The standard account of the fermentation bottleneck says the answer is contract manufacturing: someone builds tanks, everyone else rents them. Our capacity map was largely an exercise in tracking who is building those tanks and how far behind schedule they are.
For mycoprotein, that is not what is happening. Every commercial-scale plant in the set below that is running or fully funded — Enifer’s, ENOUGH’s and Planetary’s — has been built beside an existing sugar refinery, starch plant or ethanol refinery, by the company that will use it. Not one of them is a contract manufacturing arrangement. The one plant in the table that is not co-located, The Protein Brewery’s, is demo-scale and argues its cost advantage on different grounds.
And the people doing it say so explicitly. Planetary cofounder and CEO David Brandes, whose company runs a plant next to a Schweizer Zucker sugar mill in Aarberg, told AgFunderNews: “In the food space, unit economics are everything and for bulk fermented commodities, contract manufacturing is not viable.”
That is a direct contradiction of the received wisdom, from someone with a plant in the ground. It is worth taking seriously.
The map
| Company | Site | Co-located with | Disclosed capital cost | Nameplate output | Status |
|---|---|---|---|---|---|
| Enifer | Kantvik, Kirkkonummi, Finland | A sugar refinery; phase one runs on sugar-industry side streams | €33m projected | Up to 3,000 t/yr (500 kg/h at full rate) | First four-tonne commercial run completed by May 2026; operational in H2 2026 |
| Enifer + FS | Brazil | Corn ethanol refinery — thin stillage as feedstock | Not disclosed | 500 t/yr demonstration plant | Completion foreseen by end-2026 |
| ENOUGH | Sas van Gent, Netherlands | A Cargill starch plant supplying fermentable sugars; Cargill’s bioethanol plant next door takes the sugary wastewater | Not disclosed | 10,000 t/yr, working to double to 20,000 t/yr | In production since summer 2023, per February 2024 reporting |
| Planetary | Aarberg, Switzerland | Schweizer Zucker sugar mill | Not disclosed; industrial installations financed entirely with non-dilutive debt | Not disclosed | Operational nine months after first equipment delivery |
| Planetary + Dhampur Bio Organics | India (proposed) | DBO sugarcane processing site | Not disclosed | Not disclosed | Exploratory; target below $1/kg |
| The Protein Brewery | Breda, Netherlands | Not co-located, per available sources; uses non-sterile fermentation | Not disclosed | 600 t in 2027, above 2,000 t by 2029 | Demo-scale |
Methodology note. Every figure above is taken from a source we fetched and read, listed at the end. Where a company has not published a number we have written “not disclosed” rather than estimating. The ENOUGH capacity figures come from February 2024 reporting and we found no fetched 2026 source updating them — see What we could not establish.
Why the tank goes next to the mill
Four separate advantages stack, and only the first is the obvious one.
Feedstock at the gate. ENOUGH’s plant takes fermentable sugars directly from the Cargill starch facility it sits beside. Enifer’s Kantvik phase one runs on sugar-industry streams from the refinery it is located at. Nobody is trucking dextrose.
A waste sink. This is the underrated one. ENOUGH centrifuges its biomass, and the sugary wastewater goes to Cargill’s bioethanol plant next door. In a standalone plant that stream is an effluent cost. Co-located, it is a transfer.
Shared utilities. Enifer’s Kantvik factory is a brownfield project inside an existing industrial building that already had steam, electricity, process and cooling water, and waste-water treatment. Those are line items that do not appear in the capital budget because they were built decades ago.
Energy integration. Planetary’s stated case for India names it directly, alongside abundant sugar side streams and low labour cost: energy can be integrated with the sugar mill.
Practical consequence: if you are evaluating a biomass-fermentation business, “where is the site” is a unit-economics question, not a logistics question. A greenfield plant on a clean site is carrying four cost lines its co-located competitor is not.
The one capital number anybody has published
Enifer is the only company in this set that has put a capital cost and a nameplate capacity in public in the same sentence: a projected €33m for a plant making up to 3,000 tonnes a year.
That is €11,000 per tonne of annual capacity.
It is worth converting into the unit the industry argues about. Straight-line, ignoring financing cost, at full nameplate utilisation:
| Assumed asset life | Annual capital charge | Capital cost per kg of output |
|---|---|---|
| 10 years | €3.30m | €1.10/kg |
| 15 years | €2.20m | €0.73/kg |
| 20 years | €1.65m | €0.55/kg |
| 15 years at 70% utilisation | €2.20m | €1.05/kg |
This is our arithmetic on Enifer’s disclosed figures, not Enifer’s cost model. Enifer has not published a cost per kilogram and we are not attributing one to it. Straight-line depreciation with no cost of capital is deliberately generous; a real weighted cost of capital pushes every figure up. Utilisation in the first years of a ramping plant is well below nameplate.
What the table is for is calibration. Before feedstock, energy, labour, downstream processing, packaging or overhead, a plant built at Enifer’s disclosed capital intensity is carrying somewhere between roughly €0.55 and €1.10 per kilogram in capital charge alone — and Enifer’s number is the cheap end of the range, because it is a brownfield conversion inside a building that already had its utilities.
What that means for the sub-$1/kg claim
Planetary and Dhampur Bio Organics have said they expect to make mycoprotein in India at industrial scale at below $1/kilo. That figure has been repeated widely and it is the most aggressive cost claim in the category.
Run it against the capital arithmetic and the claim is not absurd, but it is demanding. At Enifer’s €11,000 per annual tonne, capital charge alone consumes the entire dollar at a ten-year life. For sub-$1/kg all-in, the Indian plant has to be materially cheaper to build per tonne than a Finnish brownfield conversion, or run at a longer asset life, or both — before a single input is paid for.
Three things could plausibly deliver that, and Planetary has named two of them: lower construction and labour cost, and energy integration with the mill. The third is that licensing the BioBlocks platform to an incumbent means the incumbent already owns the land, the utilities and the effluent infrastructure, so the marginal capital is fermentation and downstream only.
Practical consequence: when a sub-$1/kg claim appears, the question to ask is not about yield. It is what is inside the boundary. If the capital, the utilities and the effluent treatment belong to the sugar mill and only the marginal fermentation cost is being counted, the number can be true and still not be the number you would pay to replicate it.
Do not import precision-fermentation feedstock maths
One correction, because it is an easy error and we have published the underlying numbers ourselves.
Our analysis of sugar as a feedstock worked through GFI’s case-study assumptions for precision fermentation — a yield of 0.13 grams of product per gram of glucose, meaning 7.7 kilograms of sugar per kilogram of protein, and roughly $4.62 of feedstock per kilogram at $0.60/kg glucose.
Those figures do not transfer to mycoprotein. In precision fermentation the product is a single protein secreted by the organism and then purified out; most of the substrate goes into cell mass that is discarded. In biomass fermentation the cell mass is the product. The yield on substrate is far higher because nothing is being thrown away, which is the central reason biomass fermentation is economically viable at industrial scale today while precision fermentation, in Brandes’s words, “is a field requiring further improvements in productivity.”
Applying a precision-fermentation feedstock ratio to a mycoprotein cost model will produce a number several times too high. We have seen it done.
The contract manufacturing argument, properly stated
Brandes’s position deserves the full version rather than the quotable line. His case has three parts:
- Control. “Category leaders controlling the full value chain… are emerging as the winners. We have seen companies failing due to multi-party dependencies and too little self-control.”
- Ownership. “The production infrastructure needs to be owned or co-owned or at least exclusively accessible.”
- Retrofit scepticism. Retrofitting existing equipment for upstream processing “rarely works, does not safe substantial capex, and usually drives up COGS.”
Point three is the sharpest, and it cuts against a comfortable assumption in this industry — that idle brewing, pharma or industrial fermentation capacity can be repurposed cheaply. Brandes says it cannot. Enifer’s Kantvik project is a partial counter-example: it is a brownfield conversion inside an existing industrial building, and it is the cheapest disclosed build in the set. But Enifer installed a new purpose-built fermenter — over 12 metres tall and more than 4.5 metres in diameter — rather than repurposing someone else’s vessel. Brownfield site, greenfield equipment. That distinction is doing a lot of work.
The counter-argument
The strongest case against the co-location thesis is that it is an artefact of what this category currently makes.
Mycoprotein is a bulk commodity sold by the tonne at prices that must eventually compete with chicken. At those unit values, transport and effluent costs are a meaningful fraction of the price, and co-location is close to mandatory. Precision-fermented proteins are worth an order of magnitude more per kilogram — our review of the public cost models found published estimates spanning under $20/kg to $15,000/kg — and at those values the logistics arguments weaken considerably. Contract manufacturing may be perfectly viable for high-value proteins and non-viable for bulk biomass, in which case Brandes is right about his own category and wrong as a general rule. He does in fact scope his claim that way: “for bulk fermented commodities.”
There is also a survivorship problem with the table above. It lists plants that got built. Companies that tried to scale via contract manufacturing and failed are not in it, but neither are companies that tried co-location and failed, and we have not attempted to construct the denominator.
And demand is not settled. Quorn — the incumbent, and the only producer with decades of commercial biomass fermentation behind it — has been reporting lacklustre sales. Brandes’s answer is that alt-meat sales are growing in Germany, France and Italy and that hybrid products are the larger opportunity. That is a demand thesis, and every plant in the table is underwritten by it.
A process contradiction worth flagging
In June 2024 The Better Meat Co told AgFunderNews that moving to continuous fermentation had cut its at-scale cost by more than 30%, and that it could produce 68% more mycelium in the same time than a year earlier. Cofounder Paul Shapiro added: “To our knowledge, nobody aside from Quorn is doing continuous biomass fermentation in the mycoprotein space.”
In February 2024, four months earlier, ENOUGH CEO Jim Laird described his own process to the same publication: a 200,000-litre tank, inoculated, doubling every four to six hours, “and then we start to harvest on a continuous basis.”
Both statements cannot be straightforwardly true. The likely reconciliation is that the two companies mean different things by “continuous” — harvesting continuously after a batch growth phase is not the same as continuous feed and continuous draw at steady state, and the industry uses the word for both. Neither company has published enough process detail to settle it, and we are not going to adjudicate it on the basis of two interviews.
Practical consequence: in diligence, “continuous fermentation” is not a specification. Ask for the feed regime, the dilution rate and the run length between turnarounds. Those are answerable questions and the word on its own is not.
What we could not establish
- Capital cost for five of the six plants. Only Enifer has published one. Every per-kilogram capital figure in this piece rests on that single data point and is presented as arithmetic on it, not as an industry norm.
- ENOUGH’s current capacity. Our fetched source is from February 2024 and gives 10,000 t/yr going to 20,000 t/yr. Secondary summaries encountered during research cite 60,000 tonnes a year by 2027 and a cumulative one-million-tonne ambition by 2033; we could not verify the 60,000 figure from a source we fetched and have not used it.
- Planetary’s capacity and capital cost. Neither is disclosed. We know the Aarberg plant became operational nine months after first equipment delivery and that the industrial installations were financed with non-dilutive debt, but not what they cost or how much they make.
- Whether the Planetary–DBO deal exists. It was described in January 2026 as exploratory and “subject to final agreements and regulatory approvals.” We found no fetched source confirming a signed agreement.
- What is inside the sub-$1/kg boundary. Whether it is a full loaded cost or a marginal production cost at a licensee’s existing site is the single most important unknown in this article, and no public source resolves it.
- Enifer’s schedule slip. The 2024 announcement set factory completion for end-2025; the May 2026 reporting says operational in H2 2026. We have not established the cause.
What to watch
- Whether Enifer’s Kantvik plant reaches nameplate. 3,000 t/yr means 500 kg/h sustained. The first commercial run was four tonnes. The gap between a four-tonne run and a continuous 500 kg/h line is where most of the risk in this table sits.
- The Enifer–FS demonstration plant in Brazil. 500 t/yr on corn ethanol thin stillage, due by end-2026. If it works, it extends the co-location thesis from sugar to ethanol and from Europe to a low-cost geography — and it is a demonstration plant explicitly designed to answer the process-integration and economics questions.
- A signed Planetary–DBO agreement, and any cost disclosure with it. A sub-$1/kg claim with a published boundary would be the most valuable single number this category could produce.
- Whether anyone builds a mycoprotein plant that is not co-located. The Protein Brewery’s Breda facility is the closest thing to a control in this set, and it argues cost advantage from non-sterile fermentation rather than from siting. If it hits 2,000 tonnes by 2029 on that basis, the co-location thesis is a preference rather than a requirement.