Magical disruption? Alternative protein and the promise of de-materialization

A significant body of scholarship has focused on the discursive, regulatory, and ontological work that positions the new “alternative proteins” in relation to existing animal products such as meat, milk, or eggs (Broad, 2020; Chiles, 2013; Jönsson, 2016; Jönsson et al., 2019; Morris et al., 2019; O’Riordan et al., 2017; Sexton, 2016, 2018; Sexton et al., 2019; Stephens, 2013; Wurgaft, 2019). Unlike meat substitutes such as seitan and tofu, they are positioned to be simulacra of meat, eggs, or dairy, or in the case of cellular meat, “versions of rather than alternatives to” meat and animal products (Jönsson et al., 2019: 78). These ontological politics and discursive practices, also known as “promissory narratives,” are critical to convincing funders, consumers, regulators, and broader publics of the food-ness of stuff otherwise seen as non-food (Sexton, 2018) or acceptable as meat given contradictory regulatory and cultural understandings of what meat is (Jönsson, 2016).

At the same time as referencing nearly functional and hedonic identity with their conventional counterparts, these discursive practices also aim to position these products as superior to those stemming from conventional animal agriculture, whether safer, more nutritious, more sustainable, or just higher functioning (Jönsson, 2016; Jönsson et al., 2019; Sexton, 2016, 2018: 595; Sexton et al., 2019; Stephens, 2013). Unlike protein substitutes of yore, that is, these products come with higher order promises of addressing “grand challenge” environmental, food security, and animal welfare concerns, without sacrificing the pleasures and benefits of the animal product counterparts (Jönsson et al., 2019). These products, in other words, are laced with the dual promises of being both world-changing and essentially the same.

While immensely insightful, this literature is for the most part narrowly focused on cellular technologies that are deployed to grow animal products and biotechnologies that are deployed to imitate animal products. It thus fails to account for the role of, say, insect-based foods within the pantheon of alternative proteins, an uneasiness that Sexton et al. (2019) note in their categorization of the new alternative proteins into cellular, plant based, and insect based. This is because edible insects are not used to imitate meat but are generally roasted and ground into flours to be used as ingredients. We find an even broader array of products in the new tech space which are not compared to meat, milk, or eggs but simply promise protein, reflecting the extent to which protein has become what Kimura (2013) calls “a charismatic nutrient.” We thus suggest that the key referent in this boom may be “protein” and not meat, milk, or eggs. As such, innovation is responding not only to imagined ecological, public health, and ethical concerns with livestock production, but with food production methods more generally that are deemed resource intensive or wasteful in relationship to the nutritional needs they fulfill. In light of the overarching aim of these endeavors to use novel means to produce protein qua protein, the object of our study is therefore broader than de-animalization and includes a range of processes and products that are designed to deliver protein. Indeed, according to some of our subjects, imitation of meat, eggs, and dairy is a strategy that is necessary to win over consumers, and therefore the emphasis on mimicry may be temporary.

A second, and for the purposes of this paper more salient concern, is that this literature is somewhat cursory in its examination of narratives regarding how these alternative proteins are brought into being. Jönsson et al.’s (2019) focus on ontologies (72) and on categories based on usage rather than substance (78) explicitly allow them to “move beyond” these concerns (72). Sexton (2018: 587), in contrast, is certainly attentive to the material practices that go into making these products simulacra of their counterparts, as she is interested in what she calls “edibility formation.” For her, edibility formation involves technological practices that attend to foods’ molecular matter, their physical forms as end products, and their visceral attributes. Technologies that make, say, plant-based proteins feel, look, and smell like meat are important to “visceral equivalence,” just as the deployment of molecular reductionism establishes nutritional equivalence (Sexton, 2016). Here, she refers to a comment made by Beyond Meat’s CEO that meat was basically composed of amino acids, lipids, carbs, minerals, and waters—substances also available in plants (p. 67). Yet, the materiality that concerns her is the substance of food as it is experienced by consumers as tasty and nutritious, not as a set of ingredients that come together and are fabricated in laboratories, factories, and other sites and may or may not metabolize like the food to which humans are accustomed.

It is not as if these scholars are unconcerned about how these forms of protein are produced or act on eaters’ bodies. Jönsson (2016: 727) critiques work that emphasizes the instability of cultured meat as an ontological object for “overlook[ing] the materiality of cultured meat [and] the conditions under which research and development (R&D) is undertaken.” He explicitly comments that questions of labor are missing in discussions of the scaling up of cellular meat. Likewise Sexton (2018: 597) points out that the new alternative proteins “present distinct changes to some of the existing materialities of protein supply chains, such as the shift from conventional agriculture to biotechnological contexts and the emergence of new geographies of food production (e.g. Silicon Valley).” Nevertheless, other than these nods—and the important exception of cellular meat (see below)—this literature leaves representations of the supply chains and processes by which these products come to fruition largely unexamined.

Our interest is thus more in the promises related to the materializations, as opposed to the promises regarding the materiality of these alternative proteins. By that we mean we are interested in claims-making about how these products become edible protein as opposed to how they will be experienced by consumers as meat or meat-like. We find this important not only because many of the products we trace claim to make and extract protein in forms that do not necessarily simulate animal products. Rather, we also see in the promotion of these alternative proteins promises of a range of environmental benefits in relationship to conventional animal protein production that they do not, in fact, generally demonstrate. They specifically promise optimization of resource use in the making of protein, if not exactly something from nothing. Though more subtle, we also see a set of nutritional narratives that promise nutritional equivalency despite wildly different ingredients than those for which they substitute (Scrinis, 2008; Sexton, 2016). In short, we see promises that would seem to defy both industrial and bodily metabolism (Guthman, 2015; Metcalf, 2013).

To be clear, our goal here is not to demonstrate the real impact of these technologies in comparison with animal protein. Others are undertaking that project, including those conducting Life Cycle Assessment studies of cellular meat (Lynch and Pierrehumbert, 2019; Mattick et al., 2015; Tuomisto and Teixeira de Mattos, 2011). Rather, our goal is to show how industry discourses and practices both draw attention to and avoid pertinent questions about the transformation of source ingredients into edible protein.

There are hints within critical scholarship that de-materialization is entirely illusory, with processes actually requiring as many or more resources and infrastructure than their counterparts and producing more waste than they aspire to, or simply by being fantastical. Cooper’s (2011) astute observations of the bioeconomy offer a useful description of the imaginary that underpins these critiques. In her view, the sense of limits of growth led to prognostications that investment in the life sciences would allow geochemical production to be “replaced by the much more benign, regenerative possibilities of biomolecular production” and that biological production would transform into a means for generating surplus value (p. 23). Biology, that is, with its potential for limitless growth—something from almost nothing—could both address environmental limits and produce profit (see also Helmreich, 2007). The case of cellular meat is an obviously apposite case in point and the only one that scholars of alternative protein have substantially discussed in this light. Commenting on a raft of recent promissory publications touting a future of “clean” meat (Datar et al., 2016; Post, 2012; Shapiro, 2018), Jönsson (2016: 735) writes that “a fleshy cornucopia of endless meat supply is evoked through depictions of how a single biopsy could theoretically feed the world.” Some have noted that the imaginaries of cellular meat promise no impact whatsoever—“molecularly tuned flesh with no body and thus no apparent ecology” (Metcalf, 2013: 75). For Metcalf, this “de-worlding ethos” of what he calls vat meat or shmeat is precisely its point, even if fantastical:

With cultured meat and similar technologies we see that the response to messy worldliness is to engineer the world out of the technoscientific system – cultured meat operates as a strictly technical fix to the disastrous relations between humans, animals, land and sustenance. If you want meat without faeces (sic) in it, engineer a cow that has no digestive system. If you want to have meat without diseased brain matter, engineer a cow that has no brain. If abusive labour conditions in slaughterhouses result in poor food safety, then grow meat in a bioreactor factory. These imaginaries require an uncritical assumption that both ethics and engineering can operate as if technoscience exists in a closed system readily detachable from flesh, ecology and labour. (83)

Closely related to promises of something from, in this case, just a bit of animal flesh, are the de-materialization promises of drawing on abundant rather than scarce resources, often appearing in discussions about pharmaceutical uses (Cooper, 2011). Describing a laboratory run by the University of Hawaii that is using marine microbes as raw materials for bio-products and pharmaceuticals, Helmreich (2007) suggest that for such endeavors the ocean in particular is a site of both “sky high promise” and “biological fecundity” (289)—a point that is relevant to a range of technologies that attempt to draw on seemingly endless oceanic resources such as algae and kelp and make them useful as human food. Yet, as he suggests, these are not simply extracted. “Considering the university’s treasury of cyanobacteria, it becomes obvious that a lot of work—growing algae, bioactivity screening, changing compounds into units transferable between labs—is required to convert wet wealth into a viable product” (p. 293). In other words, microbes do not reproduce by themselves and it takes labor, food, and material infrastructure to get them to transform into useful products.

A third de-materialization promise involves producing useful goods from waste, sometimes referred to as upcycling. For proponents, the promise of upcycling, “the creation or creative modification of any product out of used materials in an attempt to generate a product of higher quality or value than the compositional elements,” is hindered only by the problem of scaling up (Sung et al., 2014). However, it is often scaling up that requires extensive infrastructure (Fish, 2013). For that reason, critical scholars are typically more skeptical that such processes actually reduce waste. Liboiron (2013) warns that “recycling is an industrial process that produces waste, uses energy, requires virgin (non-recyclable) materials, and often results in down-cycling, where created products are less robust than their predecessors” (p. 10). Likewise, Tracy’s work (2019) on monosodium glutamate shows that its production allows modern food manufacturers to incorporate their own waste and byproducts into the making of a new commodity, but may also lead to unintended break-down products and downstream metabolic effects scientists are just learning ways to trace (p. 559).

It may be an ironic coincidence that the best examples of recycling from food waste involve the worst aspects of industrial livestock production to which these alternative proteins respond. In her glimpse into the long history of using beet pulp, fish heads, and corncobs as animal feed, Landecker (2019) writes it “came hand in hand with new adjuvants intended to correct for the nutrient deficiencies, non-palatability, or disease susceptibilities engendered by new feeding regimes,” (531) which had a host of attendant public health and environmental consequences. For that matter, Blanchette’s (2020) profound examination of industrial pork reveals that “mission-driven” entrepreneurs are far from unique in their dedication to making use of waste. Indeed, he shows that aims of efficiency and profit lead producers to use every bit of hogs, including their shit, in co-products that include the concrete below our feet.

This scholarship thus suggests several ways in which efforts at de-materialization have unwelcome upstream or downstream effects—or simply require more resources than the promises suggest, especially when brought to scale. It also hints at the other concerns that animates this paper: that in spite of efforts among some alternative protein producers to convey an ethic of transparency (Broad, 2020), there is a great deal of silence in their promises about the ecological aspects of alternative protein production and consumption (Jönsson, 2016: 727). That efforts at transparency can also be a shibboleth has been explored perhaps most relevantly in Freidberg’s work. For example, she describes the all-windowed storefront and clear vegetable packages used in London’s Tesco supermarkets to indicate a commitment to full disclosure (Freidberg, 2003), tactics we note are frequently replicated in the food tech sector, such as in the open design of the San Francisco-based biotech accelerator IndieBio and the clear packaging of Beyond Burgers. Friedberg also writes of Walmart’s erstwhile plans to rate the environmental performance of every product it sold as another case in which a company that has been subject of much distrust committed to “radical transparency” (Freidberg, 2015). In both cases, the resulting lack of transparency was less a product of deliberate “greenwashing” than the inability to operationalize transparency in any meaningful way, given enormous difficulties in monitoring, measurement, data collection, and so forth. The effect of such transparency efforts can be further obfuscation, in other words. This article brings together the somewhat established scholarship on promissory narratives specific to the alternative protein space with a fairly amorphous literature on de-materialization, with its hints at the problem of transparency.

For our larger project investigating Silicon Valley’s foray into food and agriculture, we have amassed a database of over 250 agri-food tech companies. These have been identified and collected through existing tech databases, conference and pitch event programs, records of incubators and accelerators, and promotional media. Given the geographic scope of our project (see note 1), the companies in our database either are based in the region or have come through the area for conferences and pitch events. Our database is organized by company and includes descriptions of technologies used, aims, and products (both in development and already available), which are drawn from company websites.

For the purposes of this article, we used this comprehensive database to create a smaller database including only companies whose work is related to the production of edible protein, eventually eliminating companies that sell conventionally manufactured consumer packaged goods (like coconut chips or tofu) but showed up in our database because they appeared at tech events to market their products. While many such products are marketed for their protein content, they are not products of the Silicon Valley style technology-driven protein innovation that is the focus of this study. We focused instead on the tech-driven companies and extracted additional available data on core ingredients and production processes from their websites. It is through this exercise that we found an abundance of claims about the ills of livestock and seafood production but sparse information on how key ingredients were sourced or the transformations that non-animal source ingredients undergo to produce edible protein. We thus came to see just how much magic was being promised in the name of improving upon or eradicating the environmental problems of industrial livestock production and fishing, sometimes explicitly so.

In order to highlight the materiality of these products, we have sorted them primarily by source ingredients, though sorting by process was an approach we considered. Goodman et al.’s (1987) prescient parsing of core biotechnologies into cellular, acellular (e.g. fermentation), and extractive techniques (pp. 125–130) pointed us in the direction of sorting by process, as did the frequent use of such schemas by our subjects. However, we note that the process-based categories are extremely slippery not only because some companies use multiple technologies, but also because—to our point—it is often difficult to discern from available data what techniques are being used. Besides, for the purposes of our analysis, the core ingredients are more germane since they are what the technologies act upon to make protein without the negative externalities of animal production. For heuristic purposes, we classify them into animal stem cells, plants, microorganisms, fungi and algae, insects, and food and beverage processing byproducts. That these ingredients are abundant, mundane, otherwise wasted, or can easily reproduce are the bases of the ecological benefits that companies claim.

In what follows we draw on company self-representations of their production processes drawn primarily from (often-changing) websites, augmented with presentations at events (often involving pitches) and interviews. We note that these representations sometimes differ in emphasis depending on audience (the public, investors, or social science researchers) but the tensions we describe here, and their resulting obfuscations, are ever-present. We use these data to illustrate the promises of de-materialization—the core imperative. But sometimes even this much information is hard to obtain. Differentially navigating the competing imperatives of transparency and secrecy, some companies provide just enough detail to demonstrate the promise of protein from what they deem non-problematic ingredients; some position their products as having less material impact than meat production and fishing just by virtue of using less space (a controversial measure of impact, given that Concentrated Animal Feeding Operations (CAFOs) are part of the problem), fewer resources, or producing less waste and greenhouse gas emissions; some simply assert their environmental better-ness by reciting these typical environmental problems of livestock production without addressing their own input and outputs at all; and some quite explicitly refer to their trade secrets and patents. At the same time, virtually all companies treat the bioavailability, digestibility, and nutritional appropriateness of these highly novel foods as if all protein is equivalent and easily substitutable, and thus will not fundamentally alter the ecology of bodies. As our subjects explained, they are in the business of producing protein and are largely “agnostic” to the source; “we will jump if we find another plant,” one put it.

In keeping with the Silicon Valley ethos, alternative protein companies promise major disruption. But because of the nature of the industry they aim to disrupt, theirs is a unique kind of disruption: an aspiration to replace animal protein with substances that are either simulacra of the real thing or that deliver even more protein, while avoiding the environmental impacts of animal production. At the same time, these companies must embrace an ethic of transparency (Broad, 2020) in order to set themselves apart from agri-food incumbents and manage potential consumer skepticism about novel processes or ingredients, especially genetic engineering. Thus, the many descriptions and diagrams explaining “how we do it” that appear on company websites, avid participation in publicly accessible events, and general availability for media (and academic) interviews. These performances contrast with those of the incumbents of industrial animal agriculture to which they respond, which rarely engage in such representational tactics, and when they do, might showcase a rancher overseeing range cattle rather than a visual of a meat-packing plant. Yet, the imperative to transparency exists in an uneasy tension with the simultaneous imperative for these alternative protein companies to keep their processes a secret. From a business standpoint, secrecy is critical to producing and selling the value of these companies. It prevents technological processes from being easily replicated, thereby securing “first mover” advantages which in turn help shore up return on their substantial investments. Therefore, companies make heavy use of patents and trade secret protections, the latter of which do not require disclosure of processes but can be legally replicated through backwards engineering. Secrecy, along with their big promises of disruption, is what makes these companies investable (cf. Goldstein, 2018), and, in fact, we have witnessed investors “judges” at pitch events voice concerns about whether or not products are sufficiently protected by proprietary mechanisms.

The result of trying to meet these three, often competing, imperatives is the obfuscation and sometimes deliberate black boxing that we have shown above. Representations of production processes are carefully curated to provide what companies want publics and investors to see and hear. Indeed, we find that in discussing their sources, processes, and infrastructures, companies consistently engage in simplifications, redirections, and omissions—and the degree to which our recounting above is vague, it is a reflection of the representational practices we observed. Websites are rife with cartoonishly simple process diagrams, with several depicting a simple progression from source material to bioreactor to product. The undoubtedly complex processes that go on in the bioreactor are barely discussed and, if they are, in language that reveals almost nothing, making the bioreactor a perfect example of Latour’s (1987) black box. Websites and pitch events routinely provide detailed comparisons of their product’s land and water uses compared to animal production, often without stating the nature of the equivalencies, and remain vague or silent about their own ingredient sourcing and processing infrastructures and thus the ecological entanglements of these materials. And many companies boast of their super secretive processes and numbers of patents pending, making meaningful transparency quite the chimera.

Alternative protein’s impact on human bodies is also conveyed optimistically but abstrusely, with much emphasis on protein content, as well as that of other vitamins and minerals, and near silence about how these proteins act as food in terms of edibility, digestibility, and bioavailability. Soylent’s 2016 fiasco, in which a product reformulation undertaken in the style of software upgrade (between versions 1.5 and 1.6) introduced a change that led to nausea, diarrhea, and vomiting among consumers is telling. Entrepreneurs go to great lengths to demonstrate product differences from animal protein but then act as if these products will work in bodies as direct analogs (Jönsson, 2016).

The tensions among competing imperatives, broadly true of much of what comes from Silicon Valley, are thus constitutive of the forms of obfuscation we describe. Moreover, the particular promise of ecological de-materialization heightens the tensions, while throwing the fundamental promise into doubt. De-materialization is no easy feat, to the degree it is possible at all. For the promises to be believable, companies have to create selective vision that highlights what the source ingredients are, not their ecological embeddedness; that they are produced in specific and seemingly benign spaces, not that their supply chains are dispersed and potentially violent; that they cut out animals, not that they take all sorts of other processing; and that they will deliver usable protein to human bodies, not make them sick. Like magic, in other words, they both have to invite acute observation and distract from how the trick is actually done. It is notable, if not at all surprising, that this ecological magic does not appear to be of investor concern. As we have observed at pitch events, skepticism from judges, some of whom are investors, generally stems from questions about technological possibility and scalability, cost effectiveness, and market feasibility, not whether their ecological claims are believable.

Critically, however, the vision of Silicon Valley food tech entrepreneurs is not just of innovative technologies providing an array of novel forms of protein, but of a world remade by these innovations: an agri-food future they deem will be delicious, abundant, affordable, and ecologically sound. Our aim in calling out these de-materializations as magic is not necessarily to detract from this potential. Rather, it is to show that these practices make it difficult, if not impossible, for the public—or anyone really—to meaningfully assess the promises and their potential consequences, much less hold their proponents accountable to anything but pecuniary concerns. It is irrefutable that all of these alternatives do or will have impacts of some kind, leaving many unanswered questions about what bringing any of these alternative proteins to scale might look like ecologically—whether in mono-cropped peas, landscapes of bioreactors, or insect factories. Likewise, they leave many unanswered questions of how these products will act as food and whether consumers will eat them, digest them, get necessary nutrition from them, and not be harmed. In short, the sleights of hand we have described do not provide an adequate basis on which to make rational and reasonably democratic decisions about possible and desirable techno-futures, especially those concerning one of the most consequential aspects of the human–nature metabolism.

This paper was strengthened from the probing comments of the UC AFTeR Project members, Shun-nan Chiang, Kathy De Master, Madeleine Fairbairn, Zenia Kish, and Emily Reisman, along with three anonymous reviewers.