China’s National Health Commission approved Fusarium venenatum protein as a new food raw material on 27 November 2025, under Announcement No. 7 of 2025. The approval names two permitted production strains, sets a minimum protein specification of 50 g/100 g on a dry basis, and excludes four population groups — but it sets no daily consumption limit, which makes it the outlier among the three new food raw materials cleared in the same announcement.
That last point is the one the English-language coverage missed, and it is the one that matters commercially. It is also worth stating plainly at the top that the year is 2025, not 2026. Coverage carrying 2026 datelines is republication; the primary record is unambiguous.
The regulatory facts, pinned down
The NHC’s own interpretation document for Announcement No. 7 of 2025 is the authoritative text. It describes the substance — 威尼斯镰刀菌蛋白, Mycoprotein from Fusarium venenatum — as produced from Fusarium venenatum strain A3/5 or strain TB01, via fermentation, nucleic acid removal, inactivation and filtration, with principal nutritional components of protein (≥50.0 g/100 g, dry basis), fat and dietary fibre.
The full announcement as mirrored by Foodmate records the document as signed by the NHC on 19 November 2025 and published and effective on 27 November 2025. It covers 14 items in total: three new food raw materials, eight new food additive varieties and three new food-related product varieties.
The compliance timeline, disclosed by CIRS Group, the consultancy that filed the dossier, runs entirely within 2025:
| Stage | Date | Reference |
|---|---|---|
| Application accepted by NHC | 7 February 2025 | Acceptance No. 卫食新申字(2025)第0010号 |
| CFSA draft released for public comment | 10 July 2025 | China National Center for Food Safety Risk Assessment |
| NHC approval issued | 27 November 2025 | Announcement No. 7 of 2025 |
Nine months and twenty days from acceptance to approval. CIRS calls this “just nine months” and treats it as evidence of dossier quality; it is also a data point on how fast this pathway moves for a substance with an established international safety record.
Practical consequence: if you are modelling a China entry for a fermentation-derived biomass ingredient with existing EU, US or Canadian clearance, the acceptance-to-approval window here was under ten months — an order of magnitude faster than the six years The Protein Brewery spent in the EU novel food queue.
What the approval actually permits
| Term | What Announcement No. 7 of 2025 specifies |
|---|---|
| Approved strains | Fusarium venenatum A3/5 or TB01 |
| Process | Fermentation, nucleic acid removal, inactivation, filtration |
| Identity spec | Protein ≥50.0 g/100 g, dry basis; also contains fat and dietary fibre |
| Daily intake limit | None stated |
| Permitted food categories | None specified, and therefore not restricted by category |
| Excluded populations | Infants and young children; pregnant women; lactating women; people allergic to edible fungi |
| Safety indicators | To be applied per the announcement’s own annexed limits |
The exclusions are framed on a precautionary basis: the NHC states that safety data in infants and young children, pregnant women and lactating women are insufficient, so on a risk-prevention principle those groups should not consume it. This is a labelling and formulation constraint, not a safety finding against the ingredient.
The missing intake cap is the real finding
Chinese new food raw material approvals routinely carry a recommended maximum daily consumption. The other two raw materials in the very same announcement do:
| Ingredient (Announcement No. 7 of 2025) | Recommended daily intake |
|---|---|
| Elderberry anthocyanins (接骨木莓花色苷) | ≤3.2 g/day, on a basis of 10 g/100 g total anthocyanins |
| Olive fruit polyphenols (油橄榄果多酚) | ≤600 mg/day, on a basis of 10.0 g/100 g total polyphenols |
| Fusarium venenatum protein (威尼斯镰刀菌蛋白) | No limit stated |
This asymmetry is not an oversight. Anthocyanins and polyphenols are concentrated bioactives where a dose ceiling is the operative control. Mycoprotein is a macronutrient ingredient — eaten by the hundred grams, as forty years of Quorn consumption demonstrates — and a milligram cap would be meaningless. The regulator treated it as a food, not a functional extract.
The ≥50 g/100 g protein figure has been described in secondary coverage as a “composition limit”. It is not a limit in any restrictive sense. It is a minimum, an identity specification defining what may be sold under the name. It constrains the seller’s product, not the buyer’s inclusion rate.
Practical consequence: a formulator in China can use this ingredient at whatever inclusion level the recipe requires, in any food category, subject only to the four population exclusions and the annexed safety limits. That is a materially freer permission than most new food raw material approvals grant, and it is the difference between an ingredient that can carry a burger patty and one that can only be dosed into a supplement.
Applicant-specific, or generic?
The announcement and the NHC interpretation both describe the substance generically. Neither names Jiangxi Fushine Biotechnology. The scope is drawn around the substance, the two named production strains and the process — not around a company or a facility. Under NHFPC Order No. 1, the object of review is the raw material itself, and approved materials are published to a national list.
So the approval is generic to the substance. Any company may rely on it — provided it can produce conforming material using strain A3/5 or TB01.
That proviso is where the commercial moat sits. A3/5 is the long-established Fusarium venenatum strain underlying Quorn, per the literature summarised by FoodTimes. TB01’s provenance is not established by any source we fetched. If TB01 is proprietary to Fushine, a generic approval with a two-strain gate is functionally a duopoly of strain access, not an open door.
How the three pathways compare on the questions an operator actually asks:
| Question | China new food raw material | EU novel food | US GRAS |
|---|---|---|---|
| Who may rely on it | Any company producing conforming material from the named strains | The Union list entry applies generally, though data protection can reserve it to the applicant for a period | The conclusion attaches to the substance and its described uses, not to a licence held by one firm |
| Published instrument | Numbered NHC announcement plus annexes and an official interpretation | Commission implementing regulation amending the Union list | For notified GRAS, an FDA inventory entry and releasable notice; for self-affirmed GRAS, nothing is required to be published |
| Strain named in the public text | Yes, A3/5 and TB01 | Varies by dossier | Varies by notice |
| Intake terms published | Yes, including any cap and excluded populations | Yes, with conditions of use | Stated in the notice where one is filed |
| Elapsed time on the case at hand | 9 months, 20 days, from acceptance to approval | Six years for Fermotein, per our earlier reporting | Fushine states its expert panel concluded in January 2025; no filing date is disclosed |
Practical consequence: the Chinese file is the most legible of the three. It publishes the strain, the spec and the population exclusions in one numbered document. For a competitor considering China entry with a Fusarium venenatum product, the regulatory question is now largely answered and the binding constraint moves to strain access and manufacturing.
The industrial layer: 1,200 against 67,000
Approval is half the story. What can Fushine actually ship?
On its own site, Fushine states it “has already completed a dedicated 1,200-ton production line, which is operating steadily, and construction of the Phase I 200,000-ton mycoprotein and resource-utilization project is currently underway.” Green Queen and CIRS repeat these figures. The facility is at Changjiang District, Jingdezhen, Jiangxi.
Note what is absent from every one of those sources: the basis. Not one states whether 1,200 tonnes means wet biomass, dry biomass or protein. Mycoprotein wet paste is roughly three-quarters water and the dry solid around half protein, so “1,200 tonnes” can denote anywhere from about 150 to 1,200 tonnes of actual protein depending on the convention. We have written before about how basis mismatch corrupts alt-protein comparisons; this is the same failure in capacity units rather than cost units.
The scale check:
| Asset | Stated figure | Basis stated? | Status |
|---|---|---|---|
| Fushine, Jingdezhen line | 1,200 tonnes/year | No. No source states it | Company says completed and operating |
| Fushine, Phase I project | 200,000 tonnes/year | No. No source states it, and the figure covers a combined “mycoprotein and resource-utilization” project | Company says under construction |
| Marlow Ingredients, Belasis (UK) | 67,000 tonnes/year | No explicit basis; CEO glosses it as “the protein equivalent of 143,000 cows” | Stated in 2023 as the capacity a fifth fermenter “will take” the site to |
Marlow’s number deserves its own caveat. The 67,000-tonne figure comes from a quote by then-CEO Marco Bertacca, reported in May 2023, describing a fifth 170,000-litre fermenter then being built that would produce 375 tonnes of mycoprotein per week. It is a forward-looking capacity figure from 2023, not a verified current output, and we found no source confirming the fifth fermenter was commissioned or that 67,000 tonnes was reached. Treat it as the best-documented figure for the largest facility in the category, with that qualification attached.
Taken at face value and assuming a common basis, Fushine’s operating line is about 1.8% of Marlow’s stated capacity. Its announced Phase I would be roughly three times the world’s largest existing mycoprotein plant. Both of those numbers should shape how a buyer reads the approval.
Practical consequence: the approval unlocks a domestic market of over a billion people for a line that, on the company’s own disclosure, produces 1,200 tonnes of something. If you are an ingredient buyer, the binding question in 2026 is not whether the ingredient is legal in China. It is whether Fushine can quote you volume, and on what basis the tonnes are counted. Ask for kg of protein at a stated dry-matter percentage, in writing.
Formats and what is not disclosed
Fushine sells three formats, per its product pages: a wet form (2.5 kg packs) for structured products such as ham slices, sausages, meatballs and nuggets; a dry powder (20 kg packs) for beverages, baked goods, meal replacements and supplements; and whole cuts (300 g packs). The company’s January 2025 announcement referenced only two formats, dry and wet, so the cuts line appears to be the newer addition.
No source we fetched discloses a price per kg, a named customer, an offtake agreement or a utilisation rate on the 1,200-tonne line. For a publication that tracks who has agreed to buy the thing, that absence is the headline commercial fact.
Scrutinising the efficiency claims
Two company claims are repeated across coverage. Both come from Fushine, and neither carries a stated basis.
“Biomass doubles every five hours.” Reported by Green Queen as a property of strains A3/5 or TB01. A doubling time is an early-exponential-phase kinetic parameter and says nothing about plant output. Industrial fermentation of Fusarium venenatum is constrained by oxygen transfer, heat removal, harvest rate and — specific to this organism — the RNA reduction step, which is a hard throughput bottleneck and is explicitly named in the NHC process description (“脱核酸”, nucleic acid removal). A five-hour doubling time in a flask does not translate into tonnes at the gate.
“1,000 times more efficient than livestock-derived protein.” Also from the company. No source we fetched states what is being measured: protein output per unit land, time, feed, water, or something else. Without a denominator it is not checkable. For comparison, the sourced claims in this category are an order of magnitude more modest — the Marlow figures reported by FoodTimes are 90% less land and water than beef and 98% fewer carbon emissions, a 10x-to-50x class of claim, not 1,000x.
Practical consequence: attribute both figures to Fushine, or leave them out. Neither belongs in a technoeconomic model.
The US GRAS status: three sources, three different claims
This is worth flagging because it affects any assessment of Fushine’s export readiness, and the sources genuinely disagree.
- CIRS says it “had already supported Fushine in securing U.S. market access for the same ingredient through the GRAS Notice.”
- Green Queen says Fushine “has already self-determined FuNext as a Generally Recognized as Safe (GRAS) ingredient in the US.”
- Fushine itself, in January 2025, says FuNext “has officially passed the scientific evaluation process conducted under the U.S. FDA’s GRAS framework and received approval from an independent expert panel.”
Fushine’s own wording is decisive on the mechanism. “Approval from an independent expert panel” is the signature of a self-affirmed GRAS conclusion. It is not an FDA no-questions letter, and an expert panel is not a regulator. We could not locate a GRN number in the FDA GRAS Notice Inventory attributable to Fushine.
The distinction matters more now, given the proposed move to mandatory GRAS notification. Self-affirmation is a lawful basis for US marketing today. It is also the weakest of the three positions Fushine’s coverage claims for it, and the one the company’s own website describes.
Practical consequence: treat Fushine’s US position as self-affirmed GRAS unless and until a GRN appears in the FDA inventory. If you are a US buyer, ask to see the expert panel report and the dossier, because there is no public FDA file to read.
The counter-argument
The sceptical read of everything above is that a 1,200-tonne line with no disclosed customers and no disclosed price is not a supply story, and that a generic approval mostly benefits whoever eventually builds real capacity — which may not be Fushine.
There is force in this. The company’s Phase I figure of 200,000 tonnes would be roughly three times the largest facility on earth, announced by a firm whose current operating asset is under two percent of that same benchmark. Announced Chinese fermentation capacity has a track record of being quoted at figures that do not survive contact with commissioning, and we have documented how capex claims in Chinese fermentation get repeated without a currency or a basis. The 200,000-tonne figure also bundles “resource utilization” into the same project, so it is not even cleanly a mycoprotein number.
The stronger counter is the regulatory asymmetry. Fushine secured a national permission with no intake cap and no category restriction, in the largest food market in the world, in under ten months. That permission is now available to competitors too — but Fushine is the only firm with both the permission and an operating line, and first-mover access to Chinese manufacturers needing compliant domestic supply is a real asset even at 1,200 tonnes. Whether it converts depends entirely on offtake, which nobody has disclosed.
A further reading: the absence of an intake cap may reflect the strength of the forty-year consumption history rather than regulatory generosity — the NHC interpretation explicitly notes US GRAS status and EU and Canadian permission. On that reading the precedent may not transfer to a novel organism without that history, and other fermentation firms should not assume the same terms.
What we could not establish
- The approval year is resolved. It is 2025. The NHC interpretation is dated 2025-11-27, the announcement is numbered 2025年第7号, the internal signature date is 19 November 2025, and CIRS dates acceptance to 7 February 2025 under acceptance number 卫食新申字(2025)第0010号. Outlets carrying 2026 datelines are republishing 2025 material. We flag this because getting it wrong would invalidate the whole timeline.
- The basis of every capacity figure here. Neither Fushine, CIRS, Green Queen nor the Marlow reporting states whether tonnes are wet biomass, dry biomass or protein. Every capacity comparison above is indicative only.
- The annexed food safety indicators. The announcement’s Annex 1 PDF, carrying the numeric contaminant and microbiological limits, did not return content on repeated fetch attempts from two hosts. The terms above come from the NHC’s official interpretation, authoritative on scope but not reproducing the numeric annex.
- TB01’s provenance. No source states whether TB01 is proprietary to Fushine, publicly deposited, or licensed. This determines whether the generic approval is practically open to competitors.
- Any price, customer, offtake or utilisation figure for Fushine. None disclosed anywhere we looked.
- A GRN number for Fushine in the FDA GRAS Notice Inventory. We could not find one, consistent with a self-affirmed conclusion but not proof that no filing exists.
- Whether Marlow’s 67,000 tonnes was reached. The figure was forward-looking in 2023 and we found no confirmation of the fifth fermenter’s commissioning.
What to watch
Four testable claims, each of which will resolve one way or the other:
- A second applicant relies on the Announcement No. 7 permission. If the approval is genuinely generic, another firm should be able to launch a Fusarium venenatum product in China without its own dossier. Watch the NHC’s periodic acceptances: if competitors keep filing their own applications rather than relying on the existing entry, the strain gate is binding.
- Fushine discloses a basis for its tonnage. The most informative disclosure available is kg of protein per year at a stated dry-matter percentage. Until then, treat 1,200 and 200,000 as unqualified.
- A named Chinese offtake. A branded product on shelf listing 威尼斯镰刀菌蛋白 in its ingredient declaration, from a manufacturer other than Fushine, converts this from a regulatory story into a supply story.
- Movement on Phase I. If the 200,000-tonne project hits disclosed commissioning milestones, the scale check changes materially. If it stays a press-release number through 2027, discount it to zero. Note that co-locating mycoprotein capacity with sugar and starch infrastructure is how the rest of the industry is closing the capex gap, and Fushine has disclosed no such arrangement.
The regulatory question in China is settled and settled generously. The manufacturing question is entirely open.