Ajinomoto Foods Europe sells contract fermentation from a plant at Nesle, in northern France. Its Industrial Fermentation Services website lists what it offers under a heading called Technologies. There is one line item for culture size:
160 M³ INDUSTRIAL CULTURES
On 8 July 2026, AMSilk announced a long-term manufacturing and supply agreement under which Ajinomoto Foods Europe will establish a dedicated line for its silk proteins at Nesle. The release describes the setup precisely: it “is based on 160 m³ fermentation reactor capacity, combined with customized downstream processing capabilities, and is designed exclusively for AMSilk’s production.”
The same number, on the same site, now committed exclusively to one customer. That customer makes materials for textiles, automotive and consumer care. It does not make food.
What is actually stated, and by whom
It is worth separating what each party has published from what a reader might infer.
| Claim | Stated by | Where |
|---|---|---|
| Nesle offers “160 M³ INDUSTRIAL CULTURES” | Ajinomoto Foods Europe | AFE Industrial Fermentation Services site, Technologies section |
| Nesle “includes cultures of more than 100 m³ and their associated equipment” | Ajinomoto Foods Europe | Same site, About section |
| Large-scale industrial fermentation services offered since 2022, bioreactors above 100 m³ | Ajinomoto Foods Europe | About AFE boilerplate, 8 July 2026 release |
| The AMSilk line “is based on 160 m³ fermentation reactor capacity” | AMSilk | 8 July 2026 release |
| That line is “designed exclusively for AMSilk’s production” | AMSilk | 8 July 2026 release |
| Target of “hundreds of tons per year and beyond” | Hiroshi Kaneko, President, AFE | Quoted in the 8 July 2026 release |
| Standing Ovation’s animal-free casein is produced at Nesle under a long-term deal | Standing Ovation | Green Queen, 9 January 2025 |
| Standing Ovation is “expected to triple our production capacity with the expansion of our partnerships, and scale our production to several hundred tonnes” | Yvan Chardonnens, CEO, Standing Ovation | Green Queen, 9 January 2025 |
What nobody states is how many 160 m³ cultures exist at Nesle. The company’s own site names the size, not the count. It says “cultures” in the plural in one place and gives a single figure in another. Both statements are consistent with two vessels and with twenty.
Practical consequence: if you are evaluating Nesle as a route to commercial volume, the published capacity figure is a vessel specification, not an available quantity. Ask for the count, the current committed fraction, and the earliest uncommitted campaign window — in writing — before treating “over 100 m³ bioreactors” as capacity you can access. This is the same failure mode we found in nameplate volumes reported as output and in a 4x capacity multiple with no stated base.
Two customers, the same site, the same target
The striking thing is not that a contract manufacturer has more than one customer. It is that two of them have publicly announced the same order of magnitude of output from the same plant.
- AMSilk, per AFE’s own president, is being supported “toward hundreds of tons per year and beyond”, on a line described as exclusive.
- Standing Ovation, per its CEO in January 2025, expects to “scale our production to several hundred tonnes”, having described the Nesle location as close to its main customers, “world leaders in the dairy industry”.
A caveat on that second figure, because it matters. Standing Ovation’s CEO said the company was “expected to triple our production capacity with the expansion of our partnerships” — plural. The several-hundred-tonne target is therefore attached to its partnership network as a whole, not stated as Nesle output alone. We are not going to present it as a Nesle number when the source does not.
Even read conservatively, though, two customers are publicly pointing at the same plant with volumes in the same order of magnitude, and the arithmetic that would reconcile them has not been published by anyone. That is entirely compatible with a site that has several 160 m³ trains, which is the most likely reading. It is also compatible with a queue.
The 8 July release also contains a forward-looking clause that is easy to skim past: the agreement “further provides a framework for future expansion through the addition of precision fermentation capacity in the coming years.” Read plainly, that says the existing footprint is not sufficient for where this is going, and that new tanks will be needed. It is a capacity announcement disguised as an option.
Why non-food wins the tank
Here is the part that matters for anyone building a food protein business in Europe.
A fermentation vessel does not care what protein it makes. The economics of who gets to use it are set almost entirely by what the output sells for per kilogram, because the fixed cost of a campaign — sterilisation, media, downstream, cleaning, changeover, quality release — is broadly similar whatever the molecule.
Consider what these two customers are selling into:
| Customer | End markets named in the source | Competing against |
|---|---|---|
| AMSilk | Textiles, automotive, consumer care | Performance materials |
| Standing Ovation | Cheese, yoghurt, ice cream | Dairy casein |
A performance material for automotive interiors and luxury textiles is priced against other performance materials. A precision-fermented casein is priced against casein. Our earlier work found that formulators buying dairy protein are paying $12 to $13 per kilogram for WPC 80%, not the commodity whey price that gets quoted at conferences — and that is the ceiling a fermented dairy protein has to work under.
We do not have a published price for AMSilk’s silk protein, and we are not going to invent one. But the structural point does not require it. When a contract fermenter allocates a scarce, capital-intensive asset, the application that can absorb the highest cost per kilogram gets the campaign. Food is, and will remain, the lowest-value application of a precision fermentation tank. Textiles, cosmetics, pharma intermediates and specialty chemicals all sit above it.
Practical consequence: food protein companies competing for third-party European fermentation capacity are not primarily competing with each other. They are competing with cosmetics, biomaterials and specialty chemicals, and they lose that competition on price per kilogram every time. Any capacity plan that assumes a CMO will be available on a food-protein margin is assuming away the CMO’s own incentive.
The exclusivity word
“Designed exclusively for AMSilk’s production” is doing real work in that sentence, and it is the strongest phrase in the release.
An exclusive dedicated line is not a booking. It is a carve-out. The equipment is configured for one process, with “customized downstream processing capabilities” — cell separation, evaporation, crystallisation and drying are what AFE lists — tuned to one molecule. Reconfiguring it for a different product is not a scheduling decision; it is a capital project.
The two companies describe a multi-million joint investment in equipment, installation and facility upgrades, establishing “a purpose-built, certified manufacturing environment tailored to AMSilk’s proprietary biotechnological process”. The currency is not stated in the release. Given a German company, a French site and a wire distributed in English, euros are the obvious assumption — but it is an assumption, and we flag it below rather than printing a euro sign.
The practical effect is that this capacity has left the merchant market. Whatever the site’s total is, this portion of it is no longer available to a food protein company at any price, for the duration of a long-term agreement whose length is not disclosed.
The counter-argument
There is a reasonable case that this is good news for food fermentation, not bad.
Contract fermentation sites are underused assets with brutal fixed costs. A long-term anchor tenant paying materials-industry prices is exactly what makes a site financially able to keep serving lower-margin food customers at all — the same logic by which a pharmaceutical estate can carry a food protein tenant, which is what we found at EVERY’s arrangement inside Huvepharma’s plant. The AMSilk agreement explicitly contemplates adding capacity. If it does, the site ends up larger, better utilised and more likely to survive than it would have been.
It is also true that AFE has not exited food. Its own description of the Nesle plant says “We have served the food industry, but also some niche applications in cleaning agents, cosmetics or pharmaceutical,” and the AFE boilerplate lists food first among its applications. Standing Ovation’s arrangement is live and was reaffirmed publicly as recently as the coverage of its scale-up plans.
Both points are fair. Neither changes the allocation question, which is the only one that matters to someone trying to book a campaign in 2027.
What we could not establish
- How many 160 m³ cultures exist at Nesle. AFE publishes the size and not the count. We asked the question of the public record and it does not answer.
- What fraction of the site is now committed. Neither AFE nor AMSilk states this, and Standing Ovation’s share is not quantified either.
- The value and currency of the “multi-million joint investment.” The release gives neither a figure nor a currency. We have assumed nothing.
- The duration of the AMSilk agreement. It is described as “long-term” without a term.
- Any price per kilogram for AMSilk’s silk protein, or for Standing Ovation’s casein. Neither company publishes one, so the price-per-kilogram argument above is structural rather than arithmetic, and we have labelled it that way.
- Whether the 160 m³ named in the AMSilk release is the same asset as the 160 m³ advertised on AFE’s website, or a separate vessel of the same specification. This is the central ambiguity of the story and neither party resolves it.
- Current output volumes. Both companies state targets in “hundreds of tonnes”. Neither states what is being produced today.
What to watch
- Whether AFE publishes a vessel count. A contract manufacturer with genuine spare capacity has a commercial incentive to advertise it. Continued silence on the count, while announcing exclusive dedications, is itself information.
- The “addition of precision fermentation capacity in the coming years” clause. If new tanks are announced at Nesle with a size and a date, the site is expanding and the allocation pressure eases. If the clause stays an option, it does not.
- Standing Ovation’s regulatory timetable against its capacity. Its CEO stated in January 2025 a plan of self-affirmed GRAS in the US that year, FDA in 2026, and EU novel foods authorisation in 2027. An EU authorisation arriving in 2027 into a site whose flagship line is dedicated to a textiles customer is a specific, checkable collision. We looked at the route-to-market constraints on precision-fermented casein separately; capacity is the other half of that problem.
- Whether any European food protein company announces a dedicated line of its own. Nobody has. The dedications so far have gone to materials.
The broader pattern is one we keep running into. Europe’s fermentation capacity problem is usually described as a shortage of tanks. It is at least as much a shortage of tanks that food can afford. When the tank exists, is certified, is in the right country, and has been publicly offered as a service — and it still goes to a silk protein for car interiors — the constraint being expressed is not engineering. It is price per kilogram, and food is at the bottom of that stack.