The climate impact and land use of cultivated meat: Evaluating agricultural feedstock production

RESEARCH ARTICLE The climate impact and land use of cultivated meat: Evaluating agricultural feedstock production Hanno KossmannID*, Thorsten Moess, Peter Breunig Faculty of Agriculture, Food and Nutrition, Weihenstephan-Triesdorf University of Applied Sciences, Weidenbach, Germany * hanno.kossmann@hswt.de Abstract As global demand for meat continues to rise, alternative and sustainable methods of produc- tion are being explored. Cultivated meat (CM) is one such alternative that holds potential for sustainable production with less environmental impact. This study develops an approach to evaluate CM production based on agricultural feedstock. The specific objectives are to determine the minimum land area required to produce a certain amount of cell medium– feedstock for CM production–on agricultural land and to identify potential future land use scenarios assuming that the macro components of the cell medium are solely produced from common agricultural crops in southern Germany. A linear programming model was developed to analyze four different scenarios of CM production, considering factors such as crop rotation, nutrient sourcing, and solar energy use. The results indicate that CM produc- tion using plants as raw material for the cell medium cannot improve land use efficiency sub- stantially compared to conventional pork production. Extraction methods, crop choice, and energy sources will strongly influence future pathways for CM. We also find that there is no substantial benefit from CM in terms of climate change mitigation when feedstock is sourced

solely from plants. This study provides valuable insights into the limitations of using agricul- tural feedstock for sustainable CM production. The findings suggest that future research should focus on optimizing the land use efficiency of CM. This includes exploring alterna- tives such as sourcing cell media from precision fermentation instead of relying solely on crops, and utilizing upcycling possibilities. 1 Introduction Human diets have prominently incorporated meat for millennia, serving as a cornerstone for nutritional needs and livelihoods throughout the world [1–3]. Despite declining meat con- sumption in high-income countries, forecasts indicate a ballooning demand for meat in forth- coming decades due to global population surges and increasing incomes [4, 5]. Plant-based foods are not always available or affordable and livestock is an important source of income to poor households [6]. Globally meat is instrumental for food security and livelihoods [7, 8]. PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 1 / 27 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Kossmann H, Moess T, Breunig P (2025) The climate impact and land use of cultivated meat: Evaluating agricultural feedstock production. PLoS ONE 20(1): e0316480. https://doi.org/10.1371/ journal.pone.0316480 Editor: Anto´nio Raposo, Lusofona University of Humanities and Technologies: Universidade Lusofona de Humanidades e Tecnologias, PORTUGAL Received: October 3, 2024 Accepted: December 11, 2024 Published: January 22, 2025

Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0316480 Copyright: © 2025 Kossmann et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the manuscript.

But the consumption of protein-rich foods is marked by their substantial land and resource demands [9–12]. In addition to complex concerns surrounding animal welfare and ethical considerations [13–15], the resource utilization of animal-sourced foods has become a focal point in both everyday discourse and academic research. Increasing consumption rates in lower-income countries and the Global South also lead to dependencies and geographically detached environmental influences due to imported animal feed [16, 17]. Pork is the most consumed meat globally, and its production alone covers more than 5% of all croplands worldwide. The feed conversion ratio and environmental impact of pig farming are highly dependent on effective management practices and the quality of feed used. Profi- cient management can significantly enhance the sustainability of pig farming operations [18]. Based on feed conversion ratios, pork production efficiency lies between that of chicken and beef [19, 20]. Life cycle assessments (LCAs) accentuate the extensive environmental footprint of pork production, encompassing factors like greenhouse gas emissions, land and energy use [21]. Feed production and mixture, which influence land use, as well as waste management, which affect methane emissions and nitrate pollution, are the most critical factors impacting the environmental footprint of pork production [22]. These environmental concerns, also linked to meat production in general, have amplified the quest for sustainable alternatives, especially those with a rich protein content [23–25].

A subset of these innovations are animal proteins made from cell cultivation, which directly mimic traditional animal products in flavor, appearance, and composition. Cultivated meat (CM) represents one of those technologies [26–28]. Using in-vitro techniques, animal cells are cultivated in a controlled setting, obviating animal rearing, slaughter, and thereby animal wel- fare issues [29–31]. Notably, the CM domain witnessed significant strides with the unveiling of the first CM burger in 2013 [32]. Consumers are generally open to eating CM [33, 34]. Especially ethical advantages in terms of environmental sustainability and animal welfare are influencing consumers’ willingness to try and eat CM [35]. On the other hand, perceived risks and unnaturalness make consumers critical [36]. Existing LCAs on CM present a spectrum of conclusions, from overtly skeptical [37–39] to decidedly optimistic [40–42]. While CM emanates promise as a greener substitute for tradi- tional meat, the sector remains nascent with pronounced information voids [23, 30, 43]. Par- ticularly, there is a dearth of clarity on how emerging countries, where meat demand is rising, can develop the necessary capabilities to integrate CM technology, as most startups are based in developed countries [44]. Given the elevated demand for meat–and possibly CM–substantial input commodities will be required for the nourishment of CM cells. Consequently, agricultural production could play a pivotal role in this supply chain [45]. Central to CM’s efficacy is cell media, which is a

composite vital for in-vitro cellular growth [41, 46]. The cost-effective synthesis of cell media is pivotal for ensuring efficiency in resource and monetary terms. The cost of cell media which is 55%–95% of CM’s total cost, is mainly caused by cytokines and transferrine production [47, 48]. The cause of these high costs is mainly the lack of industry scaling [49]. Current research posits agricultural crops, predominantly maize and soy, as feasible crop-based nutrient sources for CM’s media [40, 41, 50]. But some parts of the feedstock are also added by precision fer- mentation microbial or chemical production in the case of Sinke et al., 2023 [41] for instance 25%, while 75% from soy hydrolysate. The production of cell culture medium ingredients has high influence on the environmental impact of CM [42]. Contrary to Sinke et al. (2023) [41], Tuomisto et al. (2022) [42] further recommend exploring amino acid production using hydro- lysates from legumes among other things. PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 2 / 27 Funding: Funding for the authors’ research position, which supported the completion of this study, was provided by the Bavarian State Ministry of Science and the Arts as part of the High-Tech Agenda Bavaria. This article was additionally funded by the Open Access Publication Fund of Weihenstephan-Triesdorf University of Applied Sciences. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Besides glucose and amino acids as main components of the cell media, electrical energy is another major input in CM production, since there is no complete organism regulating cell temperature like in animals [41]. So far, there is no research that investigates optimization opportunities in land use for pro- ducing the macronutrients (glucose and amino acids) and renewable electricity required to enable CM production. This study evaluates the use of agricultural feedstock and land use for cultivated pork pro- duction in terms of macro nutrients and energy demand. We use linear programming to opti- mize CM output per land area based on the cell nutrient and energy demand as well as crop yield and crop rotation restrictions. The optimized land use demand for CM is compared to land use requirements of pork production. In addition, we quantify the climate impact of land use through carbon opportunity cost. To apply this approach to a specific region we use yields and crop rotation restrictions of southern Germany. This region has the environmental conditions to grow a large variety of crops with relatively high yields and robust data. This paper unfolds as follows: Section 2 delves into the background of CM, charting its technological background, environmental implications, land use, economic viability, and impact on agriculture. Section 3 elucidates our methodology—encompassing data acquisition, analytical techniques, and the structure of linear optimization. Section 4 showcases our results,

which we subsequently discuss. The paper culminates with reflections on our findings and an outline for future research trajectories. 2. Background of cultivated meat Technical and environmental aspects of CM production, innovative approaches, regulatory challenges, and the importance of transparent and accurate information for consumer accep- tance will be described in this section. CM research is in its early stages, focusing on promising narratives, understanding consumer behavior, and exploring technical and regulatory feasibil- ity [23, 32, 36, 47]. CM provides an alternative to conventional animal farming, potentially leading to a more sustainable and ethical meat production system by eliminating the need for animal slaughter and rearing. CM’s market introduction could significantly impact monogas- tric animal production by producing fewer by-products and not using land unsuitable for human food production [29]. Around the world, many startups have ventured into the CM industry, securing significant funding [51]. However, for successful market introduction, it is crucial to address production techniques, scaling challenges, environmental concerns, regulatory hurdles, and consumer acceptance [23, 25, 41, 52, 53]. Several technical challenges hinder CM from becoming a viable alternative to conventional meat. A primary concern is the industry’s reliance on fetal bovine serum for cell culture, which is ethically contentious and not scalable. Efforts are underway to develop serum-free alternatives [32, 54]. Criticism states that Moore’s law, speaking of an

almost linear cost to doubling output relation, is not applicable to biological systems [12]. Another challenge is replicating the texture and flavor of traditional meat, as current CM pro- duction results in disorganized muscle fibers that lack the complex structure of conventional meat cuts [55]. Additionally, assessing the nutritional quality of CM, particularly its micronu- trient composition and overall health implications, requires further research and technological advancement [54, 55]. From a regulatory standpoint, CM is being managed differently across the globe. Singapore leads with early approvals and frequent updates. China and India are in the process of develop- ing their frameworks. Japan, South Korea, Australia-New Zealand, and the EU regulate under PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 3 / 27

existing novel food guidelines. Post-Brexit, the UK is creating specific regulations, and the US is coordinating between the FDA and USDA. Canada and Brazil are also progressing in their regulatory approaches [56]. 2.1 Technical background of CM production CM production encompasses various stages: cell isolation, cell expansion, differentiation, bio- reactor utilization, and harvesting. Initiated by extracting muscle or fat cells from animal tis- sues, these cells are nurtured in a nutrient-dense culture medium. Following sufficient proliferation, these cells undergo differentiation into muscle cells. Subsequently, this tissue is matured in a bioreactor until it reaches an optimal state for harvesting and processing into meat products. Preliminary prototypes spanning beef, pork, and poultry highlight the potential of this approach [37, 50, 56]. Culture media, instrumental in in vitro meat production, present economic and technical challenges. Current formulations, although general, necessitate tailoring for specific cellular metabolic states. Although fetal bovine serum, a meat industry byproduct, is commonly used in cell culture media, serum-free alternatives are under investigation to alleviate associated eth- ical concerns [47, 57]. Moreover, innovations are imperative in cellular modulation. These could be reached by physical and chemical conditioning, optimized substrate elasticity, and nutrient diffusion in expansive tissues [29, 32, 58]. Formulating cell media entails sourcing from diverse origins and processes. Core compo-

nents include glucose, amino acids, vitamins, and trace elements crucial for optimized cell growth and yields. O’Neill (2021) [46] suggests incorporating renewable feedstocks like ligno- cellulosic hydrolysates into media formulations for environmental sustainability. Vergeer’s & Risner’s LCAs emphasizes the role of eco-friendly practices in media formulation, touching on issues such as the environmental ramifications of antibiotics and the strain from glucose pro- duction [58, 59]. Research supports cell media derivation from plants like soy and maize for sugars and amino acid extraction, alongside utilization of precision fermentation or biotech- nological processes [40, 41, 50, 60–62]. This study accentuates plant-based media composition, considering its implications on agriculture and land use. Cell metabolism encapsulates the myriad chemical reactions essential for cellular suste- nance and growth, including nutrient breakdown, molecular synthesis, and waste elimination. The vital components for these reactions range from glucose, amino acids, and vitamins to oxygen and growth factors like hormones. To evaluate the demand of cells for efficient growth of CM, we assumed the following metabolism reaction taken from Humbird (2021) [37]: 0:147 Glc þ 0:378 O2 þ 0:007 Arg þ 0:004 Cys þ 0:022 Gln þ 0:003 His þ 0:007 Ile þ 0:010 Lys þ 0:002 Met þ 0:005 Phe þ 0:009 Thr þ 0:002 Trp þ 0:005 Tyr þ 0:010 Val þ 0:013 Ala þ 0:006 Asn þ 0:008 Asp þ 0:011 Gly þ 0:011 Leu þ 0:007 Pro þ 0:010 Ser ! ð1Þ DCMa þ 0:005 Glu þ 0:004 NH3 þ 0:041 Lac þ 0:455 CO2 þ 0:613 H2O

In the fed-batch process as assumed in Humbird (2021) [37], cells are cultivated in a nutri- ent-rich medium, replenished periodically. Contrary to perfusion cultures, where nutrients are provided and waste is removed continuously, with the fed-batch process the medium is depleted as cells reach certain densities. The fed-batch technique is cheaper, but cell growth may be limited to a certain extent [37, 43]. For CM to be widely accessible, scalable production is vital. Humbird (2021) [37] expounds on the challenges, emphasizing bioreactor issues during fed-batch operations. Compared to PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 4 / 27

microorganisms, scaling is complex due to slow animal cell growth rates combined with shear force-induced cell damages, metabolic inhibition, and aseptic maintenance challenges. Perfu- sion technology offers a potential solution, although it limits bioreactor sizes. Elevated hygiene demands, combined with the need for richer nutrient mediums for animal cells, especially stem cells, further amplify the challenges [37]. 2.2 Environmental implications and challenges of CM LCAs Finally, the environmental impact of CM will be crucial to its success [31]. Evaluating the envi- ronmental impact of food products requires examining land use efficiency [63] and under- standing the carbon opportunity costs associated with land use [64]. In this study, we employ land use as a metric to explore the potential environmental advan- tages of CM. While a singular focus on land use does not fully encompass all environmental ramifications of meat production, it remains an informative, albeit partial, metric for this anal- ysis due to two main reasons. First, land use encompasses a significant majority of the environ- mental costs associated with pork production, especially soy cultivation for feed production [65–67]. Second, land use serves as a potent indicator of the biodiversity impacts tied to vari- ous agricultural products [68]. LCAs for CM have been conducted to evaluate the environmental impacts of this technol- ogy [38, 41, 42, 43, 50, 59, 62, 69]. These assessments typically cover the entire product life-

Several LCAs examine different methodologies for producing cell media for CM, emphasiz- ing various biological pathways. Some focus on using crops as the base material for cell media [41, 42, 50], while others explore using bacteria or alternative microorganisms [61, 70]. Conse- quently, the specific composition and production processes of cell media significantly affect resource utilization in CM production. Recent LCAs report that the land use of CM is consistently lower than that of conventional meat, with values ranging from 0.4 to 6.9 m2/kg meat [41, 42, 43, 50, 61]. This variation is due to differences in raw material selection, databases, and case study locations. This study uses a specific example that accounts for crop rotation restrictions and local yields. Energy use and culture medium directly impact land use, considering feed and sustainable energy production. Most LCAs use agricultural modeling databases to calculate land use for CM. However, incor- porating agricultural realities and modeling crop rotation effects pose challenges [42, 43]. We PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 5 / 27

employed a case study approach to compute land use and carbon opportunity costs for CM production in southern Germany, using on-site generated feed, local yields and nutrient contents. 2.3. Economic viability and impacts on the meat industry and farming communities The economic viability of cultured meat (CM) is critical for its scalability and market adoption. Advances in large-scale CM production are closely tied to developments in the biomedical field, particularly in tissue engineering. Key innovations include the development of special- ized cell lines, plant-based cell media, advanced scaffolding materials, and efficient bioreactors. These innovations are foundational to the growth and competitiveness of the CM sector, and their ongoing development is essential for reducing production costs and improving product quality [71, 72]. Projected production costs vary significantly depending on the process, rang- ing from estimates of $5.50/kg [40] to $25/kg [37]. However, numerous assumptions make these values approximate and subject to considerable uncertainty. These scenarios require much of the necessary infrastructure that does not yet exist. Furthermore, the successful inte- gration of CM into the global value chain requires not only technological advancements but also strong business structuring, market positioning capabilities, and robust relationships with stakeholders. Addressing these multifaceted challenges and opportunities is crucial for the widespread adoption of CM, particularly in emerging countries where supportive policies and

infrastructure can play a pivotal role in fostering industry growth [73]. Historically, the meat industry and farming communities have been central to the global food system, impacting cultural, economic, and societal frameworks. Their contribution tran- scends direct agricultural activities, providing economic stability to many communities [19]. The emergence of CM could redefine these traditional roles, potentially diminishing the cen- trality of conventional livestock farming as technological advancements in CM become more pronounced [32, 71]. However, the integration of CM into existing agricultural frameworks might offer new opportunities. For instance, CM production could become decentralized, with potential for implementation at the farm level, thereby including traditional farmers in the emerging CM industry [73]. Also new and higher-skilled jobs are expected to be created [74]. The economic landscape may shift as CM becomes more established. Traditional meat pro- ducers may face challenges in maintaining revenue [74], whereas CM offers new economic avenues, particularly for those willing to diversify into this new area [75]. For example, farmers could utilize their expertise to contribute to the CM supply chain, such as in the production of cell media, demonstrating the adaptability and resilience of farming communities [45, 74]. Despite these opportunities, there are barriers to the widespread adoption of CM. Many farming communities, steeped in generational traditions, may view CM with skepticism or

resistance, questioning its authenticity and the impact on traditional farming practices. These observations could also be made for veterinarians and animal scientists in Brazil [74]. Integra- tion with CM could require significant investments in new equipment, knowledge, and infra- structure tailored to this innovative industry [30, 76]. While private funds have been used most recently, scientists and other interest groups are calling for additional public funding [49]. Morais-da-Silva et al. (2022) [74] in turn saw from interviews of alternative protein experts more positive reflections and opportunities outlined by the experts. The experts saw the poten- tial to create new, higher-skilled jobs, particularly in technical and management fields. How- ever, the transition may disproportionately affect animal farmers, especially in a rapid change PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 6 / 27

scenario, with varying levels of optimism across regions, most notably higher in Brazil than in Europe. Furthermore, understanding the future role of agriculture in CM production necessitates a comprehensive analysis of resource utilization and the sustainability of CM compared to tradi- tional meat. This includes evaluating the land use requirements of CM to ascertain its environ- mental benefits and value-added potential in the agricultural sector. 3. Methods and scenarios 3.1 Modeling structure Our model uses linear programming to maximize CM production within a 100-hectare refer- ence area, incorporating crop rotation and nutrient constraints specific to southern Germany. Linear programming is a mathematical strategy frequently used across various disciplines for optimization purposes [77]. In agriculture, it is often employed to optimize resource utiliza- tion, encompassing factors such as land and crop rotation optimization [78–81]. In the context of our research, we employed a linear programming model with several spe- cific components in mind. Primarily, we defined an objective function that represents our optimization goal: maximizing the production of CM per 100 hectares of land. In the modeling structure, the 100 hectares serve as a land area reference scale for comparison. The results are expressed in kilograms of meat per square meter. Our model also incorporates constraints reflecting the crop restrictions specific to southern Germany. We focus on southern Germany since this region has environmental conditions that allow the production of a wide variety of

crops with relatively high yields (Table 1). These constraints include crop rotation practices, solar energy use, and meeting CM nutrient requirements. Although future methods may pro- duce some amino acids, we focused on a closed, plant-based system. By integrating these com- ponents, the linear programming model utilized in this study is designed to identify the optimal land use strategies. The ultimate objective is to maximize the production of CM within the confines of limited land availability. Hence, the model offers valuable insights into efficient land practices and optimal crop mixtures for achieving the highest possible output of CM. Table 1. Average crop yields in southern Germany, Germany, the world, and continents in 2021. Crop Southern Germany (Bavaria) (t/ha)a Germany (t/ha)b Africa (t/ha)b Americas (t/ha)b Asia (t/ha)b Europe (t/ha)b World (t/ha)b Barley 6.75 6.76 1.77 2.87 1.59 3.96 2.97 Lupine 2.88 1.84 0.70 1.61 2.25 1.58 1.40 Maize 10.52 10.36 2.27 7.80 5.58 7.20 5.87 Oats 4.45 4.32 1.30 2.38 2.03 2.52 2.36 Peas (general) 2.99 3.06 1.19 1.47 1.45 2.32 1.76 Potatoes 42.26 43.79 15.02 29.44 19.15 23.62 20.74 Rapeseed 3.57 3.50 1.89 1.53 1.72 2.81 1.93 Rye 5.52 5.27 1.87 2.43 2.29 3.23 3.05 Soybean 3.17 3.11 1.44 3.27 1.43 2.09 2.86 Sunflower 2.70 2.60 1.05 1.92 2.02 2.06 1.97 Triticale 4.88 5.81 2.30 2.80 2.06 4.12 3.87 Wheat 7.54 7.30 3.03 2.83 3.39 4.28 3.49 aLfL, 2023 [82] bFAOSTAT, 2023 [83] Note: t/ha = tons per hectare.

Crop rotation restrictions are detailed in Table 2. For grain legumes, total acreage cannot exceed 20% of the farm area, with specific caps for individual crops: lupine, field bean, and pea are each limited to 15% [84]. https://doi.org/10.1371/journal.pone.0316480.t001 PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 7 / 27

The objective to be maximized is the expansion of the activity fresh mass production (referred to as the target cell). To achieve this goal, a solver in the model is tasked with adjust- ing the expansion of all activities, including that of fresh mass production. This adjustment must be carried out while adhering to certain constraints. These constraints include the full utilization of land area capacity, crop rotation restrictions (mentioned below), and the required nutrient supply for agricultural crops together with the corresponding expansion of production procedures. 3.2 Model specification This section specifies the objective and constraints of the linear programming model to opti- mize nutrient and land use for CM production. We developed a static linear programming model to evaluate the agricultural feedstock production for CM in southern Germany, apply- ing it to a 100-hectare land area reference. The model aims to maximize CM output while adhering to agricultural constraints like yields and crop rotation restrictions. The focus is on using 100 hectares to produce the macro nutrients (glucose and amino acids) and energy required for CM production. The model’s objective function and constraints are formulated as detailed below: Objective Function: Maximize M ¼ pT x ð2Þ Where: • M is the total production of CM medium (targeting maximization). • x is the vector of decision variables, representing the activities including different crops and crop rotation, as well as the potential use of solar panels for energy production.

• p is the vector of production quantities per unit of activity, namely the quantities of glucose and single amino acids produced per unit of crop or solar energy generation activity. • T denotes the transpose of the vector p. Constraints: Nutrient Recipe Constraints: These constraints ensure that the mix of inputs used matches the specific nutritional requirements for CM production: X jaij � xj � yj � bj � M ð3Þ where: • aij represents the amount of amino acids and glucose content i provided by crop j. • xj is the amount land use for a specific crop or for energy production j. • yj is the yield per ha of crop or energy production • bi is the required amount of input i per unit of CM medium, defining the feedstock recipe for CM. The goal of the model is to find the optimal combination of activities x that maximizes the total production of CM medium M, subject to the constraints of available resources, which is land and crop rotation restrictions. Crop Rotation and Land Use Constraints: PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 8 / 27

Account for crop rotation and land use practices: xj � L � CRj ð4Þ where: • xj is the land used for crop j. • L is the total available land. • CRj is the crop rotation factor for crop j, representing the maximum proportion of land that can be used for this crop to ensure sustainable rotation. Account for total land use constraint: X j2Cropsxj � L ð5Þ The model aims to identify the optimal combination of activities x that maximizes the total production of CM medium M, while complying with the available resources and crop rotation restrictions. The model’s outcome serves to calculate and compare the area of land (in square meters) required per kilogram of produced CM compared to pork meat, or conversely, the yield of CM compared to pork meat (in kilograms) per square meter of land. This provides a comparable metric for assessing the efficiency and sustainability of the CM production process in southern Germany as the case study region. We used Microsoft Excel’s Solver Add-In for linear programming. Solver’s optimization function balanced CM production nutrient requirements with crop area allocation, consider- ing crop rotation constraints. This linear programming approach provided insights into which crops were most limiting and which nutrients were in surplus or deficit. 3.3 Technical inventory and background: Linear programming for CM Our study applies strict crop rotation restrictions to mirror realistic farming practices and ensure sustainable soil management in southern Germany. Regardless of farm structure or

land ownership, these restrictions limit the share of each crop to maintain long-term yield lev- els. We examined protein plants (soybean, lupine, field bean, pea, winter rapeseed, and sun- flower) and starch plants (grain maize, winter wheat, rye, triticale, winter barley, oat, and potato). Table 1 provides crop yield statistics, showing that southern Germany’s yields are above the world average [83]. Soybeans are traditionally grown in warm regions, but advances in plant breeding and agri- cultural technology have made it possible to grow adapted varieties in more northern regions. In southern Germany, soybean farming increases crop diversity and reduces reliance on imported soybeans for livestock feed [85]. Challenges such as pest and disease risks and limited domestic market demand must be considered. We applied a soybean restriction to a maximum of 50% of the land area. A two- year maize-soy rotation system, proven feasible in Canada and the US Midwest [86–88], is assumed for southern Germany. Summer crops are limited to no more than two-thirds of the land area. Grain maize is capped at 50%, while sunflower and potato each have a 20% cap. Winter crops also follow the two-thirds limit for the total farm area, with specific restrictions: wheat at 33.33%, rye at 50%, triticale at 33.33%, winter barley at 33.33%, oats at 25%, and rapeseed at 20%. These rotation limits are based on data from Diercks & Heitefuss (1990) [89], Kolbe (2021) [84], and Reckling et al. (2016) [90] (Table 2). PLOS ONE Evaluating agricultural feedstock for cultivated meat

PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 9 / 27

Our study uses data on agricultural crops in Germany and feedstuff analyses to determine amino acid patterns and glucose contents [91, 92]. The INRAE-CIRAD-AFZ Feed Tables—a comprehensive reference for feed composition and nutritional values developed by French research institutions INRAE, CIRAD, and AFZ—are used for glucose quantity specifications [93]. Glucose content is derived from starch and sugar, however, the database does not provide detailed information on the monosaccharide composition of these fractions. For energy and oil crops, starch was considered primarily as D-glucose polymers, allowing near-complete conversion to glucose during hydrolysis. For grain legumes, only half of the starch and sugar fraction was con- sidered convertible to glucose due to non-starch polysaccharides and oligosaccharides. Based on Humbird (2021) [37], an overall amino acid yield of 70% after extraction from crops was assumed. Due to feed analyses capturing glutamine and asparagine with glutamic acid and aspartic acid, these amino acids were considered collectively (Table 3). To quantify nutrient delivery, we used average yields from 2017–2021 calculated by the Bavarian State Research Center for Agriculture [82]. Nutrient yield per hectare was calculated by multiplying crop nutrient contents (%) with per-hectare yields (t FM/ha), considering dry matter contents of the feedstuffs [83, 94]. The energy demand for CM production was sourced from Vergeer et al. (2021) [40], Mat- tick et al. (2015) [50], and Tuomisto et al. (2014) [95]. Information on solar panel efficiency

came from Bo¨hm et al. (2022) [96]. Crop productivity data was obtained from the Bavarian State Research Center for Agriculture (LfL, 2023) [82]. Finally, we assume that all raw materials, including soy, are produced on-site by the farm, which is set at 100 hectares. The linear programming model then balances the nutrient requirement (kg/kg CM) with the nutrient delivery (kg/ha) for CM production. A specific correction is applied that takes into account that the LfL data relates to 14% or 9% moisture (equivalent to 91% or 86% dry matter). This results in an adjustment of hectare yields based on the dry matter content of each analyzed feed material. Here is an example, in the case of soybeans. Dry matter content per analysis: Yield of 3.17 t/ha at 14% and dry matter following feed analysis of 89.5% results in a calcula- tion of 3.17 t/ha x 0.86/0.895, or approximately 3.05 t/ha. This corrected yield per hectare is then multiplied by nutrient content percentages, the nutri- ent extraction factor, and a conversion factor of 1,000 kg/t. This calculation gives the nutrient delivery in pure nutrient kilograms per hectare for each crop type. Table 3 presents the nutrient requirements for CM. The nutrient delivery is directly correlated with the nutrient requirement, expressed in kg/kg of CM, ensuring a consistent comparison of values in the analysis. Table 2. Crop rotation restrictions assumed in the linear programming model. Crop category Category limit (%) Individual crop restrictions (%) Grain legumes � 20% (without soy)

Soybean � 50%, lupine � 15%, field bean � 15%, pea � 15% � 50% (with soy) Summer crops � 67.77% Grain maize � 50%, sunflower � 20%, potato � 20% Winter crops � 67.77% Wheat � 33.33%, rye � 50%, triticale � 33.33%, winter barley � 33.33%, oats � 25%, rapeseed � 20% Foliage crops � 25% Soybean, lupine, field bean, forage pea, rapeseed, sunflower Rapeseed + Sunflower � 33.33% Rapeseed, sunflower Cereals � 67.77% Wheat, rye, triticale, barley, oats https://doi.org/10.1371/journal.pone.0316480.t002 PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 10 / 27

For energy and oil plants, it is presumed that starch is present as amylopectin (approxi- mately 70–80%) and amylose (20–30%), both of which are polymers of the monosaccharide D-Glucose (primarily α-1,4-glycosidic bonds). It is approximated that starch or disaccharides can be broken down completely and utilized as glucose. This assumption is justified by the compensatory effect between any decomposition losses and the mass increase of the glucose fraction caused by the addition of H2O (H+ and OH- adaption) during hydrolysis [37]. With grain legumes, the case arises where some non-starch polysaccharides or oligosaccha- rides are analytically captured in the starch fraction, which, besides glucose, have up to 50% other monosaccharides (Mannose, galactose, and fructose) as basic building blocks. There are also considerable varietal differences [91, 92]. Therefore, it’s roughly assumed that for these plants, 50% of the analytical sugar and starch fraction can be present as, or can be converted to glucose. Humbird (2021) [37] presumes a protein solubility of 88% and a hydrolysis conver- sion of 80% for hydrolysate from soybean meal, which corresponds to an overall amino acid yield of 70.4%. As stated before, an amino acid yield of 70% is assumed for all considered agri- cultural crops in this study. The data on each crop’s relative nutrient supply, as detailed in Table 4, serves as a founda- tional parameter for our calculations and is thus essential for determining the nutrient yield per crop. Our analysis distinctly categorized crops into energy and protein plants, which is a

differentiation that proved to be critical in the outcomes of our study. As stated before, the yield by crop together (as shown in Table 5) with the nutrient supply (as shown in Table 4) led to the nutrient supply by crop in kilogram per hectare (as shown in Table 6). Table 3. Relative nutrient requirements for cell mass production. g/g CM FM stoichiometric demanda Glucose 0.362 Arginine 0.016 Cysteine 0.005 Glutamine 0.044 Histidine 0.006 Isoleucine 0.012 Lysine 0.020 Mehionin 0.004 Phenylalanin 0.011 Threonine 0.014 Tryptophan 0.004 Tyrosine 0.012 Valine 0.016 Alanine 0.016 Asparagine 0.011 Aspartic acid 0.014 Glycine 0.011 Leucine 0.019 Proline 0.010 Serine 0.015 Glutamic acid + Glutamine 0.044 Asparagine + Aspartic acid 0.025 aHumbird, (2021) [37]. Note: g/g CM = gram per gram cultivated meat; FM = fresh mass. https://doi.org/10.1371/journal.pone.0316480.t003 PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 11 / 27

3.4 Comparative energy demands Energy requirements of CM production are compared with traditional meat, assessing the role of solar energy in reducing environmental impacts. Energy consumption forms a significant aspect of our study, given the high energy requirements of CM production, particularly the energy needed to maintain bioreactors at a consistent temperature of 37˚C [39]. The energy Table 4. Relative nutrient supply by crops considered in linear optimization. Protein and oil plants Energy plants Nutrient yield with hydrolysis Soybean Lupine Field bean Pea Rapeseed Sun Maize Wheat Rye Triticale Barley Oats Potato Glucose from starch (Legume) 50% Glucose from starch (General) 100% Amino acid yield (General) 70% Dry matter 89.5% 88.1% 86.0% 87.2% 92.4% 92.8% 86.3% 86.9% 86.7% 86.8% 87.2% 87.6% 89.1% Grams per 100 grams of fresh mass % % % % % % % % % % % % % Nutrients Saccharides Starch 5.3 6.2 37.1 44.7 3.5 1.2 63.8 60 53.7 58.8 52.3 36.8 65.2 Sugar 7.5 6 3.5 4.2 5.4 2.5 1.7 2.6 3.1 3 2.2 1.3 4.7 Starch + Sugar 12.8 12.2 40.6 48.9 8.9 3.7 65.5 62.6 56.8 61.8 54.5 38.1 69.9 Glucose 6.40 6.10 20.30 24.45 8.90 3.70 65.50 62.60 56.80 61.80 54.50 38.10 69.90 Amino acids Essential Arginine 2.63 3.6 2.61 1.73 1.1 1.07 0.37 0.55 0.43 0.53 0.48 0.63 0.31 Cysteine 0.53 0.55 0.34 0.28 0.47 0.26 0.19 0.27 0.2 0.26 0.23 0.31 0.12 Histidine 0.98 0.73 0.68 0.51 0.47 0.37 0.22 0.25 0.18 0.24 0.23 0.2 0.14

Isoleucine 1.68 1.55 1.09 0.85 0.77 0.59 0.28 0.39 0.3 0.38 0.36 0.36 0.28 Lysine 2.24 1.63 1.72 1.48 1.15 0.6 0.23 0.32 0.34 0.39 0.37 0.4 0.41 Mehionin 0.52 0.26 0.19 0.2 0.41 0.34 0.16 0.18 0.14 0.17 0.17 0.17 0.12 Phenylalanin 1.82 1.3 1.12 0.96 0.72 0.63 0.37 0.52 0.38 0.43 0.48 0.47 0.33 Threonine 1.46 1.24 0.95 0.78 0.89 0.57 0.29 0.33 0.29 0.33 0.34 0.33 0.32 Tryptophan 0.46 0.23 0.22 0.18 0.24 0.2 0.05 0.13 0.08 0.13 0.12 0.12 0.08 Tyrosine 1.28 1.56 0.85 0.63 0.53 0.37 0.32 0.34 0.22 0.31 0.28 0.34 0.35 Valine 1.73 1.44 1.21 0.96 0.92 0.68 0.38 0.49 0.42 0.48 0.5 0.5 0.43 Non- essential Alanine 1.53 1.13 1.1 0.9 0.84 0.65 0.56 0.4 0.38 0.42 0.41 0.45 0.32 Aspartic acid 4.01 3.56 3.21 2.37 1.39 1.19 0.49 0.57 0.65 0.66 0.59 0.81 1.45 Glutamic acid 6.34 6.94 4.72 3.32 3.07 2.46 1.41 3.15 1.92 2.4 2.24 1.61 1.38 Glycine 1.55 1.33 1.15 0.9 0.92 0.85 0.29 0.45 0.38 0.44 0.4 0.48 0.28 Leucine 2.72 2.43 2.03 1.45 1.28 0.87 0.93 0.74 0.52 0.64 0.68 0.69 0.46 Proline 1.81 1.41 1.11 0.85 1.15 0.61 0.7 1.09 0.85 0.87 1.04 0.58 0.26 Serine 1.89 1.79 1.36 0.95 0.81 0.64 0.38 0.54 0.38 0.45 0.41 0.46 0.32 https://doi.org/10.1371/journal.pone.0316480.t004 Table 5. Yield by crops considered in linear optimization. Protein and oil plants Energy plants Soybean Lupine Field bean Pea Rapeseed Sunflower Maize Wheat Rye Triticale Barley Oats Potato

Average yield per hectarea t FM/ha 3.17 2.88 2.56 2.99 3.57 2.7 10.52 7.54 5.52 4.88 6.75 4.45 42.26 Humidity of yields % 14% 14% 14% 14% 9% 9% 14% 14% 14% 14% 14% 14% 78% Adjusted average yield per hectare t FM/ha 3.05 2.81 2.56 2.95 3.52 2.65 10.48 7.46 5.48 4.84 6.66 4.37 10.43 a LfL (2023) [82] Note: t FM/ha = tons of fresh matter per hectare. https://doi.org/10.1371/journal.pone.0316480.t005 PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 12 / 27

Table 6. Nutrient supply by crops considered in linear optimization. Protein and oil plants Energy plants Nutrient yield with hydrolysis Soybean Lupine Field bean Pea Rapeseed Sunflower Maize Wheat Rye Triticale Barley Oats Potato Glucose from starch (Legume) 50% Glucose from starch (General) 100% Amino acid yield (General) 70% kg/ha kg/ha kg/ha kg/ha kg/ha kg/ha kg/ha kg/ha kg/ha kg/ha kg/ha kg/ha kg/ha Nutrients Saccharides Glucose 194.95 171.49 519.68 720.99 312.92 97.96 6866.65 4671.16 3110.05 2988.04 3628.13 1664.48 7293.76 Amino acids Essential Arginine 56.08 70.85 46.77 35.71 27.07 19.83 27.15 28.73 16.48 17.94 22.37 19.27 22.64 Cysteine 11.30 10.82 6.09 5.78 11.57 4.82 13.94 14.10 7.67 8.80 10.72 9.48 8.77 Histidine 20.90 14.37 12.19 10.53 11.57 6.86 16.14 13.06 6.90 8.12 10.72 6.12 10.23 Isoleucine 35.82 30.50 19.53 17.55 18.95 10.93 20.55 20.37 11.50 12.86 16.78 11.01 20.45 Lysine 47.76 32.08 30.82 30.55 28.30 11.12 16.88 16.71 13.03 13.20 17.24 12.23 29.95 Mehionin 11.09 5.12 3.40 4.13 10.09 6.30 11.74 9.40 5.37 5.75 7.92 5.20 8.77 Phenylalanin 38.81 25.58 20.07 19.82 17.72 11.68 27.15 27.16 14.56 14.55 22.37 14.37 24.10 Threonine 31.13 24.40 17.02 16.10 21.90 10.56 21.28 17.24 11.12 11.17 15.84 10.09 23.37 Tryptophan 9.81 4.53 3.94 3.72 5.91 3.71 3.67 6.79 3.07 4.40 5.59 3.67 5.84 Tyrosine 27.29 30.70 15.23 13.00 13.04 6.86 23.48 17.76 8.43 10.49 13.05 10.40 25.56 Valine 36.89 28.34

21.68 19.82 22.64 12.60 27.89 25.59 16.10 16.25 23.30 15.29 31.41 Non-essential Alanine 32.62 22.24 19.71 18.58 20.67 12.05 41.10 20.89 14.56 14.21 19.11 13.76 23.37 Aspartic acid 85.50 70.06 57.52 48.92 34.21 22.05 35.96 29.77 24.91 22.34 27.49 24.77 105.91 Glutamic acid 135.18 136.58 84.58 68.53 75.56 45.59 103.47 164.54 73.59 81.23 104.38 49.24 100.80 Glycine 33.05 26.17 20.61 18.58 22.64 15.75 21.28 23.51 14.56 14.89 18.64 14.68 20.45 Leucine 58.00 47.82 36.38 29.93 31.50 16.12 68.25 38.65 19.93 21.66 31.69 21.10 33.60 Proline 38.59 27.75 19.89 17.55 28.30 11.31 51.37 56.93 32.58 29.45 48.46 17.74 18.99 Serine 40.30 35.23 24.37 19.61 19.94 11.86 27.89 28.21 14.56 15.23 19.11 14.07 23.37 Glutamine + Glutamine Acid 135.18 136.58 84.58 68.53 75.56 45.59 103.47 164.54 73.59 81.23 104.38 49.24 100.80 Asparagine + Asparagine acid 85.50 70.06 57.52 48.92 34.21 22.05 35.96 29.77 24.91 22.34 27.49 24.77 105.91 Sum of amino acids 750.12 643.12 459.83 398.39 421.59 240.01 559.19 559.42 308.92 322.54 434.78 272.48 537.59 Sum of amino acids + glucose 945.06 814.62 979.51 1119.38 734.51 337.97 7425.83 5230.58 3418.97 3310.59 4062.90 1936.96 7831.35 https://doi.org/10.1371/journal.pone.0316480.t006 PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 13 / 27

demand for traditional pork production is lower than that for CM production. To estimate the energy demands, we utilized data from several sources. For CM, an average value was obtained from three separate studies: 205.50 MJ/kg [40], 106.00 MJ/kg [50], and 48.00 MJ/kg [95], yield- ing an average of 119.83 MJ/kg. We averaged data from pork production LCAs by Mattick et al. (2015) [50], Dourmad et al. (2014) [97], Reckmann et al. [98], Nguyen et al. (2011) [99], and Wiedemann et al. (2018) [100], which led to a value of 17.16 MJ/kg for pork meat. This led to a differential of 102.67 MJ/kg comparing pork and CM production. In light of the substantial energy differential between CM and pork production, we explored the prospect of integrating solar energy into the CM production process. Solar power was con- sidered as the technology is mostly independent from location factors and most easy to scale, making it more flexible than other renewable energy sources like wind energy. Projected fore- casts indicate the expansion of battery infrastructure in forthcoming years due to technological advancements and increased renewable energy integration [101, 102]. Recent trends have shown the performance of solar panels in the case study region of southern Germany to be approximately 1.35 ha MW/p [96] implying an estimated performance of 851,111.1 kWh/ha/ yr for solar panels in this region. This energy yield was then implemented as part of the crop rotation within our linear programming model. With this approach, we aim to navigate the energy challenge faced by CM production, pre-

senting a holistic picture of the energy requirements and potential strategies for optimizing resource use in this emerging industry. 3.5 Scenarios We present four CM production scenarios, each differing in crop extraction methods and solar energy integration to evaluate efficiency. Four scenarios were developed, each differing in the use of photovoltaic systems for energy supply and the extraction source for protein and sugar. Scenario 1 (max. extraction/no solar energy) involves the production of CM without the integration of photovoltaic systems. Instead, it relies solely on maximal extraction from crops for glucose and amino acids. This could be difficult to apply since it is challenging to fully extract both sugars and hydrolysates from crops. Scenario 2 (max. extraction/solar energy) enhances this approach by incorporating photovoltaic systems, harnessing solar energy to potentially reduce the carbon footprint and reliance on traditional energy sources, while maxi- mizing crop extraction. Scenario 3 (specific extraction/no solar energy), in contrast, excludes photovoltaic systems and instead adopts a more targeted but less efficient extraction approach, utilizing protein-rich plants specifically as protein source and separate energy crops for the required glucose inputs, aiming for resource optimization. Scenario 4 (specific extraction/solar energy) combines the methodologies of Scenarios 2 and 3, employing photovoltaic systems for electricity production and distinctly utilizing protein plants for amino acids and energy crops for glucose sourcing.

The results are compared to several studies that conducted LCAs for pork production. 3.6 Consideration of carbon opportunity costs The carbon opportunity cost of land use for CM is examined, incorporating forgone carbon sequestration into climate impact assessments. Food production and agriculture are responsi- ble for a significant share, namely one-third, of all global greenhouse gas emissions, with the majority emanating from land use and land use change [103]. Carbon opportunity cost is an approach that integrates the forgone carbon sequestration potential of land allocated to specific uses like agriculture rather than allowing it to remain in or revert to a natural state. Every par- cel of land has an intrinsic ability to capture and store carbon, particularly ecosystems like PLOS ONE Evaluating agricultural feedstock for cultivated meat PLOS ONE | https://doi.org/10.1371/journal.pone.0316480 January 22, 2025 14 / 27