To the naked eye, two shirts in the same product line are identical. Since they come from the same brand and from the same collection, it’s easy to assume that each shirt must be inherently the same, minus the difference in sizing.
This, however, is a mistake, according to Christopher Wai, co-founder of Sixone, a tech firm based in Canada that has been using AI to comb through textile waste for recyclable material. Items might look the same, but they do not necessarily have the same chemical composition. That difference matters.
Sixone tested for this. The company looked into close to 1,000 pieces garments from a brand partner, all in the same style. On the surface, everything looked the same and the brand even said they all came from the same manufacturer. However, Sixone found that the garments had different chemical compositions.
“What we showed was that for that 1,000 products, there were three different chemical paths that those products go down to,” he told Sourcing Journal. “That’s really important because we then have to throw away our perceived expectation that if things look the same, they will react the same. We cannot rely on that.”
This is an important premise to establish when discussing textile waste. When textile waste is recycled, it gets broken down to its chemical building blocks. If all the garments have the same chemical properties, they can all be recycled the same way. Otherwise, the chemicals can react against each other and hamper recyclability, or worse, release harmful byproducts into the environment.
“If you put something that has PFAS into a reactor and something that has urethane into a reactor or silicone into a reactor, those three will break down in different ways,” Cai explained. “Polyurethane, for example, breaks down into something where it is it is very toxic in the water. Even a part per million into the water is a big problem. You don’t want that to break down.”
The challenge is to figure out how to recycle the material in a way that leaves the PFAS unaffected or keeps the Polyurethane from breaking, he said. To find the best paths of action, Sixone turned to AI and algorithmic learning.
Sixone has been working on a solution—a pilot project—for three years now. It developed a proprietary AI called Prism that predicts how much high-value polymer is in textile waste and how to get the best possible outcome. So far, it said it has recycled 70,000 kilograms of complex textile waste, which were then converted to polyester it calls Origin, all of which has been sold to brand partners.
It is important to note that Prism predicts chemical reactions, but it doesn’t identify the actual chemical composition. “Predict” is the operative term. It just needs to know which textiles could react the same way, regardless of the chemicals inside them, so that the AI can determine the best move forward.
This speaks volumes to how complicated and vast the problem of textile waste has become. Out of 120 million tons of clothing that was discarded globally in 2024, 80 percent went to incinerators or landfills. Less than 1 percent was recycled back into new textiles, according to the Boston Consulting Group. The problem will only get bigger, with BCG estimating 150 million tons of discarded clothing in 2030.
It’s why Cai considers highly precise predictions to be the more feasible way of approaching the problem, as opposed to testing each piece of discarded textile.
“There are very expensive and long chemical tests that you can do to get to exactly what is in a piece of clothing,” Cai explained. “But the thing is, we have a lot of different types of clothing that’s out there. We have a lot of different additives, and even when you have two pieces of the same thing that look exactly the same… you can have different chemical compositions in there.”
In other words, identifying the chemicals and predicting the reactions are two ways of answering the same question: How much polyester can be recycled from this textile waste? But while they both answer the same question, they get to the answer differently, with one being more costly and destructive than the other.
Sixone’s work is far from done. Cai said he wants to have all the necessary algorithms to handle every possible scenario, a goal which means a longer runway for the startup. It’s eyeing commercial operations launching around 2030. He said Sixone is currently developing a scalable demonstration facility in Canada.
As with any innovation, there is a higher premium at the start. Cai believes in the long run Sixone will be able to become cost competitive enough to rival branded recycled polyesters made from plastic bottles. It’s a long journey that requires both patience and resources, and Cai counts his company fortunate to find partner brands that have been engaged for years.
“It’s about the brands that are willing to go down that journey and understand that there is a cost during these development cycles. Are they bearing all of it? Not necessarily, but are they willing to be on that journey together with us? That becomes quite important,” Cai said.