The Abu Dhabi Investment Office’s release of 23 October 2025 says the partnership with The EVERY Company and Vivici “will explore the establishment of a 4-million-litre industrial-scale facility for alternative protein production.” That is the entire quantitative content of the announcement. Ten months later, no capital cost, no site, no completion date, no equipment order, no anchor offtake and no final investment decision has been disclosed by ADIO, by either company, or by anyone else we could find.
Meanwhile the figure has hardened. Coverage moved from “explore” to “to set up,” to “to establish,” to — in one dairy trade outlet — “Fonterra has invested in a 4-million-litre precision fermentation facility in the UAE.” Fonterra is not named anywhere in the ADIO release.
What the release actually says, and what it omits
The release sits on the Abu Dhabi Media Office site, dated 23 October 2025. An identical copy was carried on ADIO’s own investwithabudhabi.com newsroom at the time of the announcement; that URL now returns 404, so the media office page is the surviving primary text and the one cited here. The photograph filename on the media office page — PALT_MOU — is the closest thing to a document type the release gives. The word “MoU” does not appear in the body.
| Item | Status in the primary release |
|---|---|
| Volume | “4-million-litre industrial-scale facility” — no split between number of vessels, working volume or total installed |
| Verb | “will explore the establishment of” |
| Capital cost | Not stated |
| Site or emirate sub-location | Not stated |
| Timeline or first-production date | Not stated |
| Equipment order or EPC contractor | Not stated |
| Final investment decision | Not stated; the release describes “designing, financing and commercialising” as forward work |
| Ownership or SPV structure | Not stated |
| Offtake or anchor tenant volumes | Not stated |
| Regulatory status of the products in the UAE | Explicitly forward-looking: the partnership “will support the creation of a regulatory framework for fermented proteins” |
| Multi-tenancy | Stated as a design intent: “designed to be multi-tenanted to allow for the participation of other fermented protein companies in the future” |
Note the last two rows together. ADIO is not describing a plant awaiting a permit under an existing regime; it says a regulatory framework for fermented proteins in the UAE has yet to be created, and that this partnership will help create it. An operator reading this as a near-term supply option should register that the approval pathway for the output is itself part of the workplan.
The release’s only other number is market-sizing: the protein fermentation market “valued at US$3.03 billion in 2024, is projected to expand to US$54.04 billion by 2032, at a compound annual growth rate of 43.5 per cent.” ADIO names no source for it. The arithmetic is at least internally consistent — 54.04 divided by 3.03 is a factor of 17.8, which over eight years is a 43.4% CAGR — but a figure with no attribution is not evidence of anything, and it is not a capacity claim.
Practical consequence: treat this as a pre-feasibility announcement. There is nothing in it an operator can put in a supply plan, and nothing an investor can underwrite. The single actionable line is the multi-tenancy intent, and that is a design principle, not an allocation.
The 4-million-litre number already had an owner
Here is the detail that changes how the figure reads. In April 2025 — six months before the ADIO announcement — AgFunderNews reported that Liberation Labs had received a US$1.39 million award from the US Department of Defense to conduct a feasibility study into expanding its Richmond, Indiana site “to a 4-million-liter facility.” The same piece describes Richmond as “a 600,000-L facility that could scale up to 4-million-liters.”
So 4 million litres is the published full-build target for a Richmond-type plant. And Vivici is Liberation Labs’ first commercial customer at Richmond: the 7 May 2025 announcement confirms Liberation Labs will produce Vivici’s Vivitein BLG there when commercial production begins in 2026.
We want to be exact about what this does and does not establish. No source we fetched connects the two. Liberation Labs is not a named party to the ADIO partnership, and neither ADIO nor Vivici has said where the 4-million-litre figure came from. But the number is not an arbitrary round one. It is the size of a known, engineered, four-times-Richmond expansion module circulating in exactly the commercial neighbourhood these parties occupy — which is a considerably more plausible origin for a headline volume than a bespoke Abu Dhabi front-end engineering study that nobody has mentioned.
Practical consequence: when a capacity number in an exploratory announcement matches a design template already public elsewhere, read it as an aspiration expressed in a vendor’s units, not as an output of site-specific engineering. Ask whose FEL-3 package it came from.
What 4 million litres would mean against verified reference points
Assembled from primary sources we fetched. The units column is the point of the table.
| Asset | Figure | Unit as stated by the source | Status |
|---|---|---|---|
| Proposed Abu Dhabi facility | 4,000,000 litres | “4-million-litre industrial-scale facility” — basis not specified | Agreement to explore, 23 Oct 2025 |
| Liberation Labs, Richmond, Indiana | 600,000 litres | “fermentation capacity of 600,000 litres with a fully dedicated DSP”; 150,000-litre fermenters | Under construction; production expected to begin 2026 |
| Liberation Labs, Richmond, full build | 4,000,000 litres | DoD-funded feasibility study target | Feasibility study only |
| Liberation Labs, Saudi Arabia (NEOM/Topian) | Would “mirror” Richmond | No volume stated for the Saudi site | Feasibility study signed; CEO on record that there is “no firm commitment at this point” |
| Solar Foods Factory 01, Finland | 20,000 litres | Single bioreactor volume | Operating since April 2024 |
| Solar Foods Factory 01, output | 160 tonnes/year Solein | Annual design capacity, whole-cell biomass via gas fermentation | Confirmed reached, Oct 2025; target 230 t/yr in 2026 |
| The EVERY Company, 2025 | 63,000 litres | Industrial fermentation volume EVERY was running at; see our earlier report | Historic; doubled at the start of 2026 |
| Vivici, own plant | No source states it | Vivici has disclosed no owned fermentation volume | Uses third-party European capacity plus Richmond |
Two things fall out of this.
First, the scale gap. The largest food-grade precision fermentation plant either named company can point to today is a 600,000-litre site that has not yet begun commercial production. The Abu Dhabi figure is 6.7 times that entire plant, at a location with no announced site.
Second, and more important for anyone modelling this: you cannot convert 4 million litres into tonnes of protein. Doing so requires a titre (grams of target protein per litre of broth) and a downstream recovery yield, and neither Vivici nor The EVERY Company has disclosed either. Solar Foods shows why the temptation should be resisted: a single 20,000-litre bioreactor yields 160 tonnes a year — but that is continuous gas fermentation harvesting the whole cell as the product, where essentially everything grown is sellable. A precision fermentation process expressing a single secreted protein in fed-batch, then purifying it, operates on a completely different basis. Applying Solar Foods’ litres-to-tonnes ratio to a precision fermentation plant would overstate output by a wide and unknowable margin. This is the same units problem we mapped across the sector in our capacity map, where four incompatible units were in circulation for the same question.
On capex, the only verified food-grade per-litre anchor we have is Liberation Labs’ Richmond project: approximately US$115 million for the 600,000-litre first phase. We are deliberately not multiplying that out. Capex does not scale linearly with volume, Gulf construction costs, water costs and utility tariffs differ from rural Indiana, and — as we found looking at vendor capex heuristics for Chinese fermentation assets — per-litre rules of thumb are exactly the kind of number that becomes a fact by repetition. The honest statement is: no capex has been disclosed for Abu Dhabi, and the nearest verified comparator is $115m for 600,000 litres.
Practical consequence: any model that turns “4 million litres” into tonnes of ovalbumin or beta-lactoglobulin is inventing a titre. If a banker or a broker shows you that conversion, ask which titre and which recovery yield they used, and where those came from.
The second beat: how “explore” became “Fonterra invested”
Fonterra and dsm-firmenich are Vivici’s founding shareholders — AgFunderNews describes Vivici as “a Dutch startup formed by Fonterra and DSM-Firmenich in December 2022,” and Vivici’s own Series A release of 25 February 2025 names both as “existing shareholders” who continued to support the €32.5 million round led by APG on behalf of ABP and Invest-NL. (A minor discrepancy worth noting: that release says Vivici “was founded in 2023,” while AgFunderNews dates the formation to December 2022. The European Commission clearance and the start of operations plausibly straddle the two; nothing in this article turns on it.)
What Fonterra has not done is invest in an Abu Dhabi facility. It is not named in the ADIO release. No Fonterra statement about the Abu Dhabi project surfaced in our searches, and we found no disclosure of any Fonterra investment figure attached to it.
Yet on 31 October 2025, The Dairy News published a piece headlined “Fonterra Expands into Precision Fermentation with UAE Facility Investment,” opening: “Fonterra has invested in a 4-million-litre precision fermentation facility in the UAE.” That is a compound error. It converts a shareholding in one of three partners into a direct investment; it converts an agreement to explore into a facility; and it attaches a volume to an asset that does not exist.
The drift is visible across the whole cycle:
| Outlet | Date | Headline verb | Body verb |
|---|---|---|---|
| Abu Dhabi Media Office / ADIO | 23 Oct 2025 | “partners with … to scale” | “will explore the establishment of” |
| AgFunderNews | 23 Oct 2025 | “explore 4m-liter alt protein facility” | “explore the establishment of” (quoted) |
| The National | 23 Oct 2025 | “aims to become centre” | “They plan to set up a four-million-litre production site” |
| Green Queen | 24 Oct 2025 | “to Set Up Large Animal-Free Protein Factory” | “to explore the establishment of” |
| The Dairy News | 31 Oct 2025 | “Fonterra Expands … with UAE Facility Investment” | “Fonterra has invested in a 4-million-litre precision fermentation facility” |
AgFunderNews is the only outlet in this set whose headline verb matches the primary. Green Queen’s body text is accurate and its headline is not, which is the ordinary mechanics of how this happens — headlines are written to a different constraint than the paragraph beneath them, and aggregators downstream read the headline.
Practical consequence: when a capacity figure reaches you, walk it back to the government or company release that first carried it and read the verb. In this case the walk-back takes two clicks and changes the meaning of the number entirely.
What “multi-tenanted” does and does not tell you
ADIO’s wording is that the facility “will be designed to be multi-tenanted to allow for the participation of other fermented protein companies in the future.” Read precisely, this is a statement about design flexibility at some future point — not an allocation, and not an open-access commitment.
For an operator, the questions multi-tenancy raises are all unanswered here. Is capacity contracted as dedicated trains or as campaign slots on shared equipment? Are the two announced partners anchor tenants with priority, and if so at what share? Is there segregation sufficient for GMO-organism separation and allergen control between tenants? Who owns and who operates? None of this is stated, and at the exploratory stage it would be surprising if it were.
It is worth contrasting the one comparable asset where tenancy is being tested in the real world. Liberation Labs told AgFunderNews in April 2025 that it held “letters of intent representing over 200% of the available capacity for the first five years of operations” at Richmond. That is an oversubscribed 600,000-litre plant before it opens — which is a useful signal about demand for third-party food-grade fermentation capacity, and also a reminder that LOIs are not contracts.
Practical consequence: “multi-tenanted” in an exploratory release is a marketing posture toward future tenants, not an invitation you can respond to. There is no capacity to reserve until an SPV exists.
The counter-argument, stated properly
The serious objection to this piece runs like this: exploratory agreements backed by a sovereign investment office are a different asset class from startup MoUs. ADIO is not a founder chasing a headline; it is a government body with a mandate, a budget and land. Abu Dhabi’s AGWA cluster targets AED 128 billion (about US$34.8 billion) of investment and AED 90 billion (about US$24.5 billion) of incremental GDP by 2045, with 60,000-plus jobs. When a body like that says it will explore something, the conversion rate to steel in the ground ought to be materially higher than for two startups signing a memorandum at a conference. Add sovereign offtake demand, a stated intent to build the regulatory pathway, and CEPA export routes, and the case for taking the 4-million-litre figure seriously is real.
That argument deserves weight. But it is testable, and the test is unflattering in one specific respect: AGWA launched in 2024 as ADIO’s second cluster, and we could not find a public record of any AGWA-anchored fermentation facility that has reached construction, let alone production. Its published targets run to 2045. That is a twenty-one-year horizon, which is entirely reasonable for a cluster strategy and entirely useless as a guide to whether a specific plant gets built this decade.
The sharper evidence comes from the closest available analogue. Liberation Labs signed with the NEOM Investment Fund in April 2025 to explore a precision fermentation plant in Saudi Arabia — a Gulf sovereign vehicle, a named site sponsor, an investor already on Liberation’s cap table. Asked directly about funding, CEO Mark Warner told AgFunderNews: “The agreement we have signed is to do a feasibility study, but obviously, part of the reason we partnered with NIF is they have the ability to bring capital to the table, although there is no firm commitment at this point.” On his own timeline, a 6–12 month feasibility study would be followed by a build of “a couple years.”
That is an operator who wants the deal to happen describing what a Gulf sovereign exploratory agreement is worth, on the record, in his own interest. Even on his optimistic sequencing, an October 2025 exploration agreement puts first production somewhere in 2028 at the earliest — and only if the study clears, an FID follows, and capital is committed. The counter-argument is right that sovereign backing raises the odds. It does not compress the schedule, and it has not, so far, produced a disclosure.
Practical consequence: the correct adjustment for sovereign backing is to raise your probability of eventual construction, not to advance your date of first supply. Those are different variables and coverage routinely collapses them.
What we could not establish
- Any capex figure for the Abu Dhabi facility. None disclosed by ADIO, Vivici or The EVERY Company.
- A site. No location within Abu Dhabi has been named — not KEZAD, not Masdar City, not any other zone.
- A timeline. No study completion date, FID date, groundbreaking date or first-production date exists in any source we fetched.
- Equipment. No fermenter order, vendor, EPC contractor or FEL stage has been disclosed.
- The basis of the 4-million-litre figure. Whether it denotes aggregate working volume, total installed volume or nameplate is not stated. This matters: those three numbers can differ by 30% or more for the same plant.
- Any statement from Fonterra or dsm-firmenich on the Abu Dhabi project. We found none. We also found no evidence for the investment attributed to Fonterra by The Dairy News.
- Vivici’s or The EVERY Company’s own newsroom treatment. We were unable to fetch
vivici.comdirectly in this session; the company’s own framing of the partnership could not be checked against the ADIO text. - Whether the deal is a binding agreement or an MoU. The photograph filename on the media office release contains “MOU”; the body text does not use the term.
- Any AGWA-anchored fermentation asset that has reached construction. Searched; nothing found. Absence of evidence here is weak evidence, and we state it as such.
- Any follow-up announcement between 23 October 2025 and 20 August 2026. Ten months, no update located from any of the three parties.
What to watch
Five testable claims. Each is falsifiable and each would move the picture.
- An SPV or JV entity is registered in Abu Dhabi with named shareholders and a paid-up capital figure. This is the first hard evidence that “designing, financing and commercialising” has moved past a workplan. Absent this, there is no vehicle to hold an asset.
- A UAE regulatory pathway for fermented proteins is published by ADAFSA or a federal body. ADIO says the framework does not yet exist. Until it does, neither Vivici’s beta-lactoglobulin nor EVERY’s ovalbumin has a domestic approval route, and the plant would be an export-only asset from day one.
- A fermenter order is placed. Long-lead stainless vessels at 150,000-litre-plus scale are the single most reliable public signal that a project is real. Watch for vendor announcements rather than sponsor announcements.
- The 4-million-litre figure is restated with a basis, or is revised down. Our expectation, on the evidence, is that any project that proceeds will proceed in a phase considerably smaller than 4 million litres — the Richmond precedent is a 600,000-litre first phase with the 4-million-litre number attached to a full build. A phase-one figure appearing in future coverage would confirm that reading.
- Liberation Labs’ Richmond plant reaches commercial production and Vivici’s Vivitein BLG ships from it. Production was scheduled to begin in 2026. If the industry’s most advanced purpose-built food-grade site slips, the base rate for a larger greenfield plant in a jurisdiction with no fermented-protein regulatory framework should be revised accordingly.
None of this means the Abu Dhabi project will not happen. It means that as of today the only verified facts are a date, a verb, a volume with no stated basis, and three names on a release. Everything else being written about it is extrapolation.