Beliefs form — open for post-workshop updates. If you attended or engaged with the workshop, please also complete the brief feedback survey → ($50 random draw + $50 most useful response award). Questions welcome.
The Unjournal · Pivotal Questions Initiative

Cultivated Meat Pivotal Questions

State your beliefs on cost trajectories, commercial viability, and animal welfare implications of cultivated meat.

💬 Discuss: Annotate via Hypothes.is  ·  GitHub Discussions: 💬 Workshop · 🧠 Substantive · 🎯 PQ framing · 📖 map
Workshop confirmed: Fri 8 May 2026 · 11am–3pm ET (17:00–21:00 CET) · see agendaAlso planned:
Wed 6 May, 11am–12pm ET — informal pre-session.
Fri 8 May, 9:00–10:00am ET — European morning drop-in (15:00–16:00 CET).
Async materials available to all registered participants.  ·  Express interest →

These are key operationalized questions from our Cultivated Meat Pivotal Questions project. We want to elicit expert and stakeholder beliefs — before, during, and after reviewing the TEA evidence — to see how views evolve and where consensus exists or doesn't. See our resources page for the key papers informing these questions. (All questions are optional.)

New: Watch the beliefs form walkthrough video for a guided tour of these questions before you fill this in.

Sharing policy: Response content will be shared within the participant group and published publicly. You can choose whether to keep your individual responses anonymous (not attributed to you by name) or allow others to see your name alongside your response — simply leave the name field blank to remain anonymous.

▸ Canonical PQ questions (Coda)
For more detail, context, and the complete set of operationalized questions, including those we're commissioning our PQ evaluators to address, see the canonical PQ formulations on Coda → (use the PQ filter bar and select "cell cultured meat price and cost" to see CM-specific questions)

You don't need to be a specialist to contribute. We want your honest assessment and reasoning, whether you feel highly confident or very uncertain. Your input helps us understand the range of views in the field.

Several CM cost questions are also live on Metaculus for public forecasting.

▸ Public forecasting on Metaculus [?]Metaculus provides a public interface for a broader pool of forecasters. At least one CM question is going into the animal welfare impact forecasting tournament. Metaculus also allows more detailed belief distributions than the form below. Your Metaculus forecast feeds into our aggregation — please share your username in the form below if you forecast there.

The focal cost question is live on Metaculus: CM_01 — production cost per kg →. The expert-aggregation version is at CM_03 →. If you forecast on Metaculus, please share your username below so we can link your contributions.

How to respond

Please estimate independently of the cost model. Don't look up what the model currently assumes and enter that — we would just be recycling our own starting point. Your own knowledge, judgment, and calibrated uncertainty is what's informative. [if you do use it]If you do open the model and use its sliders or central estimate as a reference, that's fine — but please note this in your reasoning field or a Hypothes.is comment, and explain what you changed and why. A comparison between your independent view and the model's assumption is useful information too.

▸ Guidance on forming your estimates and uncertainty bounds (Anca Hanea)

Anca Hanea (U Melbourne), expert in structured expert elicitation and developer of the IDEA protocolIDEA: Investigate, Discuss, Estimate, Aggregate — a structured protocol designed to extract calibrated expert judgements and combine them while minimising anchoring, groupthink, and overconfidence. See: Hanea et al. (2021), Risk Analysis., recommends this order to avoid anchoring and overconfidence when forming a credible interval:

  1. Lower bound first (your p10): Think of all the reasons this value might be small — technological breakthroughs, optimistic scenarios, best-case conditions. What is a plausible floor? This is the number below which you assign only a 1-in-10 chance of the true value falling.
  2. Upper bound second (your p90): Think of all the reasons this value might be large — setbacks, unexpected hurdles, worst-case scenarios. This is the number above which you assign only a 1-in-10 chance.
  3. Median last: Having fully engaged with both extremes, find your balance point — the value for which you judge an even chance that the true outcome is above or below.

Starting with the median and adding bounds afterwards tends to produce intervals that are too narrow (anchoring + overconfidence). When forming each bound, ask: "What factors could drive this much higher or lower — and how variable are those factors?" Fold that variability into the bounds. If there are things you genuinely don't know, fold that epistemic uncertainty in too — wider bounds are the honest reflection of your state of knowledge.

This form is a lightweight structured survey — not a formal expert elicitation protocol.

Shared Definitions

Average Production Cost (AC) = (Annualized capital charge + all operating costs) ÷ annual kilograms of cultured chicken cell biomass (wet weight, at harvestWet weight = the mass of cells as harvested from the bioreactor, including water. Typically ~75–90% water depending on cell line, density, and how much dewatering occurs during separation. We use 80% as our reference assumption. See the 'Cultured-chicken meat' definition above for full details.)
  • Capital charge: Total capital investment (bioreactors, facility) amortized over the plant life using the Capital Recovery Factor, adjusted for financing costs (WACC)
  • Operating costs include: Basal media (amino acids, glucose, vitamins, buffers), recombinant growth factors, utilities, consumables, labor, maintenance, and plant overhead
Does not include: Downstream scaffolding added during product structuring; texturization; blending with plant-based or other non-cellular ingredients; packaging; distribution; retail markups; R&D amortization; regulatory approval costs; or marketing.

Output basis note: "Wet weight, at harvest" = the mass of cells after separation from spent media, before any further processing. This is the same object as "edible kg before mixture with plant products" — at the harvest stage, wet-weight cell biomass is the edible component. This aligns with the cost modelA TEA (techno-economic analysis) framework built for this project by Jakub Kozlowski, drawing on Humbird (2021), Pasitka et al. (2024), and CE Delft (2021). It models undifferentiated cell biomass production at commercial scale. Results are illustrative — not a prediction of what any specific plant will achieve., which expresses all outputs on the same wet-weight, factory-gate basis.

Why average production cost? Standard economic theory predicts that competitive markets drive prices toward the minimum of the long-run average total cost curve. This represents the lowest sustainable price in a free market. We ask for the median estimate (your 50th percentile) across the range of plausible 2036 technology scenarios, reflecting your uncertainty about which trajectory the industry takes.

Cultured chicken cell biomass: Undifferentiated (proliferatingProliferating: actively dividing — cells in the growth and expansion phase, replicating via cell division. In CM, this is the bioreactor scale-up phase before any directed differentiation into muscle, fat, or connective tissue cells.) chicken cell mass produced in bioreactors — prior to any directed differentiation step — measured before any blending with plant-based or other non-cellular ingredients, and before texturization or downstream structuring.

Scaffold assumption: For this exercise, please assume that harvested cell biomass contains minimal non-degraded scaffolding material (<10% w/ww/w = weight per weight. Here: the mass of scaffold material as a percentage of the total wet-weight harvest mass.). Most TEAs for unstructured products (ground meat, hybrid products) assume suspension culture where this is not an issue. If your estimate involves scaffold-based production where structural material remains in the harvest, please note this and base your cost estimate on the cellular fraction only. We measure on a wet-weight basis (as harvested from the bioreactor — typically ~75–90% water depending on cell line, density, and post-harvest dewatering; we use 80% as our reference assumption). [hydration note]Water content at harvest varies substantially (potentially 75–95%) depending on: cell type and size, bioreactor density, centrifugation protocol, and how much dewatering occurs during cell separation. Published TEAs often don't state their assumed hydration, which means cost comparisons across papers may be subtly inconsistent — a 10% hydration difference can substantially alter $/kg wet weight. We use 80% as a reference. Feedback welcome: if you have a view on the right convention or your estimate assumes a specific hydration, please note it in your reasoning field or as a Hypothes.is annotation on this page.

Alignment note: "Before mixture" and "wet weight at harvest" refer to the same accounting object. At the factory-gate stage, the wet-weight cell biomass is the edible cultured component — nothing non-edible has been added yet. So "edible kg before mixture," "wet-weight cell biomass at harvest," and "pure cell mass" are all equivalent at this stage.

Why chicken? Among cultured meat products, chicken has the highest stakes for animal welfare (assuming 'like substitutes for like'). Forecasters may base estimates on bovine or generic-mammal TEAs and explain their conversion method.

Why "before blending"? Animal cells are often mixed with plant-based inputs (hydrolysates, plant protein) to reduce total product cost. We focus on the pure cell biomass cost because: (1) it is the technically challenging and costly component; (2) it is what the cost model estimates; (3) blended product costs can be derived from it given a mixing ratio and filler cost.

Large-scale plants: Plants producing more than 2 kt/yr (2,000 metric tons per year). We exclude smaller experimental/R&D facilities.

Value units: Inflation-adjusted (CPI) 2025 US dollars.

Target years:

  • Very-short-run: End of 2026
  • Short-run: December 31, 2031
  • Medium-run: December 31, 2036 focal year
  • Longer-run: December 31, 2051

Why these years? These are the years Rethink Priorities chose for their forecasting work, allowing us to harmonize and compare estimates. 2051 represents a "30-year terminal horizon" when production would have more or less achieved an efficient optimum.

When we ask for a probability, we're asking for your best calibrated subjective probability—your honest credence given everything you know.

One way to think about this: Imagine an ideal research team with unlimited resources, time, and data—perhaps even a kind of omniscience where they could perfectly understand the welfare and psychological states of everyone affected. What probability would you assign that this idealized team would ultimately conclude the statement is true?

Note: We avoid anchoring to "0% = impossible" and "100% = certain" because perfect certainty is rarely justified. [more]"Extremely unlikely but not impossible" should be a low but nonzero probability; "nearly certain but not absolutely so" should be high but below 100%. Use your own calibrated judgment — a wide interval is more honest than false precision.

CM_01 · Focal Question · Production Cost

What will be the average production cost per kg of undifferentiatedUndifferentiated cell biomass: Proliferating cell mass prior to any directed differentiation step. This is what most published TEAs (Humbird, Pasitka et al., CE Delft) model. It includes spontaneously immortalized cell lines such as those used by Pasitka et al. (2024), which proliferate indefinitely without a separate differentiation phase. Cost estimates for products requiring differentiated muscle fibers (structured whole cuts) would typically be higher; this question focuses on the proliferation/scale-up phase. cultured chicken cell biomass (wet weight, at harvestWet weight = the mass of cells as harvested from the bioreactor, including water. Typically ~75–90% water depending on cell line, density, and how much dewatering occurs during separation. We use 80% as our reference assumption. See the 'Cultured-chicken meat' definition above for full details.) in 2036As of December 31, 2036., across all large-scale plants in the world?

See definitions above for Average Production Cost, cultured chicken cell biomass, and large-scale plants. Costs in 2025 USD. We ask for your median estimate — your 50th percentile across the range of plausible 2036 outcomes. (other years?)You can provide projections for other target years in the discussion section below, or forecast a full distribution on Metaculus →, which also allows a broader range of target years.

This is the central question for assessing CM's commercial viability and animal welfare potential. It is also what the cost modelA TEA framework built for this project — useful for exploring how different assumptions affect cost. Its p50 is its median cost forecast, in the same units as your estimate here. Please don't use the model's default values as the basis for your estimate — your independent assessment is what's most informative. If you do explore it, note this in your reasoning and explain what you changed. focuses on — but please estimate from your own knowledge and expertise, not from the model's current central value. We strongly encourage completing the Technical/Process questions below — they help us understand why experts disagree and where consensus exists.

Your 50th percentile — half of plausible 2036 outcomes are above this, half below. [full distribution?]To express a fuller belief distribution (not just a median and CI), you can also forecast on CM_01 on Metaculus → — Metaculus lets you specify a complete probability distribution. If you forecast there, please share your username in the "About You" section below so we can link the two.

CM_01 · Supplementary · 2027 Near-term Target

Same question as CM_01, but… as of December 31, 2027, across all large-scale plants in the world? [why 2027?]2027 outcomes will be observable within a few years — allowing us to compare forecasts against actual data, helping calibrate longer-horizon estimates. Same definitions and basis as CM_01. [scope note]If you anticipate no large-scale plants, please give your estimate for advanced pilots and explain.

Goal-oriented & animal welfare investment questions

These questions connect the cost trajectory to the actual funding decisions animal welfare organisations face.

Technical/Process questions

The headline cost question (CM_01) depends on several technical cost drivers. Your views on these subquestions help us understand why experts disagree and identify the major remaining sources of uncertainty. Particularly valuable if you have direct expertise in bioprocess engineering, cell culture, or TEA. [independence note]Please form your own estimates for each question — not by looking up what the cost model currently assumes for these parameters. Your independent assessment is what's most informative. If you do use the model as a reference, note it in your reasoning.

CM_12 · Hydrolysates

Will most cultured meat (by volume) be produced using hydrolysates as the base media (replacing expensive purified amino acids) in 2036?

—
CM_13 · Growth Factor Costs

What will be the total cost of growth factorsCommon growth factors used in CM cell culture include:
FGF-2 (Fibroblast Growth Factor 2) — most widely used; promotes proliferation
IGF-1 (Insulin-like Growth Factor 1) — supports growth and survival
EGF (Epidermal Growth Factor) — stimulates cell division
TGF-β (Transforming Growth Factor beta) — regulates differentiation
Insulin — metabolic support; relatively cheap
Transferrin — iron transport; serum supplement
Approaches to reduce GF costs include: precision fermentation (yeast/bacteria-produced), plant molecular farming, autocrine cell lines (self-producing), thermostable variants, and small-molecule substitutes. per kg of undifferentiatedUndifferentiated cell biomass: Proliferating cell mass prior to any directed differentiation step. This is what most published TEAs (Humbird, Pasitka et al., CE Delft) model. Cost estimates for differentiated products would typically be higher. cultured chicken cell biomass (wet weight, at harvestWet weight = the mass of cells as harvested from the bioreactor, including water. Typically ~75–90% water depending on cell line, density, and how much dewatering occurs during separation. We use 80% as our reference assumption. See the 'Cultured-chicken meat' definition above for full details.) produced in 2036?

Your combined estimate of growth factor costs per kg of biomass produced. For a decomposed view, see E2 and E3 below →E2 asks for the price per gram of the dominant growth factor ($/g); E3 asks for the effective dosage (g of GF per kg of biomass produced). E2 × E3 = total GF cost per kg — which should match your CM_13 estimate. If you fill in E2 and E3 in the Detailed Technical Questions section ↓, you can leave CM_13 blank or use it as a consistency check.

CM_14 · Cell Media Costs

What will be the cost of cell media per kg of undifferentiatedUndifferentiated cell biomass: Proliferating cell mass prior to any directed differentiation step. This is what most published TEAs (Humbird, Pasitka et al., CE Delft) model. Cost estimates for differentiated products would typically be higher. cultured chicken cell biomass (wet weight, at harvestWet weight = the mass of cells as harvested from the bioreactor, including water. Typically ~75–90% water depending on cell line, density, and how much dewatering occurs during separation. We use 80% as our reference assumption. See the 'Cultured-chicken meat' definition above for full details.) in 2036?

Does not include growth factor costs — those are captured separately in CM_13 above.

CM_16 · Bioreactor Cell Density

For commercially relevant chicken cell culture in large-scale bioreactors (≥10,000L) in 2036, what is your distribution for the typical harvest cell density?

We are asking about typical commercial production — the density that most large-scale plants would realistically achieve — not the frontier maximum. The model uses density as a key cost driver: higher density = less media per kg of output. Please specify g/L (grams per liter wet weight).

Volumetric productivity captures both cell density and how often biomass is effectively extracted — for fed-batch: density at harvest ÷ total cycle time; for perfusion/continuous: density × effective harvest fraction per day. Determines bioreactor volume required per kg/year of output. State your assumed process mode in the text box below.

CM_17 · Food-grade vs Pharmaceutical

By 2036, what percent of commercial cultured meat will be produced using food-grade cell media (as opposed to pharmaceutical-grade)?

—
CM_20 · Custom Bioreactors

What share of cultured meat companies (those with capex over $10 million) will design and build their own bioreactors by 2036?

Some reports suggest companies can design and build their own bioreactors for much lower cost than purchasing off-the-shelf pharmaceutical bioreactors.

—
More detailed technical questions · Particularly valuable for bioprocess/TEA experts

More detailed technical questions

This section goes into more detail behind the estimates above — asking about specific cost-driver parameters: growth factor price and dosage separately, supplemental protein costs, process mode probability mix, and the probability of specific technical breakthroughs. These correspond to parameters in the cost model — but please estimate from your own background knowledge, not by looking up what the model currently assumes for each slider. Your independent assessment is what's most informative; the model's defaults are our starting point, not the answer we're looking for. [if you use the model]If you do explore the model's sliders as a reference, that's fine — but please note it in your reasoning and explain your own independent view. We'll compare expert-elicited distributions against model defaults in the synthesis.

▸ Show/hide detailed technical questions (E2–E7 — particularly valuable if you have direct bioprocess or TEA expertise)
E2 · Growth factor price ($/g) — dominant GF (e.g. FGF-2), commercial-scale 2036

Price per gram of the most-used recombinant growth factor. This is price only — quantity (g/kg cell mass) is asked separately below. Model mapping: GF Price ($/g) slider.

This is your 2036 forecast — if you expect a breakthrough in GF production to have occurred by then, your estimate should reflect that.

E3 · GF effective dosage (g of GF per kg of cell biomass produced), 2036

Total grams of recombinant growth factors consumed per kg of cell biomass produced (accounting for recycling, thermostable variants, and media use rate). Model mapping: GF Quantity (g/kg) parameter.

Autocrine cell lines (cells produce their own GFs) could approach 0; conventional recombinant GF use depends on recycling, thermostable variants, and media use rate.

E4 · Supplemental protein cost contribution ($/kg cell mass), 2036

Total cost of albumin + transferrin + insulin as media additives, per kg of cell biomass produced. Model mapping: Supplemental Proteins cost component.

E5 · Process mode probability mix — your 2036 expectation

What share of large-scale commercial CM production (by volume) do you expect each bioreactor process mode to account for in 2036? Should sum to 100%. Fed-batch: current default, lower capex. Perfusion: higher density, continuous media exchange. Continuous flow: frontier, highest potential throughput.

Sum: — (should equal 100%)

E6 · Probability that each approach substantially decreases GF costs by 2036

For each approach, what is your probability that it will have succeeded in substantially reducing effective growth factor costs in commercial CM production by 2036?

E. coli / yeast produce GFs at scale — target $10–100/g (GFI 2024)
—
Transgenic plants express GFs — target $1–10/g (BioBetter)
—
Engineer cells to produce their own GFs — target ~$0/g effective cost (Stout et al. 2023)
—
Chemicals that activate GF receptors — target <$1/g (Ahmad et al. 2023)
—
e.g. FGF2-G3 with 20-day half-life — substantially reduces effective dosage (Enantis)
—
E7 · Feed conversion efficiency — discussion (tentative — we don't model this yet; your input shapes whether and how we should)

At large scale, the dominant cost driver may shift from formulation price to feed conversion efficiency — how efficiently cells convert glucose and amino acids into biomass. This can be expressed as kcal of biomass per kcal of nutrient input (kcal:kcal), or g dry-weight biomass per g dry-weight nutrients consumed (DW:DW). Current TEAs largely ignore this, focusing instead on $/L of medium and L/kg of output. If feed conversion is poor, cheap media formulations may still be expensive per kg of output. We're considering whether to model this explicitly. Do you have views on: (a) expected feed conversion ratios for chicken cells at commercial scale; (b) whether current TEAs adequately account for this; or (c) how important this is relative to the cost drivers we already model?

Thank you — these inputs directly update our model priors. We'll report expert-elicited vs. model-default distributions in our workshop synthesis.

About You

You can choose whether to keep your response anonymous (leave blank) or allow others to see your name alongside your response. Response content will be shared within the group and published publicly in either case. A pseudonym (see below) lets us link your pre- and post-workshop responses without identifying you publicly.

Helps us report beliefs by subgroup — do TEA specialists and industry operators agree? We'll report subgroup distributions alongside pooled results.

A note on more rigorous elicitation

This form is a lightweight structured survey — not a formal expert elicitation protocol. A more rigorous process (calibration training, full distribution elicitation, subgroup comparison) would be methodologically superior — though it would likely require substantial compensation for experts' time. If this is of interest, please let us know below.

Your responses are stored securely and will be used to inform the synthesis report.

— saves a text file to your device before or after submitting.

▸ Form revision history — what has changed since first published, and why

These changes do not affect submissions already received — earlier responses are stored as entered. If a field was blocked by any of the issues below, please re-submit to add your complete response.

May 11, 2026

The volumetric productivity sub-question (CM_16) had a ceiling too low to enter any realistic value. All three fields — pessimistic, median, and optimistic — were capped at 5 kg/m³/day by the form. But a cell density of 75 g/L with a daily 50% harvest already implies around 37.5 kg/m³/day, so even a moderate estimate was blocked. The ceiling has been raised to 200 kg/m³/day, which covers all plausible scenarios. [max: 5 → 200; step: 0.001 → 0.1 on cm16_productivity_lo/med/hi]

Upper limits on growth factor dosage (E3) and supplemental protein cost (E4) were too tight. E3 was capped at 1 g/kg and E4 at $20/kg — both plausible enough for optimistic estimates, but they could block an expert with a more pessimistic view. Both limits have been raised to give more room. [E3 max: 1 → 10 g/kg; E4 max: 20 → 200 $/kg]

The growth factor price field (E2) only accepted whole-dollar-per-gram amounts. Entering $2.50/g or $0.50/g would cause the form to refuse to submit — the browser requires values to match the field's step size, which was set to $1. Given that 2036 commercial-scale GF prices could plausibly be fractions of a dollar per gram, this was a meaningful block. Fixed to two decimal places. The CM_13 median field had the same issue (its own CI fields already accepted tenths but the median didn't). [E2 step: 1 → 0.01; CM_13 median step: 1 → 0.1]

Most confidence interval fields were not checking whether the numbers were in the right order. If you accidentally typed a larger number in the lower-bound field than in the median field, the form would accept it silently. Live warnings now appear as you type whenever a lower bound, median, or upper bound set is out of order. This applies to CM_01 (2027 version), CM_13, CM_14, CM_16 cell density, CM_16 volumetric productivity, CM_20, and the E2/E3/E4 expert questions. The CM_01 (2036) and CM_02 fields already had this check from the start. [makeOrderingValidator() added in beliefs.js]

Selected questions from the Cultivated Meat PQ formulations (CM_01, CM_02, CM_10, CM_12–14, CM_16–17, CM_20). Last updated: May 11, 2026.

Submit Workshop Availability → About This Workshop → Resources & Readings → CM_01 on Metaculus →