How plastic-eating enzymes turn old polyester into new clothes
On November 9, lululemon launched a limited capsule in Australia and New Zealand featuring two core run styles made with enzymatically recycled polyester (image: lululemon)
Inside a 20-cubic-metre reactor at a demonstration plant in Clermont-Ferrand, central France, discarded clothes are being broken down into their raw chemical parts. Shredded scraps of bright polyester fleece swirl in warm water alongside engineered enzymes that act like microscopic scissors, snipping the plastic’s long molecular chains back into their original building blocks, known as monomers.
In trials using textile waste, the plant’s operator, Carbios, says the process broke down more than 95% of the polyethylene terephthalate – or PET – within 24 hours. Filtration and purification then recover those raw ingredients, allowing manufacturers to spin them into brand-new polyester.
Although the exact same polymer gives a clear drinks bottle its rigid shape and a fleece its softness, the two present starkly different challenges once thrown away. Clean bottles can be shredded, melted and remoulded into new bottles, though repeated heating gradually shortens the plastic’s molecular chains. A fleece, by contrast, is typically laced with dyes, chemical finishes and blended fibres that foil conventional mechanical recycling. Taking polyester apart at a molecular level allows manufacturers to start afresh with purified ingredients, free of the colour and form of the original garment.
From bioreactor to running track
Enzyme-recycled fabric has already moved from pilot plants to shop floors. On 10 November 2026, activewear brand Lululemon launched a limited run of two existing styles of running shorts in Australia and New Zealand, made with polyester recycled using enzymes from the Australian company Samsara Eco. According to Lululemon, the recycled polyester accounts for at least half the total weight of each pair. The release follows a limited-edition anorak launched in 2024.
Meanwhile, in France, Carbios completed its 100th demonstration batch in July and expanded its licensing offer to cover polyester textiles. Together, the developments show both sides of the picture: the shorts prove that recovered material can go into clothes people already buy, while the French trials show how the technology handles textile waste on a larger scale.
Packaging represents another major market. The European Union has adopted minimum recycled-content requirements for categories including single-use plastic drinks bottles, tied to a 2030 timetable. While negotiations on a United Nations plastics treaty continue, the EU rules give brands a concrete reason to secure supplies of recycled material. Carbios’s proposed commercial plant in Longlaville, north-eastern France, is designed to process up to 50,000 tonnes of prepared PET waste a year. In September, the company said most project lenders’ credit committees had granted approval, though financing still needed to be finalised before construction could resume.
How machine learning sharpened nature’s scissors
The scientific breakthrough began a decade ago with a bacterium found outside a bottle-recycling site in Sakai, Japan. In 2016, a team led by Shosuke Yoshida reported that Ideonella sakaiensis could survive on PET as a source of carbon and energy. It did so using two enzymes: the first, PETase, cut the plastic’s long chains into smaller fragments, while the second broke those fragments down further. In the laboratory, however, the bacterium took weeks to degrade a thin PET film – far too slow for an industrial facility.
To speed up the chemistry, a team at the University of Texas at Austin, including Hongyuan Lu and Hal Alper, turned to a machine-learning model called MutCompute. An enzyme is a chain of amino acids folded into a precise shape, and swapping even a single amino acid can alter how well that structure holds up. MutCompute analysed how amino acids fit into their local surroundings in a protein and suggested substitutions for researchers to test.
The resulting enzyme, named FAST-PETase, reduced some samples to their constituent molecules in as little as 24 hours. In the same study, PET samples from 51 used thermoformed products broke down within a week, though more tightly packed PET – such as sections of a drinks bottle – proved harder for the enzyme to attack without pretreatment.
Carbios took its lead from a different source: an enzyme originally identified in leaf compost. In a 2020 Nature study, its researchers and collaborators reported that an engineered version broke down roughly 90% of prepared PET waste in about 10 hours. The process is similar to taking apart a Lego castle by neatly unclicking its bricks – in this case, terephthalic acid and ethylene glycol, the two core ingredients used to make PET. Once purified, those ingredients can be joined again into new polyester. While a fast enzyme gets the process to that point sooner, the cost of recovering both ingredients in usable form helps determine whether the method can compete commercially.
Can biology compete with cheap oil?
One of the most expensive bottlenecks occurs after the enzyme has finished its work. Ethylene glycol is left dissolved in water, and separating the two requires substantial energy. In a 2025 study, researchers from the US National Renewable Energy Laboratory, the University of Portsmouth and the University of Massachusetts Lowell tested several modifications to bring costs down. By adding PET and enzymes in stages to the same reactor liquid, they allowed the glycol to build up to a concentration that made recovery more economical. They also altered how terephthalic acid was recovered, sharply reducing the need for fresh acids and bases.
In their model of an industrial plant, those changes lowered the minimum selling price for recycled PET to US$1.51 a kilogram, compared with a five-year US average of US$1.87 a kilogram for virgin PET. The researchers also estimated that the improved process would generate nearly half the greenhouse gas emissions of the earlier enzymatic method used as their benchmark. In practice, a commercial plant’s costs will depend on what it pays for waste and energy, how steadily it can operate and what buyers are willing to pay for its output.
Carbios is pursuing two routes to turn its engineering into a business. The Longlaville facility would allow it to produce recycled material directly, while licensing enables other companies to build plants using Carbios’s process designs and operating instructions. Clermont-Ferrand gives prospective licensees a working showcase where they can see how the system handles actual waste, including the less uniform textiles a commercial recycler has to buy.
Which plastics can enzymes recycle?
Enzymes can break down PET because its chains contain chemical links called ester bonds. By contrast, polyethylene and polypropylene – the plastics used in many bags, films, tubs and caps – are built on tougher carbon–carbon backbones. Enzyme-based recycling has yet to offer an industrial route for either.
A University of Konstanz study made that chemical divide visible. After nearly 15 months buried in forest soil, experimental polyester films still held their shape, yet an electron microscope revealed tiny cavities resembling bacterial cells. In separate laboratory tests using soil from the same forest, microbes converted almost all the carbon in two of the experimental polyesters into carbon dioxide within eight to 10 months. A third reached about 20%, while conventional high-density polyethylene barely changed. That experiment tracked biodegradation in soil; inside a recycling plant, operators want to harvest the plastic’s ingredients before microbes consume them.
The same selectivity creates practical work at the factory gate. A running shirt may blend polyester with elastane and carry a printed logo; a food tray may arrive with a peel-off film and traces of lunch. While the enzyme can break down the PET, facilities still need systems to sort incoming waste, prepare it and manage whatever remains.
Closing the loop
When Samsara Eco founder Paul Riley described his company’s partnership with Lululemon, he spoke of building ‘a fully circular ecosystem’. A pair of running shorts offers a tangible first step: polyester made from recovered ingredients can go into a garment designed for regular wear.
The next test comes years later, once that kit has been washed and worn until the seams give way. For the loop to close, the garment needs a route back to a recycler equipped to process it. If that becomes routine, an old running shirt could finish its useful life on the track and begin another as the raw material for the next.
Liked this article? You can support our independent journalism via our page on Buy Me a Coffee. It helps keep MoveTheNeedle.news focused on depth, not clicks.