Illustrative scenario: a founder presents a fermentation cost model in a funding meeting. The plan contains a target cost per kilogram, but the proposed production site has not been tested against its utility and occupancy assumptions. This is a teaching scenario, not a verified client case.
Then one of them - a former chemical engineer who had built a bioreactor facility in Geleen, the Netherlands - asked one question.
“What are you paying per kWh?”
Silence.
The energy tariff, occupancy quotation and delivered feedstock cost must come from the proposed site. A headline national or state average may be useful for initial screening, but it is not a site quotation.
Ask this question before committing to a production location: which inputs in the cost model depend on this site?
The pattern worth testing is straightforward: a technically promising process can have different economics in different facilities. The comparison must include the actual process, scale and commercial terms.
The technology is not wrong. The postcode is wrong. And the postcode is a board-level decision that most teams treat as a logistics afterthought.
A headquarters and a production site can serve different purposes. Investor access, manufacturing infrastructure and customer proximity need not be concentrated in one address. Compare locations using the requirements of the actual operation.
One-minute brief
Innovators - Your technology does not live in a vacuum. It lives in a building, connected to a grid, fed by a supply chain. The geography of that building is a cost driver that your TEA model either captures or lies about.
A published techno-economic model has a defined scope. Check its assumptions before applying it to a particular site, rather than treating the original model as a claim about every location.
For a simple illustration, a process using 15 kWh per kilogram changes its electricity cost by EUR 0.15 per kilogram for each EUR 0.01/kWh tariff difference. This is arithmetic for stated inputs, not a measured benchmark for every fermentation process.
Geography is not a branding decision. It is a COGS decision. Treating it as anything else is the kind of assumption that collapses a cost curve in the first ten minutes of a serious due diligence conversation.
Compare like-for-like site quotations. National averages, state averages and selected industrial parks are not interchangeable, and country names alone do not establish a cost advantage.
The geography question belongs in your standard diligence checklist, not as an afterthought once the term sheet is signed.
What the evidence needs to show
Use EIA and Eurostat data to screen energy conditions, then request the actual tariff and connection terms. Match currency, period, consumption band, taxes and recoverable charges before comparing figures.
Now compare that to where the best food biotech infrastructure is actually being built.
Investor proximity and manufacturing economics are different criteria. An established food-processing region can be worth investigating, but it does not automatically offset every other cost or risk.
GFI’s review identifies feedstocks, capital costs and process performance as important cost drivers. Energy also enters fermentation, sterilisation, separation and drying. Its share depends on the process; it is not automatically the largest component.
Biomass fermentation and precision fermentation describe different product architectures. Cost ranges from one should not be transferred to the other without checking the product, recovery, scale and boundary of the calculation.
Put location-dependent assumptions directly into the model. A postcode is not a cost estimate; it identifies the site whose inputs you must obtain.
Three locations dissected
Solar Foods, Lappeenranta, Eastern Finland (not Helsinki, not San Francisco)
Solar Foods describes Factory 01 in Vantaa as its demonstration-scale Solein facility. Demonstration capacity and achieved annual output are different measures and should be reported separately.
Solar Foods selected Lappeenranta for the planned Factory 02. Its public planning and the city’s June 2026 update distinguish site selection and financing from the final construction decision and subsequent operation. Do not treat planned capacity as current production.
The city’s public project update describes a planned investment whose final construction decision was still being prepared. Treat milestones separately: financing, construction, commissioning and achieved production.
The city’s June 2026 update described grant and loan financing for the planned project while the investment decision was still being prepared. This is evidence about project development, not proof that a new factory is already producing.
This is what a geography decision made on infrastructure logic looks like. Not a founder choosing a postcode for status. An operator choosing a site for survival.
Nourish Ingredients, Leiden, Netherlands (not Sydney, not San Francisco)
Nourish Ingredients is an Australian company producing animal-free fats via precision fermentation. When the company decided to build its global commercial headquarters, it did not choose Sydney, where it was founded. It did not choose San Francisco, where its investors are. It did not choose London, where the press is.
InnovationQuarter reported that Nourish Ingredients chose BioPartner 5 in Leiden for a commercial hub, near Vivici. A commercial hub, a demonstration laboratory and a production facility are different capabilities; confirm which one your project needs.
The Netherlands offers an established food and biotechnology ecosystem. The useful site-selection task is to identify relevant customers, operators, skills and infrastructure, rather than infer a universal advantage from the number of nearby corporate names.
Investigate the specific laboratories, demonstration facilities and manufacturing infrastructure available in a proposed cluster. A corporate innovation centre is not automatically an open-access production plant.
NIZO announced the BFF-Ede scale-up project with capacity up to 10,000 litres and EUR 12.5 million in grant support. Its later kick-off announcement described a future site. Check current commissioning, compatibility and campaign availability directly.
Shared infrastructure can reduce the capital a startup has to own, where equipment, scheduling, safety and commercial terms fit the process. Its benefit must be established in the project comparison.
Compare the whole operating arrangement: access to customers, suitable equipment, qualified people, transport, utility connections and the capital required to reach the next milestone.
Quorn, Billingham, Teesside, UK (not London, not Munich)
Established fermentation production in industrial regions such as Teesside illustrates why utilities and experienced operators matter. It does not establish that the same location is optimal for another organism, product or production system.
Billingham is not glamorous. It sits within the Tees Valley industrial corridor, surrounded by chemical and petrochemical infrastructure. It is not on any startup influencer’s conference circuit.
But Billingham has three things that matter for fermentation at scale: industrial-grade utilities infrastructure (reliable power, water, effluent handling), proximity to established chemical and food-grade engineering contractors, and a workforce trained in industrial biological processes.
A historic acquisition price does not isolate the value of a location decision. Assess infrastructure on its operating requirements rather than implying that a postcode caused an exit outcome.
The location preceded the exit by three decades. The economics were built into the postcode before anyone started writing deal terms.
The geography-COGS table: what this looks like in a real TEA
The following calculation uses hypothetical sites and stated inputs. It illustrates sensitivity, not a real quotation, a full techno-economic model or a ranking of countries.
Here is what geography does to the top three cost variables that GFI’s meta-analysis identifies as the leading cost-of-production drivers.
Variable 1: Electricity
Assume identical electricity use of 15 kWh/kg. At hypothetical Site A, electricity costs EUR 0.20/kWh; at Site B it costs EUR 0.10/kWh. Electricity cost is EUR 3.00/kg and EUR 1.50/kg respectively, a EUR 1.50/kg difference before other costs.
Keep the energy-use assumption explicit. Different aeration, sterilisation, recovery and drying requirements can change it substantially.
Variable 2: Occupancy
Assume Site A costs EUR 12/m²/month and Site B EUR 6/m²/month for the same 5,000 m² floor area. These are hypothetical quotations in the same currency, period and area unit.
Annual occupancy is EUR 720,000 at Site A and EUR 360,000 at Site B. The difference is EUR 360,000 per year.
At a hypothetical annual saleable output of 1,000 tonnes, that occupancy difference is EUR 0.36/kg. Lower utilisation would increase its cost per saleable kilogram.
This comparison excludes construction, fit-out, maintenance, effluent, taxes, financing and location-dependent differences in yield. Add them before making a site decision.
Variable 3: Feedstock Access
Port or processor proximity can affect freight, handling and supplier options. Obtain delivered quotations and quality specifications. Port throughput does not establish a fixed percentage saving for a fermentation business.
National harvest totals cannot establish regional feedstock availability at a specific site. Confirm the quantity, composition, seasonality, pre-treatment, competing uses and delivered price with actual suppliers.
A site near agriculture is a lead to investigate, not proof of a suitable low-cost substrate. Transport and supplier margins are only part of the input specification.
The combined arithmetic
For these hypothetical inputs only, the electricity and occupancy differences sum to EUR 1.86/kg. The calculation does not establish a California-versus-Europe cost gap or a full production cost.
Recalculate with the process-specific energy use, saleable output, utilisation, capital requirements and logistics of each real offer. A partial comparison cannot establish overall price parity or financing readiness.
The location question is part of the operating model. Resolve it alongside biology, recovery and customer requirements, rather than treating it as a universal alternative to scientific improvement.
The real pattern: where value is created vs where decks are written
Location choices need evidence at the site level, not a ranking of cities.
A headquarters can support fundraising and customer relationships. A production site must also meet the requirements of the process.
For each proposed location, investigate the following conditions:
Site-specific utility tariffs, connection capacity and reliability.
Compatible shared facilities, where available, with confirmed equipment and campaign terms.
Delivered feedstock specification, quantity, price and alternative suppliers.
Applicable financing programmes and their actual eligibility conditions.
Access to experienced operators, engineering support and maintenance.
Technical, institutional and commercial constraints can all limit fermentation scale-up. Shared facilities can help teams test processes, but their existence does not establish readiness for a particular product.
A location plan should consider what happens if the required equipment, people, funding or permits are not available on time.
Compare those dependencies before committing to a site.
The EU-CEE grant trap: same problem, different form
There is a CEE-specific version of this geography mistake that deserves its own paragraph.
A capital-city address does not establish evaluator preference, and an industrial address does not establish low cost. Compare the actual requirements and available sites.
Compare specific sites, including industrial regions and capital-city locations, on the same operating requirements.
Grant conditions vary by programme and call. Check eligible costs, geography, funding rate, milestones, procurement obligations and reporting requirements. Do not assume that European programmes reward one type of patent or address over commercial progress.
For a CEE business, manufacturing, customer access and distribution can point to different locations. The comparison should be supported by supplier, operator and buyer information, rather than inferred from a brand’s fundraising or retail footprint.
The diagnostic: seven questions before you sign a lease
This is the site selection filter to apply before any food biotech company commits to a location for its first real production infrastructure. Run this against your current postcode. Run it against the postcode you are considering. Be honest about what the numbers say.
Question 1: What is the industrial electricity rate at this site?
Get the actual site tariff, connection capacity, reliability provisions and charges. For screening data, compare the same period, currency and consumption band; do not mix a selected industrial park with a national average.
This number belongs in your TEA before it belongs in your deck.
Question 2: What is the farm-gate cost of your primary fermentation feedstock from this site?
For sucrose, glucose, molasses or another substrate, ask for the delivered specification and cost. A national crop total or a supplier’s proximity does not establish suitability, availability or a usable price.
Question 3: Can you access shared pilot infrastructure within 60 minutes of this site?
Compare the cost of compatible shared capacity with ownership using quotations, availability and process constraints. No universal capital saving or vessel-size cutoff applies. Check whether upstream and downstream equipment are both included.
If suitable shared capacity is unavailable, identify the alternative and its implications for capital, schedule and technical risk.
Question 4: What EU or national co-financing mechanisms are accessible from this site?
Check project-specific co-financing and its conditions. NIZO’s BFF-Ede announcement is one example of public support for shared infrastructure, not evidence that the same grant is available to every company or location.
Eligibility depends on the award’s geographic and project conditions.
Question 5: What is the industrial space cost per square metre per month at this location?
Use comparable quotations: the same currency, floor-area unit, period, fit-out scope and service charges. In the hypothetical example above, EUR 12 versus EUR 6 per m² per month produces a EUR 360,000 annual difference across 5,000 m².
At 1,000 tonnes of saleable annual output, that difference is EUR 0.36/kg. It is not an observed rent gap between named regions.
Question 6: What is your permitting timeline, and does this site have food-grade or pharmaceutical-grade precedent?
Check the permitting pathway, required classification and likely timetable with the relevant authority. Existing industrial use may help identify constraints, but it does not guarantee a shorter approval.
Site precedent is one input to that assessment.
Question 7: Does your B2B customer have procurement operations within commercial logistics distance of this site?
Identify where the actual customer tests, approves, buys and receives the ingredient. A corporate headquarters or commercial hub does not necessarily identify its procurement or manufacturing location.
Confirm the relevant customer team and logistics requirements directly.
The Geo-Screen for investors: three filters
When evaluating a food biotech startup’s geography decision, apply these three filters before accepting “we’re based in SF because our investors are here.”
Filter 1: Energy tariff vs COP model alignment.
Check the cost model against a current site quotation. If it uses screening data, label that assumption and show sensitivity to plausible site-specific variation.
Filter 2: Feedstock sourcing distance and margin stack.
Ask for the delivered feedstock specification and cost, then test sensitivity. For illustration only, an 18 percent reduction in a feedstock line representing 40 percent of a EUR 5/kg total would reduce that total by EUR 0.36/kg if every other input stayed constant. Those percentages are scenario inputs, not a measured port-access saving.
Filter 3: Capex per litre vs shared infrastructure access.
Compare compatible shared capacity, co-location and ownership without a universal volume or saving threshold. Include scheduling, yield, quality, downstream requirements and commercial terms in the comparison.
Diagnostic Questions
For founders: If your cost curve in your investor deck uses a national electricity average rather than the site-specific industrial tariff, how many cents per kWh is your model off by, and what does that do to your COGS in Year 3?
Run the number before someone else runs it for you.
For investors: Of the food biotech companies in your current portfolio, how many have cost models built from site-specific utility costs rather than national averages?
For those based in San Francisco or London, has anyone stress-tested the geography premium against the cost structure presented at Series A?
For corporates: If a target company’s cost curve assumes utility costs that are structurally incompatible with the site they have chosen, what is the valuation adjustment at close, and is your current diligence process designed to find it before you sign, or after?
What is the actual industrial electricity tariff at your production site today, and how does it compare to the number currently in your cost model?
Sources and scope
This is an edited adaptation of the author’s full Substack essay, first published 2026-03-31. Examples and editorial frameworks are distinguished from measured evidence. Source checks and editing: 16 October 2026.
- GFI, techno-economic insights
- Lappeenranta, planned Solar Foods project, June 2026
- InnovationQuarter, Nourish commercial hub
- NIZO, BFF scale-up project
Read the original Substack version. Read the FoodEdge editorial standards or suggest a correction.
Cover photograph: Photo by Trnava University / Unsplash. Context image, not a photograph of a company or experiment discussed here. Photo licence.
