3D Weaving Apparel Manufacturing at Arvind: OneWeave, Explained for Buyers

On 9 November 2026, American fashion-technology company unspun and Indian textile major Arvind Limited announced a partnership that moves 3D weaving out of the start-up’s own micro-factory and into a tier-one manufacturer’s production environment. Arvind will install two OneWeave machines at its Ahmedabad plant, with the first units expected to start running in December 2026. A phased plan could add more than 30 machines across India and North America in 2027–28 — if global apparel brands place the orders to justify them.

For buyers, the announcement raises a practical question that the press-release coverage has largely left unanswered: what can you actually order, from whom, and on what terms? This article sets out what the partnership covers, how the technology works, what the European regulation behind it really says, and a framework for judging the first trial orders.

What the partnership actually covers

The structure matters. Arvind is the first machine customer for OneWeave, and the first manufacturer outside unspun’s own operation to run the technology at industrial scale. Until now, unspun has produced through a micro-factory in Emeryville, California, which had seven machines as of July 2026 (The Interline, interview with unspun’s co-founder, 16 July 2026). Its customers include Walmart and Decathlon, according to the press release (9 November 2026). The Arvind deal is therefore a test of whether the technology can operate inside a large, multi-product manufacturer — not just inside a start-up’s controlled environment.

Arvind is a substantial partner for that test. The company reported consolidated revenue of ₹9,303 crore for the financial year ended 31 March 2026, up 12% year on year, with denim fabric output of 59.5 million metres, woven fabric of 135.8 million metres and garments of 41.6 million pieces. Its customer list, as set out in the announcement, includes GAP, PVH, Zara, H&M and Levi’s. That scale is the point: if OneWeave works at Arvind, it has a credible route to volume. If it stalls there, the technology’s commercial case weakens considerably.

Two distinctions should be kept in mind from the outset. First, the 30-plus machines are a conditional pathway, not a confirmed order. The press release describes a phased partnership “with a pathway” to additional machines, dependent on demand from global brands. Trade coverage that presents 30 machines as a done deal overstates the announcement. Second, December 2026 is an expected start date, not a completed milestone. The first real test will be whether the two machines are running and producing saleable garments on that schedule.

How OneWeave 3D weaving works

OneWeave weaves yarn directly into a three-dimensional garment shape, bypassing two-dimensional fabric production and most cutting and sewing. The process uses circular weaving: roughly 3,000 yarns are woven simultaneously into a tube of variable diameter, which becomes a trouser leg without a side seam. That figure is unspun’s own, given by its co-founder Arne Arens in an interview with The Interline in July 2026; it describes the company’s own machine, not an independently measured specification. Arens has also said that a pair of legs takes about eight minutes to weave (The Interline, July 2026). That is a time per unit, not an annual output: converting it into trousers per year would require assumptions about uptime, changeovers and defect rates that have not been published, so this article does not annualise it.

The immediate consequence is a shorter production chain. In conventional cut-and-sew manufacturing, fabric is woven or knitted, cut into pattern pieces, and then sewn together. OneWeave removes the fabric stage and most of the cutting and sewing stages for the parts of the garment it can form. Finishing, inspection and any remaining assembly still take place downstream — the machine does not deliver a finished, saleable garment on its own.

The product range is narrow for now. As of early 2026, the technology was suited to woven trousers — casual trousers, workwear and chinos — with denim described as “soon” to follow. The tube structure of the weave maps naturally onto a trouser leg; it does not yet map onto shirts, jackets or dresses, and no public evidence shows it doing so. Buyers should treat the current offer as a trouser technology, and judge any claim of a wider range against demonstrated output rather than a roadmap.

That narrowness is not a flaw in itself. It is the normal shape of a manufacturing technology in its first industrial deployment: one product family, one factory, two machines, and a set of questions that can only be answered by running orders through them. The relevant comparison for buyers is not with the whole of cut-and-sew manufacturing, but with the specific trouser programmes where a woven, seamless leg is an advantage — and where it is not.

Why now: the EU ban on destroying unsold clothing, accurately stated

Much of the coverage links the partnership to the European Union’s ban on destroying unsold clothing. The link is real, but the regulation is routinely overstated, so it is worth stating what it actually says.

The ban sits in Article 25 of the Ecodesign for Sustainable Products Regulation, Regulation (EU) 2024/1781. A separate provision, Article 24, imposes its own annual disclosure obligation on economic operators that discard unsold consumer products. The two provisions differ in scope, timing and effect, and should not be conflated:

Two corrections follow. First, the ban is not universal: small businesses are exempt, medium-sized businesses have until 2030, and the derogations are broad enough to cover genuinely unsaleable stock. Second, the ban does not create demand for 3D weaving by itself. What it does is remove the cheapest disposal route for unsold stock. A large enterprise that can no longer destroy unsold garments must instead route them through reuse, preparation for reuse or remanufacturing, recycling, donation, or destruction under a derogation — and each of those routes carries handling, logistics or processing costs, on top of the reporting cost of the Article 24 disclosure. That is the mechanism by which the regulation can raise the cost of overproduction: not by a fine on producing too much, but by making the disposal of the surplus more expensive and more visible. The effect is conditional, not automatic — it depends on how much surplus a company actually holds, which routes are available to it, and whether the derogations apply. What 3D weaving offers in response is production flexibility — smaller runs, closer to demand — not a solution to the forecasting problem that creates overproduction in the first place. A buyer who commissions a OneWeave run and still over-orders has simply moved the waste upstream.

That said, the timing is not coincidental. The regulation changes the economics of the most wasteful part of the conventional model: producing large runs of a style, then destroying what does not sell. Published research on conventional cutting (Rissanen & McQuillan 2016, drawing on earlier studies from 1997 and 2004) estimates fabric loss at the cutting stage at around 15% on average — about 10% for trousers, higher for tops, within a range of roughly 10–20% depending on the product. Those are literature estimates of marker efficiency in conventional cutting, not measurements from unspun’s machines, and they describe the conventional process, not OneWeave’s. But they explain why a technology that skips the cutting table attracts attention from manufacturers whose margins are squeezed by both material cost and disposal rules. Readers who want the detail on where cutting waste arises can refer to our guide on fabric waste in garment manufacturing, and to our fabric consumption calculation guide for how marker efficiency is normally estimated.

3D weaving vs cut-and-sew: what can (and cannot) be compared yet

The honest summary is that the technology’s advantages are plausible but largely unmeasured in an industrial setting. The cutting-waste argument is the strongest, because it follows directly from the process: no cutting table means no marker waste for the woven body. But cutting waste is one waste stream, not the whole picture. Yarn is still lost in preparation and at the machine (yarn breaks, start-up and changeover waste), the woven body still needs finishing and trimming, and defective pieces are still waste. The correct claim is that OneWeave removes the cutting stage for the woven body — not that a OneWeave garment is made with zero waste. Until yarn waste, weaving waste, finishing loss and the defect rate are published for a real production run, no total material-efficiency figure can be stated. The speed and lead-time claims are company statements, and the per-machine output figures in circulation contradict each other. The cost question is the most important for buyers and the least answered: the only public statement on cost is that per-unit costs will initially be higher, and that utilisation — how fully the machines are kept running — is what will bring them down.

That last point is worth underlining, because it determines who should trial the technology first. A machine that is expensive per unit at low utilisation becomes competitive only when it is kept busy. A brand that books a small, one-off trial and walks away learns little and pays the highest unit cost. A brand that commits a programme — a repeating trouser style, in volume, over several seasons — is the customer the economics are built for. The first-come, first-served allocation of initial slots should be read in that light: early access is real, but it is access to a production line that needs committed volume to make sense.

Utilisation, however, is not simply a matter of promising volume. It depends on order batch sizes, how often styles change, changeover time between styles, and how much capacity Arvind reserves for reorders rather than first orders. A buyer should therefore distinguish three different numbers: the forecast volume in a planning document, the committed volume in a contract, and the volume actually delivered. Only the second and third affect utilisation, and only the third is certain. The safer structure is a phased commitment — a small trial, a measured evaluation, then a rolling forecast with conditionally reserved capacity — rather than a long-term volume promise made before the trial has been validated. A buyer who promises volume first and validates second has taken on the procurement risk without the evidence.

A buyer’s evaluation framework for the first trial orders

This is the part the press-release coverage does not provide: how a buyer should actually judge whether to place a trial order, and what to ask before doing so. The framework below is written for a sourcing or production manager considering a first OneWeave programme with Arvind.

Start with the right styles. The first trials should be straight or tapered woven trousers with a simple construction — chinos, casual trousers, workwear — in stable, solid colours. Avoid styles that depend on complex panel construction, heavy topstitching detail, special washes or multi-fabric combinations for the first run. The tube structure is the technology’s strength; asking it to reproduce a heavily constructed garment in the first order tests the wrong thing and produces a misleading failure.

Ask five questions before committing. First, which yarn counts and fibre compositions can the machines currently run, and what is excluded? Second, how is size tolerance controlled across the woven tube, and what is the tolerance on leg length and waist? Third, what colourfastness and wash performance has been demonstrated on the exact fabric and shade you intend to order — not on a different programme? Fourth, what is the minimum order quantity, the reorder lead time, and the capacity reserved for reorders rather than first orders? Fifth, what is the defect rate on the current production, how are defective pieces handled, and what is the replenishment mechanism when a run falls short? None of these five has a published answer yet. That is not a reason to avoid the technology; it is a reason to ask the questions in writing before placing an order, and to treat any answer given verbally as unconfirmed until it appears in the order terms.

Move the quality checkpoint. In conventional cut-and-sew, quality control concentrates on cutting accuracy and sewing — stitch density, seam strength, needle damage and measurement tolerance. OneWeave shifts the defect risk forward into the weaving stage: uneven yarn tension, variation in tube diameter, uneven weaving density, and the dimensional stability of a seamless structure. A trial order’s inspection plan should therefore add weaving-defect categories that a standard garment inspection does not cover, and should define them in the order specification rather than discovering them at final inspection. Our guide to quality control for small orders sets out the discipline of sealed samples and written tolerances that applies here with particular force, and our fabric defect classification guide covers the defect categories that a weaving-stage inspection should draw on. Note also that the small-order logic applies: a first trial is a small order, with the same thin margin for error.

A trial-order acceptance framework. The five questions above establish what to ask; the framework below establishes what to check, and when. It is a suggested framework for the buyer’s own order specification — none of these values is a published OneWeave standard, and any figure used in a contract should be agreed in writing with Arvind, not taken from this article.

What to agree before production. Yarn specification and fibre composition for the exact style, including the yarn counts the machines can currently run; dimensional tolerances for waist, leg length and tube diameter, with the measurement method stated; colourfastness requirements, with the grades to be met — the applicable methods to be agreed with Arvind, for example ISO 105-C06 for colourfastness to washing, ISO 105-X12 for colourfastness to rubbing and ISO 105-B02 for colourfastness to light; wash performance, including dimensional change after washing by the ISO 3759 procedure for preparing, marking and measuring specimens and ISO 5077 for determining dimensional change in washing and drying, with the permitted change stated; and the defect list, classified as critical (a defect that makes the garment unsaleable or unsafe), major (a defect that affects wearability or appearance) or minor (a cosmetic defect that does not affect use).

How to sample and judge. Agree a sampling plan and an AQL (acceptable quality level) in the contract — for example, AQL 1.5 for major defects and AQL 4.0 for minor defects are common starting points in general apparel inspection, but they are suggestions, not OneWeave standards, and the right levels depend on the product and the price point. State the inspection lot size, the sample size and the accept/reject numbers, and whether inspection is by the buyer, a third party, or Arvind’s own team with buyer sign-off.

Four inspection gates. First, a sealed-sample approval: a physical sample of the exact style, yarn and shade, signed by both sides, against which production is judged. Second, first-article inspection of the first pieces off the machine, checking tube diameter, length and weave density before the run proceeds. Third, in-line inspection during production, checking for weaving defects — uneven yarn tension, variation in tube diameter, uneven weaving density — that a standard garment inspection does not cover. Fourth, final inspection before shipment, including the post-wash dimensional check. Each gate needs a named decision-maker and a rule for what happens to rejected pieces: rework, replacement, or a credit against the order.

Who is responsible for what. The brand is responsible for the specification — the style, the yarn, the tolerances and the acceptance criteria. Arvind is responsible for producing to that specification and for the quality of the finished garment. Unspun, as the technology provider, is responsible for the machine’s performance and for the installation, training and software support it has undertaken to provide. Where a defect is caused by the weaving process itself rather than by the garment specification, the order terms should say which party bears the cost of rework or replacement — and the buyer should not assume that question has an answer until it is written into the order.

Negotiate for utilisation, not for a headline price. Since per-unit cost is initially higher and utilisation is the stated driver of cost reduction, the negotiation that matters is not a one-off unit price. It is the structure of the commitment: a phased programme with agreed volumes, a defined reorder window, and a mechanism for sharing the cost of the learning curve. But the commitment should be staged: a trial first, an evaluation of the trial’s quality and cost, and only then a volume commitment — with the volume expressed as a rolling forecast and a conditionally reserved capacity, not as a fixed long-term promise made before the technology has been validated on the buyer’s own styles. A buyer who can commit a repeating style across seasons has something Arvind needs — utilisation — and should negotiate accordingly. A buyer who wants a single small trial should expect to pay for it, and should treat the trial as a learning cost rather than a production saving.

Verify, do not assume, the sustainability claim. A OneWeave garment eliminates cutting waste for the woven body, but that is one waste stream among many: yarn preparation waste, weaving waste, finishing and trimming loss, and defective pieces all remain. Ask for the figures that would support a sustainability claim on the specific order: yarn waste at the machine, energy use per garment, water use in finishing, and the defect rate. If those figures are not available, the honest description of the order is “produced without a cutting stage”, not “sustainable”. The distinction matters because the EU regulation discussed above rewards verifiable disclosure, not adjectives.

What to watch next

Three dates and one condition will tell buyers whether this partnership is progressing or stalling.

First, December 2026: whether the two machines at Ahmedabad start production on schedule, and whether Arvind confirms the first output. A quiet delay here would be the most informative signal of the year. Second, early 2027: an investor briefing from SOSV’s HAX programme expects the first orders to ship in early 2027, and whether named brands appear as customers will show whether the first-come, first-served slots have been taken by serious programmes. Third, the 2027–28 expansion: the 30-plus machine pathway depends on brand demand, so the announcement of further machines — or its absence — is itself the evidence. Watch for a firm order for additional machines, not another restatement of the pathway.

The condition to watch is utilisation. If Arvind’s two machines run at high utilisation on repeating trouser programmes, the cost curve should improve and the case for expansion strengthens. If they run sporadically on one-off trials, the technology will remain a demonstration. That is the honest test, and it is the one buyers can apply to their own decision as well: the question is not whether 3D weaving is the future of apparel manufacturing, but whether a specific trouser programme, at a specific volume, is better made this way today.

Frequently asked questions

Can I place an order for OneWeave garments now? Brands contract with Arvind, not with unspun. Initial capacity is limited and allocated on a first-come, first-served basis. No confirmed minimum order quantity or unit price has been published, so the first step is a commercial discussion with Arvind, not a standard order form.

Can OneWeave make denim jeans? Not yet, as far as public information shows. As of early 2026, the technology was suited to woven trousers such as chinos, casual trousers and workwear, with denim described as coming soon. Treat denim as a future capability until Arvind or unspun demonstrates it.

How is 3D weaving different from 3D knitting? Both form a garment shape directly, but they use different processes: 3D weaving interlaces two sets of yarns at right angles in a circular loom, while whole-garment knitting interloops yarns into knitted structures. The result is a woven fabric with woven properties — stability, structure and a woven hand — rather than a knitted one. Readers who want the woven-versus-knitted distinction in full can refer to our knit versus woven fabric guide and our warp knitting explainer.

Is OneWeave suitable for a small brand? The honest answer is not yet, or not on standard terms. Initial capacity is limited, slots are allocated on a first-come, first-served basis, and the economics depend on utilisation. A small brand with a repeating woven-trouser style and a willingness to commit volume may find a route in; a small brand wanting a single small run should expect high unit costs and should weigh whether the trial is worth the learning cost.

Is there a published minimum order quantity? No. No confirmed minimum order quantity or unit price has been published. The technology is suited to smaller runs than conventional cut-and-sew, but “smaller” has not been quantified, and the first step is a commercial discussion with Arvind, not a standard order form.

How are acceptance standards set for a trial order? There is no published OneWeave acceptance standard. The buyer and Arvind must agree the specification in the order terms: yarn and fibre composition, dimensional tolerances, colourfastness grades, wash performance, defect classification and the sampling plan with AQL levels. The framework in this article is a suggested starting point, not a standard.

Who is responsible if a trial order fails inspection? That depends on the order terms, which is why they should be written before production starts. In principle, the brand owns the specification, Arvind owns production to that specification, and unspun owns the machine’s performance and its installation, training and software support. Which party bears the cost of rework or replacement for a weaving-process defect should be stated in the order, not assumed.

How much do OneWeave garments cost, and can I get a quotation? No published price exists. Arvind’s vice-chairman has said per-unit costs will initially be higher, according to a report citing a Forbes interview, but no price list, minimum order quantity or unit cost has been published, so no expected price can be quoted here. A quotation can only be obtained by a commercial discussion with Arvind.

Sources

• PR Newswire, “Landmark Partnership Between Unspun And Arvind Limited”, 9 November 2026. prnewswire.com
• Textile World Asia, “Landmark Partnership Between Unspun And Arvind Limited”, 9 November 2026. textileworldasia.com
• Innovation in Textiles, “unspun and Arvind partner to scale 3D weaving”, 10 November 2026. innovationintextiles.com
• The Interline, “Can 3D weaving make domestic production viable?” (interview with Arne Arens), 16 July 2026. theinterline.com
• Inc.com, “The Former Head of North Face Just Bet His Reputation on This Tiny Startup”, 3 March 2026. inc.com
• SOSV HAX, “Can 3D weaving change the economics of fashion?”, 6 November 2026. hax.co
• MakhanaMornings, “Arvind–unspun 3D weaving factory economics”, 10 November 2026. makhanamornings.com
• Regulation (EU) 2024/1781, Articles 24 and 25 (destruction of unsold consumer products), EUR-Lex, accessed 12 November 2026. eur-lex.europa.eu
• Arvind Limited, press release on Q4 FY26 results, 15 May 2026. arvind.com
• SpringerProfessional, “Analysis of zero-waste patternmaking approaches for application” (Rissanen & McQuillan, 2016). springerprofessional.de

Illustrative image: this article uses an AI-generated illustration as its featured image. It is not a photograph of the OneWeave machine or of Arvind’s factory.