From high school science project to $18.3M: AI-accelerated enzymes are coming for fast fashion’s plastic waste

A U.K. startup, originating from founder Jacob Nathan’s high school science project on using enzymes to break down plastic waste, has secured an oversubscribed $18.3 million in Series A funding.

Founded in 2019 in London, Epoch Biodesign now a 30+ strong multidisciplinary team of chemists, biologists and software engineers. It will be using the new funding to scale up production of their plastic-eating enzymes. This means transferring the biorecycling process from the labs where they’ve been developing it to their first production facility this year, which he says will be able to gobble through 150 tonnes per year of waste once it’s up and running.

Thereafter, the first production runs of commercial-scale capacity are expected by 2028 if not sooner, as Nathan says the startup is looking for ways to accelerate the scaling. They’ll be roughly doubling the size of the team over the next 12 months as they work on switching to a higher gear, he tells TechCrunch.

Plastic not-so-fantastic

Stepping back for a second, the world’s plastic waste problem is staggeringly vast, with some 400 million tonnes of the stuff produced annually, according to the UN. Only a tiny fraction of which gets recycled currently being as, in crude cost terms, it’s far cheaper to pump out more virgin plastic than deal with processing the stuff we’ve already produced.

At the same time, the environmental and health costs of unchecked plastic pollution are stark. So there is growing pressure on regulators to act on plastic pollution and on businesses that use plastic in their products to clean up their act.

There are also a growing number of startups working on technologies targeting plastic waste from various angles — including startups applying AI to speed up sorting plastics for recycling and others developing non-fossil fuel-based plastic alternatives. But biorecycling, so leaning on biological entities to help break down resistant waste, is where Epoch Biodesign hopes to make its mark on plastics.

The biotech is developing a library of plastic-eating enzymes with the goal of disrupting the plastic pollution cycle by powering up biorecycling-based circularity — starting with a handful of plastics that are used in common synthetic fabrics. The first materials they’ve developed enzymes to tackle are polyester and two types of nylon (nylon 6 and nylon 66).

A graphical animation of the process on its website depicts waste garments going in at one end, being industrially sorted and/or pre-treated, depolymerized, purified and repolymerized, and then ready-to-use nylon (extrusion) or polyester (pellets) coming out the other end.

GenAI to the rescue?

While some plastic-eating enzymes have been discovered existing in nature, the catch is they are very slow at digesting this stuff — far too slow to help humanity escape its plastic waste mountain on any useful timescale. It’s also the case that we have produced far more types of plastics than enzymes have been found in the wild that can break them down, as yet. And as the plastic keeps piling up, the need for speed increases.

Epoch wants to lend a helping hand to evolutionary ingenuity by using technology tools to accelerate the discovery of biological catalysts that can tackle plastic waste fast. And key to unlocking this mission are developments in generative AI — specifically the rise of powerful large language models (LLMs) — that are helping accelerate the search for biological agents that can be precision targeted at this problem.

“The challenge with biology is that it’s just too complicated,” explains Nathan. “Humans don’t understand how it works. We’ll never be able to rationalize it. Most of these biological questions that we have remain unanswered. So the big shift here has been our ability to understand large, complex data-sets — which is effectively AI.”

“We’re just sort of un-baking the cake and then putting things back together at the other end,” he also says of what this biorecycling process boils down to. He adds that it only takes a “matter of hours” to go from waste fabrics to reclaiming molecularly identical material (nylon or polyester) in a form that’s ready for reusing to make new clothes or other products.

He describes enzyme design as a “ridiculously large search problem” to tackle. But by turning to GenAI, the startup’s scientists have essentially been able to shortcut sifting through possible combinations of amino acid and proteins to land on potentially useful agents — fine-tuning LLMs with information on proteins and amino acids but also feeding in “proprietary data” from its own lab work on plastic-eating enzymes.

“We’ve been able to generate tens of thousands of plastic-eating enzymes in our lab that are unique,” he says, explaining that after querying the AI models to yield promising candidates they switch to lab tests and then feed in more data from their results on the “predicted enzymes” to keep iterating the model until the search turns up “an enzyme that performs in the way that we want.”

“What we’re effectively doing is we’re concentrating hundreds of millions of years, billions of years of evolution into a few cycles in the lab that happen over the course of days, weeks, months,” he adds. “We’re making big evolutionary jumps that would be very unlikely to happen just naturally based on random mutations, natural selection.”

Epoch’s AI-driven enzyme design search has also enabled it to “pretty regularly” get speed improvements on enzymes in the region of 25x, according to Nathan.

“That means we can use less enzyme in our process,” he notes. “We can make less of it. The [capital expenditure] associated with manufacturing that enzyme in the first place goes down. And ultimately, all of that translates into a lower cost of goods for output.”

“We’re not the only company out there which is trying to design biology to do different things … but we really think we’re quite unique in the approach we’re taking in applying these tool sets to recycling — and then to our flavor of recycling: biorecycling,” he adds.

Focus on cost and commercial scale

So far, the startup has built three “best-in-class processes to recycle three very chemically distinct types of plastics” — and scaling those to commercially useful volumes is next on the slate with the new Series A cash.

“We’re building our first production facility in the U.K. this year for our first nylon process,” he says, claiming: “These technologies use entirely new biochemistries. They completely shift the cost bases of recycling into new areas that basically makes recycling the cheaper option compared to virgin.”

A key part of why Epoch is able to drive down recycling costs is the fact its process doesn’t require high temperatures to run — saving on energy costs compared to other forms of recycling which require the waste to be heated and/or melted. Nathan also points out that this means a lower capex for this (lower power) recycling facility — shrinking overall project costs.

The biological recycling process is also “incredibly high yield” compared to industrial recycling — he says they’re getting upwards of 90%, meaning most of the waste that’s fed in is coming back out the other end in a reusable state.

Plus, there’s no “unwanted side products” from biorecycling — which, again, reduces the cost and complexity of recycling the plastic.

“All of these things add up, basically, to reduce cost across the board of the process and get us into a position where — at that commercial scale — we’re reaching cost competitiveness with the materials that are on the market today made from fossil carbon,” he suggests.

Production of the enzyme itself involves a microorganism that’s been genetically engineered to include the DNA for making the enzyme and housed in a fermenter so it can replicate and churn out lots of the plastic digesting stuff — a synthetic biology technique that’s used for many other types of applications, from producing chemicals to novel foods.

Epoch’s approach to recycling plastic could have some additional benefits as Nathan suggests it could incorporate additional purification — by having the enzymes also “scrub” undesirable chemicals — since some plastics contain chemicals that can cause concerns for recycling the material.

Although he concedes that even biorecycling of plastics won’t fix the problem of microplastics where tiny pieces of plastic can wash out of garments that are made from synthetic fabrics and find their way into the environment — posing a hazard to biological life.

Still, he argues we’re going to be stuck needing to use synthetic plastic for decades, adding: “I think it’s really important that that new synthetic plastic is made from old materials, not from newly extracted fossil carbon.”

Designing enzymes to digest other types of plastic waste — such as packaging — is a wider goal for the startup. Although Nathan says they are focused on fabrics first as it’s a huge problem that’s also been getting more public attention. The business case also looks cleaner.

Notably, the startup’s Series A includes a strategic investment by Spanish fast fashion giant Inditex, owner of clothing brand Zara, which has inked a multi-year “joint development agreement” with Epoch — clearly with an eye on improving the sustainability of its business at a time of rising public awareness vis-a-vis the fashion industry’s role in the global plastic crisis.

“We want to produce material that’s actually useful,” notes Nathan. “We want to produce something for brands that is, you know, indistinguishable from the stuff that they’re using today — so in order for that to be true, we need to go through various tests. We need to do this at larger and larger and larger scale. And so having, effectively, the machinery of a business like Inditex with the scale that they have just helps us accelerate that process.”

The Series A round is led by the climate-focused fund Extantia Capital, with Day One Ventures, Happiness Capital, Kibo Invest, Lowercarbon Capital and others also participating alongside Inditex, and a $1M grant from the U.K. government. Epoch Biodesign’s total capital raised to date is now $34 million, including the latest raise.