The fashion business thrives on the conceit that it is modern, on-trend, and all about what’s happening right now. But behind the scenes—in the textile mills and garment factories that supply major brands—is an industry that runs much as it did in the 1970s. Its supply chains are splintered, chaotic, and based in warehouses that are so old-fashioned they may not even be on the internet, according to Max Easton, CEO of Smartex.AI.
Most of the work is done by hand. Checking the quality of the fabric, for instance, will be performed by a worker holding a handful of color swatches, watching a roll of cloth go by all day.
“They’re looking for simple physical defects like holes, needles, the wrong yarn, something that would be noticeable in a garment, right?” Easton said to Fortune at Web Summit in Lisbon this year. For most brands, there is no such thing as a “white” T-shirt. It needs to be exactly the right shade of white—which might be a brilliant white or a creamier off-white.
But because the work is inspected by humans, fashion brands can never be quite sure that the quality of the fabric they’re buying is going to be consistent, or even match what they ordered.
“If you go to any kind of [fashion] supply chain, whether it’s fabric production or garment production, most people are just kind of amazed by how chaotic it is,” Easton said.
The chaos ends now, Smartex hopes.
The company has been working with Tony Fadell, the legendary co-designer of the iPod, iPhone, and Nest doorbell, to launch a physical AI product that can perform fabric quality checks continuously, at three times the speed of a human. Fadell—who has invested roughly $10 million in the company—is coming aboard as Smartex’s chairman. H&M, also an investor in Smartex, is one of its launch customers. Smartex has about 65 employees, revenues in the low millions annually, and has installed 800 of its older monitoring systems at clients in 10 different countries. The company’s valuation four years ago was $55 million; Easton declined to give an updated number.
The company is rolling out two new products: Smartex Inspect is a physical AI camera system that attaches to existing textile machinery to monitor the quality of the material as it goes by. And Smartex IQ generates intelligence from the supply chain data that comes from Smartex’s suite of products.
Fabric inspection is “very, very technical, and it’s highly prone to human error,” Fadell told Fortune. “Depending on whether it’s in the morning or in the afternoon or on Friday afternoon … it’s all subjective. There’s no Pantone color for quality inspection.”
“If you walked into a mill or a garment facility 50 years ago, and you walk into it today, it’s almost exactly the same, same processes, same equipment, same everything. It’s not been upgraded in any way. They might have email. In some cases they don’t even have Wi-Fi,” he said.
Companies are at the mercy of independent operators in multiple foreign countries. “They have zero visibility, even when [the product] arrives. They don’t know when it’s going to arrive,” Easton said.
For any player in the apparel chain, 50% of their cost-base is fabric. But because the product is so variable in quality, a company like H&M will typically over-order about 5% more fabric than they need—in the sure knowledge that some of it won’t be good enough to use.
“So globally, [at H&M] that’s like 2.5 billion meters of fabric every single year, just in case,” Easton says. At $4 a meter, that adds up to $10 billion in fabric costs. Easton reckons Smartex might save H&M $100 million annually simply by reducing its waste.
There are four things that H&M cares about, Fadell said: price, speed, quality, and sustainability. “A lot of times they commit to making this garment. They pre-sell it online, but they can’t deliver it because the fabric shows up at the garment house, and they can’t make the garment because the fabric’s wrong in some way,” he said. “Reducing waste means better margins.”
The challenge isn’t as simple as pointing a camera at a textile processor and sending it to a computer. The system has to take account of changes in lighting, the speed at which the fabric is moving, small variations in color, and calibrate all that against a client’s own standards. Smartex has a library of 50 billion images of fabric to do that.
It is, admittedly, somewhat odd to hear Fadell—famed for creating some of the most iconic consumer products ever made—enthusing over loom quality and stitching. But, he said, “This is something I’ve been investing in for a decade now, because I believe in very, very applied AI in the physical world, but you’ve got to make it really easy for the customers to use.”
This story was originally featured on Fortune.com
.png)
1 hour ago
3




English (US) ·