The machine that makes the fiber
Nearly every vegan burger, nugget and strip on a supermarket shelf passes through the same piece of equipment: a twin screw extruder. It looks like an industrial pasta press, a long steel barrel with two interlocking screws turning inside it. Dry protein powder and water go in one end. Two screws push the mixture forward while the barrel heats it, in stages, up to around 150C. The combination of heat, pressure and mechanical shear unfolds the protein's native structure, a process called denaturation, and forces the unfolded strands to bump into each other and reform new bonds. What comes out the far end depends almost entirely on how much water was in the mix and what shape of opening, the die, it was forced through.
Low-moisture extrusion: TVP
Keep the water content low, roughly 20 to 30 percent going into the barrel, and the extrudate puffs up as it exits into normal air pressure, the way puffed cereal does. It comes out porous and spongy rather than fibrous. Manufacturers dry this down to about 6 to 10 percent final moisture, and the resulting dry granules, chunks or crumbles are textured vegetable protein, TVP, sometimes called textured soy protein when soy is the base. TVP is what goes into a bag of dried "vegan mince" that a cook rehydrates with hot water or stock, and it is also the filler used inside sausages, nuggets and some burger patties to add bite and structure without doing the expensive work of building real muscle fiber. It has been made this way, largely unchanged, since Archer Daniels Midland scaled up commercial soy TVP production in the 1960s.
High-moisture extrusion cooking: the whole-muscle texture
To get a texture that pulls apart in strands the way chicken breast or a steak does, manufacturers use high-moisture extrusion cooking, sometimes labeled HMEC or HMMA (high moisture meat analogue). The mix going into the barrel carries 40 to 80 percent water. Instead of exiting into open air, the hot, still-liquid protein mass is pushed through a long, narrow cooling die, sometimes a meter or more, that is actively chilled while it holds the material under continued pressure. As the protein cools inside that die, the shear forces from the extruder screws have already aligned the unfolded protein chains into parallel sheets, and the cooling locks those sheets in place before the mass can relax back into a random tangle. The result is called anisotropic structure, meaning the material has a directional grain, exactly the property that makes an animal muscle fibre pull apart lengthwise rather than crumble. This is the process behind Beyond Meat's whole cuts and most of the branded vegan chicken strips and pulled pork sold in supermarkets. It is a genuinely different piece of engineering from TVP, not a wetter version of the same thing, because the fibre only forms while the material is under pressure inside that cooling die.
Three routes that skip extrusion entirely
Extrusion dominates the category but it is not the only way to build muscle-like texture from plant protein.
Shear cell technology, developed at Wageningen University in the Netherlands and commercialized by the spinoff Rival Foods, uses a heated cylinder rotating inside a second heated cylinder rather than a screw barrel. The protein slurry, usually soy, pea or wheat based, sits in the gap between the cylinders and gets stretched and folded as the inner cylinder turns. That lower, more controlled shear builds layered, fibrous structure without needing a binder to hold smaller pieces together, and it can produce larger single pieces, like a chicken fillet, in one pass rather than needing to be reassembled from strands.
Mycoprotein is a different process altogether: fermentation, not extrusion. Quorn grows the filamentous fungus Fusarium venenatum in large air-lift fermentation tanks, feeding it glucose the way a brewery feeds yeast sugar. The fungus grows as long, branching threads called hyphae, and those threads are naturally fibrous, so the texture Quorn is known for comes from the organism's own growth structure rather than from mechanically forcing protein into alignment. After fermentation the biomass is heat treated to reduce its RNA content, then centrifuged into a paste, combined with a binder such as egg albumin (Quorn's vegan lines use alternatives) and formed into the final product.
3D printing is the newest and smallest-volume route, aimed specifically at whole cuts like steak that extrusion still struggles to replicate convincingly. Barcelona-based Novameat uses what it calls micro-extrusion, depositing extremely fine plant-protein fibres layer by layer to build up a steak-shaped structure with visible marbling. Redefine Meat, an Israeli company, 3D prints a formulation of soy, pea and chickpea protein with beetroot and coconut fat into steak and lamb flank products served at restaurants in the UK, Germany, the Netherlands and Israel. Both remain restaurant and foodservice products rather than supermarket staples, because the printers are slow and expensive compared with an extrusion line running continuously.
Why isolates, not flours, and where the fat comes from
The protein source is almost always sold to manufacturers as an isolate rather than a flour or a concentrate. Soy flour is roughly half protein by weight and still carries the bean's fiber, starch and oils. Soy protein isolate is processed to strip almost everything else out, reaching around 90 percent protein, and pea protein isolate is made the same way, by washing pea flour with water to separate the soluble protein from the starch and fibre before drying it back into a powder. Manufacturers pay for that extra processing step because extrusion behavior depends on protein concentration: a flour's leftover starch and fibre interfere with the unfolding and realignment that builds fiber, and batch-to-batch consistency matters enormously on a production line that a small viscosity change can jam. Pea protein has grown fast as a soy alternative in this category specifically because it dodges the allergen and it is comparatively high in the amino acid leucine.
Fat delivers mouthfeel and carries flavor, and the near-universal choices are coconut oil and cocoa butter, both solid at room temperature and melting close to body temperature the way animal fat does, which matters for how a patty feels as it cooks and cools. Manufacturers blend these with a smaller amount of a liquid oil, often sunflower or canola, to fine-tune the melting profile and keep the fat from seizing into hard lumps.
Methylcellulose: the binder nobody has replaced yet
Extruded fibres alone do not hold a patty or a nugget together, especially through the shrinkage of cooking, so most products add a binder. The industry default is methylcellulose, a plant-cellulose derivative labeled E461, prized because it gels when heated rather than when cooled, the opposite of gelatin. That means a raw vegan patty holds its shape as it goes onto a hot grill, exactly when a burger most needs structural help. Methylcellulose has also become the category's most visible reformulation target: it reads as a synthetic-sounding additive on a label, and both retailers and brands are moving away from it. Beyond Meat's newer lines, including Beyond IV and Beyond Ground, have dropped methylcellulose in favor of shorter ingredient lists, part of a broader clean-label push that accompanied the company's 2026 rebrand from Beyond Meat to Beyond. Ingredient suppliers are racing to fill the gap: researchers have tested enzymatically treated plant fibres and sodium alginate as substitutes, and companies including ICL and DSM have launched proprietary alternatives built from ingredients like curdlan and rapeseed protein. None has yet matched methylcellulose's combination of cost, heat-set gelling and regulatory familiarity closely enough to fully displace it.
Leghemoglobin: Impossible's heme, made in yeast
Impossible Foods' signature ingredient is soy leghemoglobin, a heme-carrying protein that soybean plants naturally make in their root nodules, in tiny quantities, to manage oxygen for the bacteria that fix nitrogen there. Impossible could never harvest enough from actual soy roots to supply a burger line, so the company inserted the soybean gene that codes for leghemoglobin into the yeast Pichia pastoris (now formally renamed Komagataella phaffii) and grows that engineered yeast by fermentation, the same basic technique used to make insulin. The yeast is then broken open and the leghemoglobin protein purified out. The heme group is chemically the same molecule found in animal myoglobin and hemoglobin, and it is what makes an Impossible patty smell and taste more like cooked meat, and appear to bleed, as it browns.
The regulatory record here is genuinely contested, not settled quietly in the background the way most food ingredients are. Impossible filed a GRAS notice for soy leghemoglobin (GRN 000737) with the FDA in 2017. The FDA issued a letter in 2018 saying it had no further questions about the company's self-determination of GRAS status, but the agency was explicit that a no-questions letter is not an FDA finding that the ingredient is safe, only that it saw nothing in the submission to challenge. Impossible separately sought and won approval of soy leghemoglobin as a color additive, since the ingredient also affects the product's appearance, and the Center for Food Safety filed formal objections to that color additive approval in September 2019 and a lawsuit in March 2020, arguing the FDA had not required long-term animal studies and pointing to preliminary data Impossible itself submitted that flagged findings in rats worth further study. An appeals court upheld the FDA's approval. Soy leghemoglobin remains on the market in the US under that approval, and the dispute over whether the FDA's process was thorough enough has not fully gone away.
Cultivated meat is a different technology altogether and is not vegan: it is animal muscle and fat cells grown in a bioreactor rather than raised as a whole animal, and it still comes from an animal cell line. It gets mentioned alongside plant-based meat because it aims at the same texture and flavor problem from the opposite direction, growing the real tissue instead of building an imitation of it. The approval record, company by company and country by country, is tracked on Vegan Brief's cultivated meat timeline.
Is vegan meat ultra-processed? Yes, and here is the honest nuance
Under the NOVA classification system, the one most nutrition researchers and journalists use, extruded vegan meat qualifies as ultra-processed. It is built from isolated protein rather than an intact food, it uses industrial processes no home kitchen has, and it typically includes additives like methylcellulose, flavors and colors that exist to fix problems the processing itself created. Vegan Brief is not going to pretend otherwise or hide behind the technical argument that heme or fibre makes it a special case.
TVP, HMMA, mycoprotein and cultivated meat compared
| Process | Texture | Typical protein (per 100g, cooked basis varies) | Common products |
|---|---|---|---|
| Low-moisture extrusion (TVP) | Porous, spongy chunks or crumbles, rehydrated before use | Roughly 50 to 55g per 100g dry weight | Dried vegan mince, sausage and nugget fillers, meat extenders |
| High-moisture extrusion (HMMA/HMEC) | Layered, anisotropic fibres that pull apart lengthwise like whole muscle | Roughly 15 to 20g per 100g finished product | Beyond Meat and other burger patties, vegan chicken strips and pulled products |
| Mycoprotein fermentation | Naturally fibrous from fungal hyphae, firm and chewy | Roughly 11 to 13g per 100g finished product | Quorn pieces, fillets, mince and nuggets |
| Cultivated meat (not vegan) | Depends on scaffold and cell type; aims to match native muscle exactly | Comparable to conventional meat, varies by cut | Restaurant-only chicken and beef products in a handful of markets |
Vegan meat's manufacturing has kept improving while its US sales have not followed. Retail dollar sales of plant-based meat and seafood in the United States fell around 10 percent in 2025 compared with 2024, with unit sales down roughly 11 percent, according to SPINS retail data tracked by the Good Food Institute, and refrigerated plant-based burgers specifically dropped about 26 percent year over year in the same period. The picture outside the US looks different: global plant-based food sales rose about 3 percent from 2024 to 2025, reaching roughly 28.9 billion dollars, suggesting the contraction is concentrated in the American market rather than universal. Analysts covering the category through late 2025 and into 2026 point to price relative to conventional meat and inconsistent eating quality as the two most cited reasons shoppers have pulled back, which is part of why methylcellulose removal and other reformulation work described above has become a competitive priority rather than a marketing footnote.

