic mill space can be used more fully for production.

Making locally gives Minus33 a benefit beyond a country-of-origin label. Glidden can program several versions of a sock, test them, and adjust the fit without sending specifications back and forth to a distant contractor (similar to what Fourbital does). He can make a custom run for an organization such as the Mount Washington Observatory in weeks, he says. That speed matters in a business where holding excess inventory can be costly.

It does not mean every part of a domestic product comes from America. Minus33’s machines came from Italy, and Glidden said much of the fine merino wool suitable for next-to-skin garments comes from Australia or New Zealand. Strobeck was equally direct about Fourbital’s supply chain: “We get fabrics from all over the world.” Her beanies are often acrylic, chosen for price, color retention and fit, though the company is developing wool versions. The distinction between where a product is assembled and where its fiber, yarn and fabric originate is central to any honest account of American manufacturing and to keeping the product affordable for customers.

Rebuild the chain, not just the factory

Anne Wiper, chief executive of the NSF Textile Innovation Engine of North Carolina , looks at the problem from the other end of the supply chain. The Engine brings together growers, mills, researchers, brands and workforce programs across the Southeast. After an initial $15 million investment, NSF renewed its support with $16 million in funding that could extend to $45 million over three years . The Engine says its network now includes more than 500 partners and has attracted $70 million in additional funding from public, private and philanthropic sources. Its work spans textile hemp, recycling used textiles into new materials, automated manufacturing, biomedical textiles and education.

Wiper cautions against treating all textile production as if it had vanished. North Carolina still has materials manufacturers, experienced workers and NC State’s Wilson College of Textiles , she says. Some mills, she adds, have invested in sophisticated equipment. The larger gap is in turning materials into finished products at scale. “That is the area that needs, I would say, the most investment and will take the most time.”

One test is hemp. Growing the crop became possible under a federal framework following the 2018 Farm Bill , but a usable domestic textile chain requires more than farmers. Fiber must be separated from the stalk, processed, characterized, spun, and made consistently enough for brands to adopt it. Wiper said the Engine has helped convene farmers, processors and buyers, supported work on cleaner “degumming” chemistry and identified a need for better fiber standards. Furniture company MillerKnoll has participated in the Engine’s hemp supply chain discussions , and is keen to support more sustainable and local sourcing for its furniture.

Wiper describes a planned “textile test bed” linking regional facilities so companies developing new materials can navigate equipment and expertise more easily. For now, she says, materials made in the U.S. might still be sent elsewhere to become finished goods while domestic assembly capacity develops. The Engine is also a partner in FutureTEX , a separate defense textile initiative led by NC State. That program has $36 million in first-year funding and a ceiling of $480 million over ten years.

The people who operate these systems matter as much as the equipment. At Fourbital, Strobeck says workers learn several machines and tasks rather than spending their careers on one seam ("Most people don’t want to devote their career to a left sleeve," she says aptly).

To encourage a new workforce that does have the right skills for the future of textile manufacturing, the Engine helped develop a high school textile curriculum that includes chemistry, technical work, and manufacturing as possible careers. The goal is to show young people a future in towns where earlier generations watched mills decline.

Nor does automation eliminate craft. Wiper expects machines to take over more repetitive work while skilled people handle demanding sewing and materials. Strobeck’s team still prototypes garments, checks the fit on different bodies, and looks for ways to reduce waste. Glidden still programs and evaluates the socks his machines produce. Each is using technology to make a smaller American operation responsive enough to compete.

A machine that makes a sock or beanie competitive in the U.S. may support a handful of technicians rather than recreate a mill employing hundreds. For Glidden, even that foothold is worth expanding. For Strobeck, it supports a factory where workers can move between production and creative work. And for Wiper, those businesses are evidence of what is possible, and reminders of the missing links between a fiber and a finished garment.

America is unlikely to remake every garment it buys. Instead, these three efforts point to a narrower, more practical ambition: identify products that can be made competitively, train people to make them well, and keep adapting with the market.