Textile mills

The pre-screen called this mill plant that has left the continent. Half right. The commodity end is closed, and the filing shows why: Unifi, the listed US maker of textured polyester and nylon yarn, reported fiscal 2026 net sales of $531.3M, down 7.0%, a 5.7% gross margin and a $24.6M net loss [A], having sold a manufacturing plant the year before and agreed after year-end to sell $60.0M of real estate. When the scale player cannot clear its costs at the import price, a new spinning frame will not. But the Canadian group is not that business. It booked $276.1M of revenue against $253.9M of expenses in 2024 [A], and its revenue has exceeded its expenses in every year since 2013 — these are specialty survivors, not commodity spinners. And half of the 44 establishments have fewer than ten employees, with eleven in Alberta and British Columbia, far from any textile cluster. The likeliest reading is custom fibre mills processing wool and alpaca for small flocks — that is inference from the size bands, not something the count states. That is the enterable proposition: second-hand carding and spinning equipment, a rural building, fleece arriving by the bag. The cut is market size. A custom mill is paid by the pound on what local flocks shear once a year, its ceiling is the throughput of one small line, and no published figure suggests the niche is wider than the handful of mills already in it.

The pre-screen filed this under structural exit. The Canadian numbers say otherwise, and the difference matters. Fabric mills are the largest and steadiest of the textile groups: $1,305.9M of revenue in 2024 against $1,093.1M of expenses, revenue between $1.19B and $1.35B every year since 2015, and a surplus over expenses of 9–16% of revenue in each of them [A]. Nonwovens alone are $410.4M. These are technical survivors — filtration, geotextile, protective and industrial fabric sold on specification. What has left is the commodity work, and Culp's filing is the local proof: it closed its Quebec mattress-fabric plant, sold the building for C$8.6M, moved part of the knitting to North Carolina and now buys the jacquard it used to weave in Canada from a supplier in Turkey — and reported a $18.4M operating loss on $213.2M of sales in the year of the closure, narrowing to a $7.2M loss on $203.5M the year after [A]. So the enterable question is the technical mill, and the cut is capital intensity. Thirteen of the 92 Canadian establishments have 100 or more employees; US mills average 53. A nonwoven or broad-woven line is sunk before the first metre is qualified, and qualification with a filtration or protective-fabric customer is measured in seasons. The margin the survivors earn is the return on plant and approvals already paid for. An entrant with ordinary resources cannot buy the line, and could not fill it while waiting to be specified.

The pre-screen read this as a finishing plant with an environmental permit attached. The size bands complicate that: 75 of the 107 Canadian establishments have fewer than ten employees, so small shops plainly exist — dye houses, garment finishers, short-run laminators. The group splits in two, and the halves are moving apart. Commission finishing is derived demand: it processes other mills' cloth and other makers' garments, and as those customers thin out so does the work — $110.0M of revenue in 2024, from $120.2M in 2019 [A]. Coating and laminating is the half that grows — $290.0M in 2024, from $239.9M in 2019 [A] — because a coated fabric is a product with its own specification, sold into protective clothing, transport and industrial uses. So the reachable half is shrinking and the growing half is the expensive one. A coating line is a coater, a curing oven, solvent or plastisol handling and an air permit, all sunk before a customer has qualified a single roll, and the whole group ran expenses above revenue in 2022 and a surplus of only 3–8% in the years around it [A]. The cut is capital intensity: the entry an ordinary entrant can afford is the dye house whose customers are disappearing, and the entry worth having costs a plant.