N. Regional comparators and approvals

This appendix holds the regional detail behind Region. It compares Vietnam with nine other jurisdictions on novel-food rules, approvals, public money, open pilot plants and corporate activity. It sets out the trade flows Vietnam depends on, and it reads the GFI APAC x Hawkwood Biotech "Where to Build" study in full, with our critique. It ends with lessons from how Singapore, Thailand and Korea built their frameworks.

#N.1 Summary

• Vietnam has no novel-food framework and no approvals. The GFI APAC x Hawkwood study scores its regulatory readiness at 0 out of 100 ¹ VN-directCao.
• Singapore has approved 18 novel foods, including the world's first cultivated duck and beef ² VN-adjacentCao. It has also stepped back from producing alternative protein itself ³ VN-adjacentCao.
• The US FDA completed its review of Wildtype's cultivated coho salmon on 28 May 2025, the first cultivated seafood cleared in the US ⁴ generalCao.
• Outside Singapore, no Asian regulator had approved cultivated meat by September 2026, even where a route exists ^5,6,7 VN-adjacentTrung bình.
• Vietnam's cost advantage is real but acts on a small share of cost, and Indonesia matches or beats it on every cost factor in the study ¹ VN-directCao.
• Vietnam ships its fermentable carbohydrate to China as starch, not as fermentation-ready glucose ^8,9 VN-directCao.
• Thailand is responding directly to the same ranking with a 10-year biotechnology framework that names alternative protein ¹⁰ VN-adjacentTrung bình.

#N.2 Novel-food frameworks, country by country

From countries.csv. "PF" is precision fermentation. The Hawkwood regulatory score is 100 for novel-food guidance plus an approved fermentation-derived ingredient, 50 for a framework without such an approval, and 0 for no framework or guidance ¹ generalCao.

For Vietnam's own rules in detail, see G. Regulation.

#N.3 Approvals to date

#N.3.1 Singapore: 18 approvals on the list dated 14 August 2026

Dates are the SFA "date allowed to be used as food" ² VN-adjacentCao.

By category: 6 cultivated, 6 fungal or bacterial biomass, 3 precision-fermented, 2 microalgae and 1 synthetic compound ² VN-adjacentCao.

Notes:

• Aleph Farms' beef is dated 22 July 2026 on the SFA list and was announced on 4 August 2026. We treat the SFA date as the approval date (DG-071) ^2,23 VN-adjacentCao. The product is 10 to 20% cultivated cells in a plant-protein base, with a foodservice launch targeted for the first half of 2027, produced with Cell AgriTech ^24,25 VN-adjacentTrung bình.
• Some press describes Solein as precision fermentation. It is gas-fermented bacterial biomass (DG-074) ^2,26 VN-adjacentCao.
• No Wolffia (Mankai), duckweed or insect entry is on the list. Insects are handled under a separate SFA framework. Some GM-microbe ingredients may sit on SFA's separate GM food list, which we did not check ² VN-adjacentCao.

#N.3.2 Other jurisdictions

We found no novel-protein approval in Thailand, Korea, Japan, Malaysia, Indonesia, the Philippines or Vietnam as of September 2026 ^15,5,7 VN-adjacentTrung bình. Singapore's own Islamic authority has also set out a halal path for cultivated meat ²⁸ VN-adjacentTrung bình.

Reading. Having a framework does not by itself make approvals fast. Korea opened applications in February 2024 and Thailand received its first dossier in December 2024; neither had approved cultivated meat by September 2026 ^5,6,19 VN-adjacentTrung bình. Singapore's first cultivated approval is dated 26 November 2020 on the SFA list, the year after its novel-food framework was set up in 2019 ^2,11 VN-adjacentCao.

#N.4 National strategies and public funding

Amounts are as found and in different currencies and years; they are not comparable and must not be added (DG-079).

The "Greater Food Approach" (drawing "calories and protein from plant, animal, and microbial sources") appears in the GFI China report as a paraphrase of a 2022 speech; we could not confirm it from a primary source ¹⁶ generalThấp.

#N.5 Pilot and contract manufacturing facilities open to third parties

From facilities_regional.csv (16 rows). "Food grade" means a food manufacturing licence or equivalent was found.

Further points:

• ScaleUp Bio's 2024 clients included Nourish Ingredients (Australia), C16 Biosciences (US), Ultimeat (Malaysia), Allium Bio, Allozymes and Algrow Biosciences (Singapore) ^36,37 VN-adjacentTrung bình. Whether Nurasa (Temasek), its original joint-venture partner, has exited is not confirmed ³⁴ VN-adjacentTrung bình.
• Typical Chinese contract terms: the client keeps its strain and operating procedures, while the manufacturer may own process improvements; clients are expected to visit monthly or keep staff on site. One client said scale-up elsewhere would have taken "two times as long and double the cost" ¹⁶ generalTrung bình.
• No Vietnamese facility publicly offers food-grade pilot fermentation to outside firms (see L. Labs and talent) ^45,46 VN-directTrung bình.

For a Vietnamese startup today the realistic contract manufacturing options are ScaleUp Bio (food-licensed, 10,000 L), EECi in Thailand (larger tanks, food status to confirm) and Chinese pharma contract manufacturers (largest and cheapest, with the most IP exposure). Singapore's shift away from local food production makes ScaleUp Bio's widened, non-food focus a risk to watch. This is our assessment.

#N.6 Regional corporates and their Vietnam links

Only items with a source read in this study are shown.

Not verified in this study: C.P. Vietnam's alternative-protein links, Wilmar, ofi, Japfa, Sojitz, Fuji Oil, and the capacities of Chinese methanol, ethanol and gas-based SCP makers. Chinese gas-fermentation protein (Clostridium autoethanogenum protein) from steel-mill off-gas is in commercial feed trials ⁵⁶ generalTrung bình.

#N.7 Trade flows Vietnam depends on

#N.7.1 Imports of food-protein ingredients are small

From trade_flows.csv. Vietnam-reported tonnages are WITS estimates (every partner shows the same unit value), so we quote values and use China's mirror data for tonnage ^57,58 VN-directTrung bình.

• Combined, Vietnam-reported imports under HS 3504 and 2106.10 were about USD 56 million in 2023 (our sum) ^57,58 VN-directTrung bình. Vietnam also imported USD 31 million of wheat gluten in 2023, 59% from China ⁶³ VN-directThấp.
• China-reported unit values of USD 1.0 to 1.6/kg for HS 3504 are far below soy protein isolate prices, so much of this trade is probably feed-grade protein substances. The two data sets also disagree on value (USD 10.1 million vs USD 16.3 million from China for HS 3504 in 2023) for reasons we could not establish (DG-075) ^60,59 VN-directTrung bình.
• These food-ingredient imports are tiny next to feed protein imports: in 2025 Vietnam imported 2.61 Mt of soybeans (98.4% of supply) and 5.70 Mt of soybean meal, and about 99% of the soy protein in Vietnamese feed is imported ^64,65 VN-directCao.

#N.7.2 Exports that already feed China's bio-industries are large

• Vietnam's cassava starch sells at about USD 0.49/kg (2023, Vietnam-reported unit value) ⁶⁶ VN-directCao.
• The glucose gap. China imported USD 114 million of modified starch and USD 104 million of dried cassava from Vietnam in 2024, but only USD 0.16 million of glucose. The step that turns starch into glucose for fermentation happens in China. Keeping that step in Vietnam is the practical way for Vietnam to plug into regional biomanufacturing ^67,68,9 VN-directCao. The interpretation is ours.
• The 2023 Vietnam-reported and 2024 China-reported starch figures are different years and reporters; we do not average them.
• Vietnam is also a net exporter of fishmeal, mostly to China: USD 430 million out (93.5% to China) against USD 260 million in (49% from India) in 2023 ^69,70 VN-directTrung bình.

#N.8 The GFI APAC x Hawkwood "Where to Build" study (October 2025)

#N.8.1 What the study is

• Authors and title. Jennifer Morton and Dean Powell (GFI APAC) and Tony Day (Hawkwood Biotech), "Where to Build: site selection and competitiveness in APAC fermentation manufacturing", October 2025. A companion GFI article names Thailand, Vietnam and Australia as the "three APAC countries with the right ingredients" ^1,71 generalCao.
• Countries. Australia, Indonesia, Japan, Malaysia, the Philippines, Singapore, South Korea, Thailand and Vietnam ¹ generalCao.
• Drivers scored. Feedstock (sugar), capex, utilities, labour, talent, business environment and regulatory readiness, plus incentives. The text says incentives and feedstock cost "emerged as the two most critical drivers" ¹ generalCao.
• Assumptions ¹ generalCao:
  • Representative plants of 80,000 t/yr (submerged biomass) and 10,000 t/yr (precision fermentation), modelled at a US Midwest site on corn glucose "due to limitations in data and modelling capacity".
  • Sugar index: 2024 exports (double-weighted), production and import reliance.
  • Talent: engineering, manufacturing and construction graduates, vocational enrolment, and R&D researchers per million people.
  • Business environment: the World Bank Ease of Doing Business index (2021 data vintage per the table header), which was discontinued in 2021.
  • Regulatory readiness: as in section N.2.
• Scores are min-max within the nine countries, from 0 to 100, with 100 best ¹ generalCao.

#N.8.2 Vietnam's inputs and scores

Source for the whole table: ¹ VN-directCao. The study also claims Vietnam had "the region's highest sugar yield per hectare" in 2024; we did not verify this (DG-082).

#N.8.3 All nine countries

Order: sugar, capex, utilities, labour, talent, business environment, regulation. The cost rows (study Figure 18) and enabling rows (Figure 21) were extracted verbatim in separate reads; an earlier machine read gave different sugar numbers for Indonesia and Australia, and we use the verbatim read (DG-081) ¹ generalCao.

Reading across: no country combines low cost with regulatory readiness. Thailand comes closest. Indonesia scores the same as or better than Vietnam on every cost factor ¹ generalCao. We checked the business-environment index by arithmetic against World Bank Doing Business 2020 scores, and it matches.

#N.8.4 Cost structure and incentives in the study

• In the study's own representative plant models, labour is 7% (biomass, 80,000 t/yr) and 17% (precision fermentation, 10,000 t/yr) of production cost, and utilities 4 to 5%. Depreciation (capital cost) is 37 to 56% ¹ generalCao.
• The same report prints feedstock at 10% (biomass) and 18% (precision fermentation) in its representative model, and about 35 to 65% (biomass) and 30 to 45% (precision fermentation) across other TEAs. Both are in the report; for Vietnam the point stands either way: labour and utilities are a minor share (DG-076) ¹ generalCao.
• Incentives. The highest return to government and the highest gain in company net present value both come from a 50% capex grant or non-dilutive funding (a reported NPV gain of USD 73 million). A 10% feedstock cost cut and a loan guarantee come next. Tax holidays rank near the bottom ¹ generalCao.

#N.8.5 What the study says about Vietnam

• Vietnam "presents one of the region's most cost-competitive environments" and "a potential first-mover opportunity" ¹ VN-directCao.
• Obstacles: "constraints in feedstock supply, the absence of a clear regulatory pathway for novel foods, and limited experience with industrial-scale fermentation" ¹ VN-directCao.
• Its three recommendations for Vietnam: "Operationalise Resolution 36 via funded programs with specific agency mandates"; "Clarify incentive eligibility criteria, timelines, and application processes for biomanufacturing sites"; "Publish regulatory procedures and statutory timelines for novel food approvals" ¹ VN-directCao.
• It cites an unnamed "significant investment in a large-scale biomanufacturing facility" in Vietnam, with no reference ¹ VN-directThấp.
• A GFI APAC article summarises the study as citing Vietnam's incentives and bioeconomy strategy but also its feedstock limits and regulatory immaturity ⁷² generalTrung bình.

#N.8.6 Our adversarial critique

1. It rewards Vietnam for what matters least. The index gives Vietnam top scores for wages and power, but the study's own cost model says labour and utilities together make up less than a quarter of cost; capital is 37 to 56% ¹ generalCao. Cheap labour and power alone will not win projects.
2. It measures the wrong carbohydrate. The feedstock index counts sugar only and scores Vietnam 26, ignoring cassava starch, Vietnam's real fermentable carbohydrate. Vietnam exported 2.15 Mt of cassava starch in 2023 ^1,66 VN-directCao. But the correction cuts both ways: that starch is already sold to China, and China imported only USD 0.16 million of glucose syrup from Vietnam in 2024 (section N.7.2).
3. It scores the business environment with a discontinued index. The World Bank stopped the Ease of Doing Business series in 2021 ¹ generalCao.
4. It gives no country-specific cost per kg. All plant costs are modelled at a US Midwest site on corn glucose ¹ generalCao.
5. The scores are relative ranks among nine countries. A score of 100 means "best of nine", not "good". They are not investment-grade cost estimates.
6. Its regulation score for Malaysia conflicts with GFI's own State of Play (section N.2).
7. Two claims about Vietnam are unsourced in the study: the unnamed large biomanufacturing investment and the "highest sugar yield per hectare".

What survives the critique: Vietnam is a low-cost site without a regulatory route, and the study's three recommendations (fund Resolution 36 with agency mandates, publish incentive criteria, publish a novel-food procedure with statutory timelines) match our own findings in G. Regulation.

#N.9 Lessons from Singapore, Thailand and Korea

Sources for the table: ^{2,15,6,19,13,20} VN-adjacentTrung bình.

What Vietnam could transfer (our assessment) ^15,5,6,13 VN-adjacentTrung bình:

1. Korea shows a route that does not need a new law. Add a "new food ingredient" safety assessment step inside an existing instrument (for Vietnam, the decree that will replace Decree 15/2018), with a published fee and a statutory review time. Korea took about nine months (May 2023 to February 2024).
2. Thailand shows how to borrow assessors. Designate an existing technical body as the safety assessment unit instead of building a new agency. The National Institute for Food Control is one candidate (see G. Regulation and policy option PO-019).
3. Singapore shows the value of transparency. Publishing each approval with its specification and risk controls lets later regulators rely on earlier assessments. Vietnam could accept or fast-track dossiers already approved by SFA, FSANZ or the US FDA.
4. A framework is not enough. Thailand and Korea have routes but no cultivated approvals after 21 to 31 months. Staff capacity and review timelines matter as much as the legal text.
5. Production and regulation can be separated. Singapore now acts as regulator, research hub and financier more than as a production base, which leaves room for lower-cost neighbours to host manufacturing ^3,29 VN-adjacentCao. Aleph Farms says it will use capacity in Penang to supply the region ²⁴ VN-adjacentTrung bình.

#Gaps and how to close them

#Related data files

• countries.csv: 11 country rows behind section N.2 and N.4.
• approvals.csv: the 33 approval rows in section N.3.
• facilities_regional.csv: the 16 facility rows in section N.5.
• trade_flows.csv: the 12 trade-flow rows in section N.7.
• disagreements.csv: DG-071 to DG-082 cover the regional conflicts cited here (see R. Disagreements).
• Working paper: working-papers/wave1/regional/regional.md.