The extruder decides

Biology rarely loses to petrochemistry on chemistry. It loses at the extruder. The bio-based materials that reached scale arrived as a molecule the plant already knew; the ones that asked the plant to learn a new one are still waiting for someone to pay the tuition.

Two companies, one sugar

In December 2020 Danimer Scientific listed in New York at an enterprise value of about $890 million. Its product, PHA, is the textbook case for biology: bacteria ferment sugar or plant oil, store the polymer as granules inside the cell, and the finished part biodegrades. In March 2025 Danimer filed for Chapter 11. Two months later its plants, more than 480 patents and applications and its operating assets went to Teknor Apex for $19 million.

Over the same years Braskem did something unglamorous. At Triunfo in southern Brazil it dehydrates sugarcane ethanol to ethylene and polymerises it into polyethylene. In 2023 it spent US$87 million to lift that unit from 200,000 to 260,000 tonnes a year; by then it had made more than 1.2 million tonnes of bio-based polyethylene since 2010.

Both start from sugar. One ended in a write-off, the other in a capacity expansion.

What a new molecule costs before it is sold

Start with the resin. Industrial PHA costs $4–6 per kg to produce, against $1–2 per kg for polyethylene and polypropylene. The same review puts the carbon source at 40–60% of production cost and downstream recovery — breaking the cells and extracting the polymer — at up to 50%. Neither line falls the way a cracker's costs fall with size. Sugar is priced by the sugar market, and a fermenter's output is capped by how fast the cells grow and how much oxygen you can drive through the broth. Scale helps; it does not change the order of magnitude.

Then the change at the converter, which never appears in the price per kilogram. PHB, the base PHA, melts at 170–180 °C and starts to degrade at around 180 °C; the acceptable residence time in the barrel is a few minutes. A polyethylene line runs well above its melt and forgives a stoppage. A PHB line punishes dead spots in the die, long purges and every unplanned stop. Crystallinity sits between 40% and 80%, and secondary crystallisation keeps going at room temperature, so a film that passes inspection on delivery can turn brittle on the shelf.

Every one of these is solvable. Copolymers, blends and nucleating agents widen the window. But each fix is a new grade, and each new grade goes through qualification again: line trials, scrap, regrind that cannot go back into the hopper, then the brand owner's own testing. That bill lands on the converter and the brand. The producer does not pay it, and the producer's investor deck does not show it.

Now Braskem. Ethanol dehydration is one catalytic step upstream. After it, the ethylene is ethylene. Same polymerisation, same grades, same settings on the converter's line, same recycling stream at the end. Qualification shrinks to a certificate. The premium a brand pays buys a carbon claim, not a process change.

Avantium sits in the middle and is worth watching for that reason. PEF is a new molecule, so FDCA needed its own plant: 5,000 tonnes a year in Delfzijl, financed in December 2021 to run in 2024. Faulty titanium welds added €7 million and pushed start-up into 2026; the first commercial-scale FDCA came out in September 2026, with full capacity 12–24 months later. The novelty sits in Avantium's own plant, where Avantium controls it and pays for it. Downstream, PEF is a polyester aimed at the bottle and film lines the PET industry already runs. Whether those lines take it with minor changes is the real test of the next two years, more than the plant itself.

Somebody pays for novelty

Read the three cases together and a rule appears. Every piece of novelty in a material is a bill, and the question is who receives it. Biology scales when the novelty stays inside the fermenter or the reactor and the molecule that leaves the gate is one nobody downstream can tell apart from the old one. It also scales when the property gained is worth more than the change it forces — but only in the applications where that is true. Danimer had the second argument for a narrow set of products and was valued as if it had the first.

Biology rarely loses on chemistry. It loses at the extruder.

If you specify, buy or scale materials

1. Price the change, not the resin. Add line trials, scrap, lost regrind and requalification to the price per kilogram before you compare two materials.
2. Ask where the novelty lives. New feedstock with an old molecule is your supplier's risk. A new molecule is your risk.
3. Buy a new molecule only for the property only it delivers. Do not put it on commodity duty where it competes on price with something your line already runs.
4. Ask the recycler before the brand does. A drop-in inherits an existing end-of-life stream. A new molecule needs one built, or has to prove it does not spoil the one next to it.

What I cannot resolve

The drop-in route moves the risk rather than removing it. Bio-based polyethylene is tied to ethanol, and ethanol to sugar harvests and fuel policy; I cannot tell you whether biology can deliver bulk molecules at a price that survives a bad year in both. And a drop-in inherits the old molecule's end of life: bio-based polyethylene in a field is still polyethylene in a field. Integration buys speed. It does not buy a better fate for the product.

The question I keep coming back to, in The Plastic Shift and since, is what we do not have to change. For bio-based materials the honest answer is: as much as possible, downstream of the gate.

The most radical thing a new material can do is go unnoticed by the machine that runs it.