The press release always comes first.
A company announces a plant that will turn the plastic you cannot recycle - the greasy film, the multilayer pouch, the candy wrapper - back into oil, or fuel, or brand-new plastic. There is a ribbon. There is a big number, tens of thousands of tons a year. And there is a word doing quiet, heavy work in the sentence: advanced. Advanced recycling. Chemical recycling. The plant in Hebron, Ohio was one of the largest of its kind in the country. It shipped its first barrel of plastic-derived oil to a Shell refinery in April 2025.
By the middle of 2026 it was in “care and maintenance,” the industry’s phrase for switched off, after the state cited it for violations and stripped its recycling status for failing to convert enough plastic for three straight quarters.
That is not an unlucky plant. That is the genre.
The roll call
Once you learn to see it, the pattern is hard to unsee. Fulcrum BioEnergy’s flagship waste-to-fuel plant in Nevada opened in 2022, shut down in 2024 after its gasification system kept failing, and filed for bankruptcy owing hundreds of millions. Brightmark’s plastic-to-fuel plant in Ashley, Indiana filed for Chapter 11 in March 2025 with $178 million in debt, and now reportedly runs at about 5 percent of capacity. Encina cancelled a $1.1 billion plant in Pennsylvania. Alterra pulled out of a Pennsylvania site in 2025, where the landowner decided a data center was the safer bet. Braven’s North Carolina plant was ordered closed, its equipment carted off. In 2026 the pattern went transatlantic, with Viridor moving to exit its European chemical-recycling business entirely.
None of this is new. A 2017 report by the environmental group GAIA found that more than $2 billion had already been sunk into gasification and pyrolysis projects that closed or were cancelled before they ever ran commercially. A 2020 tally by the same group, covering at least 37 US chemical-recycling projects announced over 20 years, found that only about three were ever operational, and not one had been proven to recycle plastic at commercial scale. Six years later, that sentence still holds.
So here is the question worth sitting with. Why does a technology that keeps failing keep getting built, celebrated, and subsidized?
The pitch, and why it lands
The pitch is genuinely seductive, because it answers a real anxiety. You rinse your yogurt cup and you still suspect, correctly, that most plastic is not being recycled. Along comes a promise that sounds like the fix: a process that can take the hard stuff, the contaminated and mixed and multilayer plastic that mechanical recycling chokes on, and break it all the way down to molecules, then build it back up. Infinite loops. Nothing wasted. Keep buying, keep tossing, the machine handles it.
The two main routes are pyrolysis, which cooks plastic without oxygen into an oil, and gasification, which cooks it hotter into a synthesis gas. A newer variant chases hydrogen. The chemistry is real. Pyrolysis oil exists. You can hold a jar of it.
The trouble starts the moment you ask the plant to be a business.
What the evidence says when you check it
Strip out the marketing and read the peer-reviewed work, and the picture is consistent to the point of monotony.
Pyrolysis is energy-marginal. It has to burn a chunk of its own output just to keep the reactor hot, so the gross fuel yield badly overstates the net energy you actually get back. Only 38 to 55 percent of the carbon in the input plastic ends up in usable oil. The rest leaves as gas, char, and emissions. On lifecycle carbon it comes out better than incineration - though by how much is a moving target that studies disagree on sharply - and here is the finding the industry never puts on a banner: once you account for the quality of what comes out, pyrolysis does not beat plain mechanical recycling on climate at all, and it does worse on other environmental measures.
Gasification looks profitable only in models, and only when you feed the model a clean, pre-sorted plastic stream rather than the actual mixed garbage the technology is sold to solve. Even those flattering studies quietly ask for subsidies and tax credits to make the numbers work.
Plastic-to-hydrogen is the clearest tell. Without carbon capture, hydrogen made from mixed plastic is, at best, no cleaner than hydrogen from natural gas, and by some estimates dirtier - the peer-reviewed studies split on which. Add carbon capture and it gets clean, but that entire advantage vanishes the moment fossil hydrogen adds carbon capture too. You are paying a premium to arrive roughly where you started.
Meanwhile the lone bright spot the industry points to, ExxonMobil’s expanding operation in Baytown, Texas, publishes only its own nameplate capacity and forward targets. There is no independent, metered figure for how much plastic actually goes in and how much usable output actually comes out. In this field, self-reported capacity is not evidence. It is the press release again.
What these plants actually produce
Put the failure record next to the physics and a simpler explanation appears. Whatever any given company intends, the fuel cannot be the point, because the fuel is a rounding error. What these plants reliably produce, running well or barely running at all, is something else, and it has a far better market: permission.
Every announcement of an advanced-recycling plant is a reason for a brand to keep making virgin plastic and a reason for a legislator to not pass a law that would make less of it. The plant does not have to run well. It barely has to run. It has to exist, be photographed, and appear in a sustainability report as evidence that the waste problem is being handled downstream, so that no one has to touch the far more expensive question upstream: making less plastic in the first place.
Watch where the policy has moved and the function becomes obvious. In the United States, something like two dozen states - the count keeps climbing as more pass, and depending who is counting it sits somewhere between the mid-twenties - have reclassified pyrolysis as “manufacturing” rather than waste disposal, which loosens the rules it has to follow. In March 2026 the federal EPA proposed to do the same nationally: a proposed rule opened for public comment, not settled law, that would redefine these facilities as factories rather than incinerators and relax the pollution controls that came with the older label. The technology cannot clear a commercial bar, so the bar is being lowered to meet it. At the same time, the Global Plastics Treaty has stalled through round after round, with the central fight being exactly this: whether the agreement should cap plastic production or merely tidy up its end of life. The petrochemical position is end-of-life, and “advanced recycling” is the centerpiece of that argument. A technology that does not work is doing real, load-bearing political work.
That is the pattern. It is worth naming plainly, because it is not unique to plastic.
What actually works
The honest answer is not “nothing.” The research points clearly to what reduces plastic harm. It just points nowhere near where the money and the ribbons are.
The order matters, because leverage is highest at the top and the industry spends the most to keep your attention at the bottom.
Make less of it. Reduction, reuse, and refill are the only moves that touch the problem at its source. Everything downstream is damage control on a volume problem. This has no venture-scale exit, which is exactly why it is under-sold.
Recycle the mechanical way. Ordinary mechanical recycling already matches or beats chemical recycling on carbon, and unlike chemical recycling it runs at scale today. It has real limits. Quality degrades with each loop, and it cannot take the truly contaminated stuff. But it is the proven workhorse the “advanced” story is designed to distract from.
Design the waste out before it exists. Single-material packaging instead of multilayer laminates, refillable formats, and less packaging overall. You make plastic recyclable at the drawing board, not by trying to un-cook it later.
Keep the one honest niche honest. For genuinely non-recyclable, contaminated residue, pyrolysis can be a real improvement over landfill or incineration - if it is labeled truthfully as subsidized waste management, not sold as “circular” and not counted as recycling.
What you can do, at every level
Here is the part the ribbon-cuttings would rather you skip: what any of us can actually do. It runs from your kitchen to the treaty table, and the honest version comes with a warning.
You, the individual. Your instinct to help is real, and it is not misplaced - the care that makes you rinse the yogurt cup and sort the bin is exactly the right instinct. What was misplaced, deliberately, is where that care got aimed. The idea that plastic waste is mostly your fault - your litter, your sorting - was manufactured. The packaging and beverage industries funded the campaigns, going back to the 1950s and the famous 1971 ad of a man weeping over roadside trash, to move the crisis off their ledger and onto your conscience. You were not the problem. You were the alibi. So take the care that was aimed at your guilt and aim it where it works. Do the real things, and hold them at their real size. Refuse and reduce single-use plastic where you can. Reuse. Recycle mechanically and correctly, without “wishcycling” contaminated items into the bin, which makes recycling harder, not easier. Buy less packaging. But know that your single largest lever as an individual is not the bin. It is your voice as a citizen and your money as a signal. Back deposit-return laws. Support producer-responsibility rules. And when you see “advanced recycling” on a label or in a press release, let it raise your eyebrow rather than lower your guard.
Your household and community. Deposit-return schemes work where they exist; use them and vote for them. Support refill and reuse infrastructure where it appears, because it only survives if it gets used. And local permitting is where these plants live or die - a community that asks a proposed pyrolysis plant for independently metered throughput before it approves anything is asking the one question the industry cannot answer.
Companies and designers. The highest-impact decision is made long before the product ships: design for one material, for refill, for less. Label honestly. And stop buying permission - the offtake deal with an advanced-recycling startup that lets a brand keep virgin production flat while claiming progress is the corporate version of the whole trick.
Investors and allocators. The graveyard is your due diligence, free of charge. The opportunity was never in turning plastic into gas. It is in reduction, reuse systems, mechanical-recycling improvement, and material redesign - lower-margin, less glamorous, and actually real. When a pitch deck leads with self-reported capacity and a subsidy assumption, you have seen this movie, and you know how it ends.
Policy and regulators. Make producers pay for what they make, through extended producer responsibility. Redirect subsidies away from thermochemical plants that keep failing and toward the reduction and mechanical recycling that keep working. Do not relabel incineration as recycling to help a struggling industry clear a bar it cannot meet. And a real global treaty has to cap production, not just manage the end of the pipe, because the end of the pipe is where this entire illusion lives.
The tell that travels
Strip the chemistry away and what is left is a shape. A technology is sold as virtue. It is celebrated at the ribbon-cutting. It collects subsidies and rewrites the rules in its favor. And it works beautifully right up until the moment someone checks the claim against the primary source, at which point it turns out to have been producing permission all along.
If that shape feels familiar, it should. The same shape shows up, on a different substrate, in the governance of AI: the policy document that is not architecture, the audit trail built for the regulator instead of the person it is about, the ethics board with no authority to stop a launch. Greenwashing and governance-washing are not the same failure - one fails in a reactor, the other in an argument - but they are the same shape: buy the appearance of a solution so you can keep doing the thing the solution was supposed to stop.
Which gives you one test that works on both, and on nearly everything sold to you as responsible: ask what happens to the claim when you check it against the source. Ask for the metered number, not the nameplate. Ask what the thing actually does, not what it is named.
The solution that cannot survive an audit was never the solution. It was the cover story. And the plastic will keep coming until we stop applauding the cover and start funding the fix.
Sources and further reading
Every factual claim above was checked against a primary source before publication. The load-bearing ones are here so you can check them yourself.
The evidence (peer-reviewed):
- Petrik et al., life-cycle analysis of Dutch mixed plastic-waste pyrolysis, Waste Management & Research (2025) - carbon recovery of 38 to 55 percent: journals.sagepub.com
- Jeswani et al., Science of the Total Environment (2021) - pyrolysis compared with mechanical recycling and incineration: sciencedirect.com
- Lan & Yao, Communications Earth & Environment (2022) - gasifying mixed plastic waste for hydrogen, with and without carbon capture: nature.com
The track record:
- GAIA, Waste Gasification and Pyrolysis: High Risk, Low Yield (2017): no-burn.org
- GAIA, All Talk and No Recycling (2020): no-burn.org
- Fulcrum BioEnergy bankruptcy, Waste Dive: wastedive.com
- Brightmark Ashley, Indiana bankruptcy, Resource Recycling: resource-recycling.com
- Encina cancellation, Inside Climate News: insideclimatenews.org
- Alterra Sugarloaf withdrawal, Waste Dive: wastedive.com
- Freepoint Hebron shutdown, C&EN: cen.acs.org
- ExxonMobil Baytown’s missing metered figures, Inside Climate News: insideclimatenews.org
The policy:
- EPA’s proposed reclassification of pyrolysis, C&EN (March 2026): cen.acs.org
- State “advanced recycling” manufacturing laws, Waste Dive: wastedive.com