40,000 Tonnes a Year, Locked for a Decade: Why Plastics-to-Fuels Lives or Dies on Feedstock Security

Two announcements landed within a month of each other this spring, and read side by side they tell you most of what you need to know about why one plastics-to-fuels project gets built and another gets written off. In May 2026 Viridor moved to shut down Quantafuel, the chemical recycling platform it had assembled for around 90 million pounds [per Plastics News]. The Skive plant in Denmark wasn't broken. Its pyrolysis line was running dry yields of 70 to 75 percent, well above the 30 percent the industry had penciled in when the platform changed hands [EUWID Recycling]. Good technology. Dead project.

Three weeks later, on 9 June, a Texas developer did the opposite of folding. Abundia Global Impact Group signed a ten-year binding agreement with Frankfort Plastics for 40,000 tonnes a year of low-grade polyolefin film, about half the feedstock its first Cedar Port plant near Houston will need [GlobeNewswire]. Nothing has been poured yet. The definitive agreements aren't even due until Q3 2026. They locked the feedstock first, and that sequence is the whole argument for treating feedstock security, not the reactor, as the real gate for plastics-to-fuels.

The reactor was never the hard part

I've commissioned three waste-to-energy facilities and lost money on one of them, and the pattern I keep seeing is that the engineering is the part we already know how to do. Pyrolysis of polyolefins is chemistry from the 1970s, and the pyrolysis systems that run it are mature. Skive proved the yield. So did dozens of pilot lines before it. What Skive could not prove was an economic case, because two things sit at either end of the plant that no reactor can fix: what you pay for feedstock going in, and what you get for the product coming out.

Quantafuel got squeezed on both. Virgin polyethylene selling near $1,100/tonne sets the ceiling on what anyone will pay for pyrolysis output [market range, 2026], and when virgin resin is cheap, your low-carbon naphtha (the actual product, despite the "fuels" branding) competes against it on price and loses. Clean, sortable feedstock, meanwhile, is scarce and contested. Most failed waste-to-energy projects fail at offtake, not technology, and chemical recycling bolts a feedstock-side version of the same trap onto the front of the plant.

Advanced plastics recycling in the UK and Europe is "commercially uninvestable without policy changes," Viridor said in proposing the closure, citing weak demand, regulatory uncertainty, and competition from low-cost virgin plastics.

Policy is the long game. But you can't sign a lender up to "maybe the rules change." What you can do is take the two variables that killed Quantafuel and make at least one of them contractual before you ask anyone for money. That's what the Abundia deal does on the feedstock side.

What a ten-year feedstock contract actually buys

So why would a developer lock a decade of feedstock before pouring a foundation? Because lenders don't underwrite reactors. They underwrite cash flows, and a cash flow is only as good as the contracts beneath it. On the output side that means an offtake agreement; on the input side it means a feedstock supply agreement with enough term and volume that the debt-service model doesn't wobble every time a commodity index moves. I walked through how lenders actually price that risk in a piece on how waste-to-energy project finance gets underwritten, and the short version is that a binding ten-year supply deal is worth more to a credit committee (and to any ESG-compliant projects screen) than another two points of reactor efficiency.

Abundia's CEO, Ed Gillespie, put it plainly: long-term feedstock supply is "one of the core building blocks required to move from project development toward commercial execution" [GlobeNewswire]. That isn't marketing. It's the order of operations. You secure the molecules, then you finance the steel.

But a tonne contracted isn't the same as a tonne the reactor can use, and film is where that gap bites. Low-grade polyolefin film is filthy by design (it's the stuff too contaminated for mechanical recyclers to bother with). Reactors have a spec. Mura's Hydro-PRT process, for one, wants polyolefin content above 80 percent and contaminants held under roughly 5 percent [per published process reviews]. Getting there from dirty film means a pre-treatment line, optical sorting on something like a Tomra unit, shredding, washing, drying, before a single molecule sees the reactor. None of that shows up in a feedstock headline number. All of it shows up in your operating cost.

There's a regulatory reason a Gulf Coast project can pencil while a Scandinavian one couldn't. More than 20 US states have reclassified pyrolysis as manufacturing rather than waste processing, which changes the permitting path, while federal treatment under RCRA keeps shifting on whether the output counts as a fuel or a waste-derived product. Cheap Gulf gas and an existing petrochemical market next door don't hurt. Abundia got its feedstock locked, and it sited Cedar Port where the output molecules already have buyers within trucking distance.

The assumption that moves

Here's the part I'd stress-test if this were my project. Feasibility is mostly figuring out which assumption will move, and a contract doesn't freeze an assumption so much as relocate the risk. So which assumption moves here? Frankfort Plastics is a real operator, recycling film since 2017 and handling more than 13,000 tonnes a year today [GlobeNewswire]. The deal asks them to deliver 40,000. That's roughly triple their current throughput, on spec, every year for ten years. The price risk just turned into a delivery-and-quality risk, and that kind doesn't surface until year two or three.

I learned to watch that gap the expensive way. On the Monterrey MSW-to-energy project in 2017 I assumed the diesel backup generator would never run past the hours the OEM warranty baked in. It ran 380 hours in year one because grid stability was worse than the utility's own data showed, and I had badly underestimated it. Cost us about 74,000 dollars in unplanned fuel and warranty argument. Small number against a project that size. But it was the assumption nobody had stress-tested, and that's almost always the one that moves.

None of this makes the Abundia model a template, and it doesn't hold everywhere. A project without a Frankfort-sized aggregator within trucking distance has no one to sign a ten-year deal with, which describes most of the developing-world markets I've advised in, where waste conversion facilities stall for want of an aggregator, not for want of waste. Below maybe 50,000 tonnes a year of plant capacity the pre-treatment line quietly eats the economics, because sorting and washing don't scale down cleanly; a 20,000-tonne plant still needs most of the same optical sorters and washers a 200,000-tonne one does. And a feedstock contract does nothing for the output side. If virgin resin stays cheap and no recycled-content mandate pulls demand, you've de-risked the input to a plant whose product still can't clear. Quantafuel had decent feedstock too.

So I'm not watching Cedar Port's reactor. The technology will work; it usually does. I'm watching one number between now and Q3 2026: whether the definitive agreements actually bind Frankfort to triple its output on spec, or quietly soften to "commercially reasonable efforts." Hold the first, and Abundia has done something the European platforms didn't. Soften to the second, and they've signed a press release. We'll know which by autumn.

Sources & Notes

The deal terms, volumes, the Q3 2026 timing, and the Gillespie quote come from Abundia's 9 June 2026 announcement. For the Quantafuel closure, the "commercially uninvestable" line and the acquisition history, I leaned on Plastics News' reporting on the proposed shutdown and EUWID's coverage, which is also where the Skive 70-75 percent dry-yield figure surfaced. The feedstock-spec numbers (polyolefin content, contaminant limits) are summarized from a 2025 review of polyolefin chemical recycling. The Monterrey figures (the 74,000-dollar overrun, the 380 generator-hours) are from my own project records, not a published source.