Last week we set out three European mycelium capacity announcements and found that none of them said which basis its tonnes used — wet biomass, dried powder, or finished food. Ranking them by capital per announced tonne gave one answer; ranking by tonne of protein gave the opposite.
Two producers do publish the number that makes the conversion possible. This piece does the conversion.
Angel Yeast states 11,000 tonnes a year of yeast protein at a protein content exceeding 80%. The Protein Brewery states 600 tonnes of Fermotein in 2027 at about 50% complete protein. On announced product tonnes, Angel Yeast’s line is roughly 18 times the size. On protein, using each company’s own stated content, it is roughly 29 times.
That is the whole finding, and it runs the wrong way for anyone who assumed protein-basis conversion would flatter the smaller producer.
The two disclosures
| Angel Yeast — AngeoPro | The Protein Brewery — Fermotein | |
|---|---|---|
| Site | Baiyang Biotechnology Park, Yichang, Hubei, China | Demo-scale factory, Breda, Netherlands |
| Announced capacity | 11,000 tonnes per year | 600 tonnes in 2027; over 2,000 tonnes by 2029 |
| Stated protein content | Exceeding 80% | About 50% complete protein |
| Other stated composition | 5g fibre per 100g; all nine essential amino acids; PDCAAS 1.0 | About 35% fibre; vitamins and minerals |
| Form | Powder, from brewer’s yeast | Powder, whole mycelium biomass |
| Process | Fermentation, autolysis, separation, drying — fully automated | Non-sterile fermentation |
| Organism | Yeast | Rhizomucor pusillus mycelium |
| Implied protein tonnage | Over 8,800 tonnes per year | About 300 tonnes in 2027; about 1,000 by 2029 |
The implied protein tonnage row is our arithmetic, not either company’s claim. It is simply announced tonnes multiplied by stated protein content. For Angel Yeast we have used 80% as a floor, because the disclosure says “exceeding”; the true figure is higher and the gap therefore wider.
Angel Yeast’s own release describes the line as having “built-in expansion capacity”. The Protein Brewery’s 2029 target of over 2,000 tonnes converts to roughly 1,000 tonnes of protein — still less than an eighth of what the Yichang line produces today at its stated floor.
The practical consequence: if you are sourcing protein rather than powder, ask for protein content in the same breath as tonnage, and do the multiplication before comparing suppliers. Two disclosures that look 18 times apart are 29 times apart on the thing you are actually buying.
Why 80% against 50% is not a quality judgement
It would be easy to read the protein-content gap as one product being better. It is not, and a formulator will already know why.
AngeoPro is a protein isolate route: brewer’s yeast taken through autolysis, separation and drying, with the non-protein fraction largely removed. Fermotein is a whole-biomass route: the mycelium is dried more or less intact, which is why about 35% of it is fibre. The Protein Brewery’s CEO Thijs Bosch told AgFunderNews the powder “comprises about 50% complete protein, 35% healthy fibers and is full of vitamins and minerals so is mainly used as a nutritional ingredient” — the fibre is the product, not a yield loss.
Those are different ingredients sold into different formulation problems. What they have in common is that both companies published enough to let a buyer normalise them, which is more than the rest of the field does. Our earlier note on crude protein against true protein in microbial biomass sets out why even a stated percentage needs its method attached; neither of these disclosures says whether the figure is crude protein by nitrogen conversion or true protein, and that is the next question to ask both.
The practical consequence: a high protein percentage is a processing choice with a cost attached. Isolating to 80% means discarding or downstream-selling the rest, which shows up in price per kilogram of finished ingredient. Do not treat protein content as a proxy for value; treat it as the unit conversion it is.
The date on the Angel Yeast line
The Yichang line is not new capacity. FoodBev Media reported the inauguration on 7 November 2025, and Green Queen on 5 November 2025. Nutrition Insight also carried the opening. Angel Yeast’s own release image assets are filed under a November 2025 path.
That matters for two reasons. First, the line has had roughly nine months of operating history that nobody outside China appears to have reported on — no utilisation figure, no output figure, no customer. Announced capacity and produced tonnes are different quantities, and the gap between them is where most fermentation-capacity claims fail. Second, anyone building a capacity map from the trade-press feed will date this line to 2026 if they are not careful.
Our fermentation capacity map is maintained rather than reissued for exactly this reason. Angel Yeast has been added at its stated figure with the November 2025 date attached.
What Angel Yeast changes about the comparison set
The venture-backed Western fermentation-protein companies are the ones that get covered, and they are not the ones with the tanks.
Green Queen reports Angel Yeast already accounts for around 15% of global yeast production and nearly 60% of the Chinese market. FoodBev puts its facility count at 33 worldwide. The company describes itself as a listed high-tech company with 12 holding subsidiaries, supplying more than 170 countries and regions. AngeoPro won FoodBev Media’s Best Ingredient Innovation award in the 2025 World Food Innovation Awards.
This is an incumbent fermentation business adding a protein line to an existing industrial base, not a startup building first-of-a-kind capacity. The cost structures are not comparable, and neither are the risks. It is the same structural point we made about China’s first mycoprotein approval: the constraint that binds a Western scale-up — access to tanks at a price that works — is not the binding constraint for a company that already owns tanks.
The practical consequence: if your capacity map only contains companies that have raised venture rounds, it is measuring the wrong population. The tonnage exists; it is mostly owned by incumbents who do not issue funding announcements.
The counter-argument
Three objections deserve stating properly.
Announced capacity is not shipped tonnes. Correct, and this piece does not claim otherwise. Angel Yeast has published a nameplate figure. Neither company has published a utilisation rate. The one supply-side fact The Protein Brewery has disclosed — that its 2026 capacity is sold out and 2027 is booking out — is a demand signal Angel Yeast has not matched with anything public.
Regulatory scope, not capacity, is the real constraint. Also fair. Fermotein cleared the EU in June 2026 after a six-year dossier, holds Singapore approval and self-affirmed GRAS in the US, and has dossiers moving in the UK, Canada and Australia/New Zealand — a route-to-market position we examined when its US notice closed without a decision. We could not establish AngeoPro’s regulatory status in the EU, US or UK from any source we fetched. Eight thousand eight hundred tonnes of protein you cannot sell into a given market is not capacity in that market.
The products are not substitutes. True at the formulation level. But a nutritional-powder buyer choosing between a yeast isolate and a mycelium whole-food powder is a real buyer, and Fermotein’s own stated lead application — ready-to-mix active nutrition — is the segment Angel Yeast names for AngeoPro too.
What we could not establish
- AngeoPro’s regulatory status outside China. Nothing we fetched states whether the ingredient is approved, notified or self-affirmed in the EU, US, UK or Singapore. Capacity without a route to market in a given jurisdiction is not supply to that jurisdiction.
- Whether either protein figure is crude or true protein. Neither disclosure states the method. A nitrogen-conversion crude-protein figure and a true-protein figure are not interchangeable, and the difference is larger for microbial biomass than for most feedstocks.
- Actual output from the Yichang line. Only nameplate capacity is public. No utilisation, production or shipment figure was found.
- The exact commissioning date. Two trade reports place the inauguration in the first week of November 2025. We did not find a company statement giving a commissioning date, and we have not asserted one.
- Price for either ingredient. No per-kilogram figure is public for AngeoPro or Fermotein. Bosch’s remark that non-sterile fermentation “structurally lowers capital and operating costs” is a directional claim with no number attached.
- The yeast protein market size. Green Queen cites a third-party forecast of $1.5bn today rising to $2.3bn by 2030. We have not used that figure in this analysis because we could not read the vendor’s methodology, and market-forecast totals are not comparable to disclosed plant capacity.
- Detail from the Nutrition Insight report. The page fetched successfully but exceeded what we could read in full. We have cited it only for the fact its own summary states — that the line has begun operating at Baiyang — and drawn no figures from it.
What to watch
- Whether Angel Yeast publishes an output figure. A listed company adding a flagship protein line usually reports something. Nine months in, nameplate is all there is.
- Whether AngeoPro appears in a Western regulatory register. An EU novel food application, a GRAS notice, or an FSA dossier would convert Chinese capacity into addressable supply. Nothing we found says one exists.
- Whether The Protein Brewery hits 600 tonnes in 2027. The 2026 book is reported sold out. The 2027 figure is a commitment against a demo-scale plant, and demo-scale plants miss.
- Whether other producers start publishing protein content. Two out of the field is a low bar. If stating the basis becomes normal, capacity comparisons stop being guesswork.