Three European mycelium producers announced new production capacity in 2026. Pacifico Biolabs will “scale production in Saxony to 200 tons per month.” The Protein Brewery will expand “to more than 2,000 MT.” Millow will run “roughly 430+ tons per year by the end of this year.”
Not one of those announcements says what a tonne contains.
That would be a pedantic complaint if the three products were similar. They are not. Pacifico sells a wet, whole-muscle-structured mycelium; The Protein Brewery sells a dried, milled powder; Millow sells a finished food that is roughly half oats. Their published protein contents are 30%, 50% and 14% respectively. A tonne of one is not a substitute for a tonne of another in any sense that matters to a buyer, a plant engineer, or an investor sizing European supply.
The consequence is not abstract. Divide capital raised by announced tonnes and Millow looks like the second-cheapest capacity in the set. Divide capital raised by tonnes of protein — using each company’s own published protein content — and Millow becomes the most expensive of the three, by a factor of more than three against Pacifico.
Same three announcements. Opposite ranking. The variable is a definition nobody published.
What each company announced
| Pacifico Biolabs | The Protein Brewery | Millow | |
|---|---|---|---|
| Site | Saxony, Germany (one brewery partner site) | Breda, Netherlands (called the “Mijkenbroek facility” by the company, “demo-scale factory in Breda” by AgFunderNews) | Stenkullen, Sweden (a former LEGO production hall) |
| Round | €7m Series A ($8.1m) | €18m Series B extension ($20.5m) | €2m ($2.2m–$2.3m depending on source) |
| Announced | 27 May 2026 | 29 June 2026 | 13 August 2026 |
| Capacity claim | “200 tons per month” | “>2,000 MT” (company); “600 metric tons in 2027… >2,000 tons by 2029” (AgFunderNews) | “roughly 430+ tons per year by the end of this year” |
| Target date for that capacity | not stated | 2029 per AgFunderNews and Green Queen; no year in the company’s own release | end of 2026 |
| Product form | Wet whole-cut mycelium, “ready to use by our customers as meat” | Pasteurised, dried, milled powder | Finished food, ~50% fungi / ~50% oats, shipped fresh or frozen |
| Published protein content | 30% | 50% | 14%, “with the ability to increase up to 20%” (CEO to AgFunderNews); 27g per 100g (Green Queen) |
| Basis of the tonnage stated? | No | No | No |
| Total funding to date | €10m equity plus €1.5m grants (one source) | “more than €70 million” (company) | not stated in sources read |
A few of those cells deserve unpacking.
Pacifico’s figure is well attested but undated. It is not a single-outlet misreport: it appears in AgFunderNews, Green Queen, EU-Startups, Cultivated X, Protein Production Technology International, Tech Funding News and MycoStories, and cofounder Zac Austin is quoted directly on it — “The 200t/month capacity ‘will be unlocked at one site, however we’re looking at other partnerships as well.’” What no source carries is a date by which 200 t/month is reached. Every outlet attaches “later this year” or “by late 2026” to the retail launch, not to the capacity. One outlet covering the same round on the same day — Tech.eu — reports no tonnage figure at all.
The Protein Brewery’s 2029 date is not in its own announcement. The company’s press release says it will “expand production capacity to more than 2,000 MT” with no year attached. The “by 2029” and the “600 metric tons in 2027” both come from AgFunderNews and Green Queen reporting CEO statements. If you are building a supply forecast, note which figure carries a company signature and which does not.
Millow’s tonnes are nameplate, and the company says so. CEO Fredrik Öhrn told AgFunderNews: “We’ll operate three S-units with a combined capacity of roughly 430+ tons per year by the end of this year,” and separately, “We expect to reach around 90% utilization during late 2027 rather than running flat out from day one.” That is unusually candid and should be read as a point in Millow’s favour, not against it. Almost nobody publishes the gap between nameplate and expected output.
The arithmetic, and why it should not be taken at face value
Everything in this section is our calculation, not a figure any company published. We show it precisely so the incomparability is legible.
Annualising Pacifico’s monthly figure requires assuming twelve months at rate, which no source supports; we do it only to put the three on one axis.
| Capital raised in the round | Announced annual tonnes | Capital per announced tonne | Protein content | Implied annual tonnes of protein | Capital per tonne of protein | |
|---|---|---|---|---|---|---|
| Pacifico Biolabs | €7m | 2,400 (200 × 12, our annualisation) | ~€2,900 | 30% | ~720 | ~€9,700 |
| The Protein Brewery | €18m | 2,000 (by 2029) | ~€9,000 | 50% | ~1,000 | ~€18,000 |
| Millow | €2m | 430 | ~€4,700 | 14% | ~60 | ~€33,300 |
On announced tonnes the order is Pacifico, Millow, The Protein Brewery — cheapest to dearest. On protein tonnes it is Pacifico, The Protein Brewery, Millow. Millow moves from second to third and its ratio against The Protein Brewery inverts.
And even the second table is not a fair comparison, because it silently assumes the three tonnes have the same moisture content, which they cannot: one is a dried powder and one is shipped fresh or frozen. A dry-matter comparison would move the numbers again. We do not have the moisture data to build it. Nobody publishes it.
There is one more distortion worth naming. Using the round size flatters companies that have raised before. The Protein Brewery’s €18 million is an extension on top of more than €70 million of prior funding, and it is expanding an existing facility that already produced the material EFSA assessed. Against total funding, its capital per announced tonne is above €35,000 — an order of magnitude worse than the round-only figure, and a fairer reflection of what the capacity actually cost to reach.
Practical consequence. Any table of European fermentation capacity — including ones we have built — is summing quantities that are not commensurable. When we mapped who has tanks, we counted vessel volume, which at least has a defined unit. Announced product tonnage does not, and should not be aggregated across companies without asking each one for the basis.
Why the basis goes unstated
The uncharitable reading is that a bigger number is better and wet weight is bigger. That is probably true at the margin, but it is not the main reason.
The main reason is that these companies are not selling into a market with a settled unit. A commodity protein trade quotes on a specified basis because buyers demanded one; whey protein concentrate is sold as WPC 80 precisely because the number in the name is the specification. Mycelium products are sold business-to-business, product by product, on bespoke specifications that are agreed privately. The public announcement is aimed at investors and hiring, not at buyers, and investors have not asked.
This is the same failure we found when four different cost-per-kilogram figures for the same cultivated duck turned out to share no basis, and when a fermented protein “cost parity” claim collapsed on a modelled cost being compared to a market price, and when the same fungal biomass was 50.8% protein and 40.2% protein depending on the assay convention. It is the defining measurement problem of this sector, and it recurs because no regulator or exchange forces a convention.
Practical consequence. If you are evaluating an offtake, the first question is not price per tonne. It is: a tonne of what, at what moisture, at what protein, measured how. Ask before the term sheet, because the answer moves the effective price by multiples.
What the announcements do not measure at all
Announced capacity is not output, and the gap is where this sector actually lives.
In the same AgFunderNews piece that carries Pacifico’s 200 t/month figure, ENOUGH’s former CEO and non-executive director Jim Laird is quoted saying the firm’s Netherlands plant is producing high-quality product “at meaningful levels but not yet at its design capacity.” That is a plant that exists, running below a number that was published years ago. The same article notes Meati Foods’ property was seized for non-payment of taxes.
Millow’s disclosure of a 90% utilisation target in late 2027 against 430 t of nameplate by end 2026 is the only one of the three announcements that acknowledges the gap exists.
There is also a regulatory gate that the capacity numbers ignore entirely, and The Protein Brewery is the case in point. Its Fermotein cleared the EU — the first mycelium ingredient authorised there, under the designation Rhizomucor pusillus mycelium, six years after filing — while its US route went backwards. AgFunderNews reports the FDA advised the company to withdraw its GRAS notice in February 2026 after a letter listing deficiencies including “significant design flaws in animal studies that compromise data interpretation” and “absence of organism-specific secondary metabolite characterization.” The company self-affirmed GRAS in March 2024, filed in March 2025, withdrew in February 2026, and expects a no-further-questions letter in Q1 2027.
Capacity announced for 2029 assumes markets that are open in 2029. Two of the three markets The Protein Brewery names for its 600 tonnes in 2027 — Europe, the US and Singapore — are currently open to it; one is not.
The case that this does not matter
Worth stating properly. Three arguments run against the complaint.
First, buyers already know. No industrial purchaser signs a supply agreement without a specification sheet. The basis is agreed in the contract, at the point where money changes hands. A press release is not a trading document and was never meant to be one.
Second, capital per tonne is the wrong metric anyway. None of these rounds is a capacity capex line. Pacifico’s €7 million funds equipment, team expansion and commercial launch; Millow’s €2 million funds a commercial launch. Treating a Series A as a plant budget overstates precision in a way that is arguably worse than the basis problem it is meant to expose. We agree — which is why we show the calculation as an illustration of instability rather than as a result.
Third, the asset-light model is the actual claim, and it is a different claim. Pacifico’s argument is not that it builds cheap tonnes; it is that it does not build at all. Austin says the capex saving versus other approaches is “more than 95%” — a figure he has given three slightly different ways across three interviews. If the model works, the relevant number is not capital per tonne but how many idle brewery tanks exist and what it costs to convert one. Neither is public.
What we could not establish
- The basis of any of the three tonnage figures. Nine sources on Pacifico, including two direct CEO interviews, none stating wet, dry or protein basis. Same for the other two.
- When Pacifico expects to reach 200 tonnes per month, and whether that figure is nameplate, contracted, or a target. The three published wordings — “scale production to”, “expand production capacity to”, and “will be unlocked at” — are not equivalent.
- Pacifico’s own figures from a company source. The company website returned no readable content; everything here comes from trade coverage.
- Pacifico’s fungal species and strain, explicitly withheld: “Austin did not share what fungi type and strain Pacifico Biolabs is using.” He said only, “We’re growing a strain that we can sell in Europe today.”
- The identity of the brewery partner, unnamed in every source, including those quoting the company.
- Millow’s protein content. 14% per the CEO quoted in AgFunderNews; 27g per 100g per Green Queen. Both published the same day, both apparently from the company. We used 14% because it is attributed to a named executive as a direct quote, and flag that the choice changes our protein-tonne figure by roughly half.
- Whether “Mijkenbroek” and “Breda” name the same Protein Brewery facility. Both are used; no source we read states the relationship.
- Moisture content for any of the three products. Not published anywhere we looked.
- Whether The Protein Brewery’s “>2,000 MT” and the “by 2029” date originate in the same company statement. The date appears only in journalist-written text.
- Several smaller conflicts we did not resolve: Pacifico’s headquarters (Leipzig or Berlin, depending on outlet), its grant funding (€680,000 from Saxony’s development bank per one source, €1.5 million in grants per another), Fermotein’s fibre content (30%, 35%, or 30–35% depending on source), and Millow’s feedstock yield (2.4 or 2.5 kg output per kg of feedstock).
What to watch
- Whether anyone publishes a basis. The cheapest possible fix is one line in a press release: tonnes of dried ingredient at X% protein. The first company to do it will make every competitor’s number look evasive, which is exactly why the first one is hard.
- Pacifico’s first disclosed output figure. The gap between 200 tonnes a month announced in May 2026 and whatever the company reports having actually shipped will be the most informative number this cohort produces. ENOUGH’s “meaningful levels but not yet at its design capacity” is the template for how that disclosure usually reads.
- The Protein Brewery’s US GRAS resubmission. A no-further-questions letter in Q1 2027, as the company expects, would validate a three-market plan built on 600 tonnes. A second round of deficiencies would strand a third of the addressable demand behind the announced capacity.
- Whether a European capacity register appears. A mandated fermentation capacity assessment is one of the asks now in front of the Commission for a second Biotech Act. If it happens and it specifies a basis, this article stops being necessary.