Abstract
Using Pig Networks—a platform ecosystem providing big-data-driven farm management, e-commerce, and agrifinance services—as a case study, this article examines the role of the platform in the financialization of China’s hog-agriculture sector. It illustrates how the 2018–2019 national pork supply shortage, caused by the African Swine Flu outbreak, provided political legitimacy for industrial upgrading—namely, the technoeconomic restructuring of traditional sectors through technological innovation, in which industry-finance integration plays a prominent role. This article argues that financializing China’s hog agriculture is operationalized through platformizing the end-to-end lifespan of a hog farming enterprise. The process constructs a feedback-loop system that embeds farming operations within the platform ecosystem, making them susceptible to data-driven, algorithm-aided control. The dual processes of platformization and financialization can lead to two key consequences. First, they create structural incentives for industry actors—particularly producers—to adopt platform-mediated services such as enterprise data verification and loan cycle completion. These services and tools translate farming into discrete, data-driven modular components. Second, they normalize labels and valuation criteria that define optimal farming outcomes largely through the lenses of risk containment, financial performance, and market stability, which promotes a biomechanical, quantitatively conclusive, and inherently partial understanding of farming and life.
Introduction
In November 2020, the Chinese government halted Ant Group’s initial public offering (IPO) on the Shanghai Stock Exchange (Liu et al., 2020). Ant Group is the fintech arm of China’s tech giant Alibaba. It characterizes itself as a big-data-enabled financial “disruptor” that can perform many of the functions traditionally associated with the banking sector. However, according to the Chinese government, the IPO would have led to tremendous overvaluation, worrying gaps in regulatory oversight, and major issues that the group “may fail to meet information disclosure requirements” (Feng, 2020).
The abrupt suspension of Ant Group’s IPO signaled the state’s tightened scrutiny of domestic internet titans and the resubjection of equity markets to state surveillance. Nevertheless, digital finance in traditional sectors like agriculture has been welcomed, exemplified by the case of Beijing Nongxin Hulian Technology Co. (hereafter, Nxin). Formally established in 2015, Nxin is a “Global Top 1000 Unicorn” agritech and platform company with a valuation over 1 billion USD (CBInsights, 2024). Its flagship product is Pig Networks (zhulianwang). The company does not raise a single pig. Instead, it provides big-data-based services for smart farm management, e-commerce, and agrifinance. 1 Pig Networks is a multisided platform ecosystem integrating all services pertaining to the hog-farming value chain. It is officially endorsed by the Ministry of Agriculture and Rural Affairs (hereafter, MARA) of the People’s Republic of China as an outstanding case of agritech innovation and is ranked the second most influential internet of things (IoT) project in China (X Chen, 2022).
China has encouraged the financialization of a traditional sector for various reasons. Since the 2007–2008 global financial crisis, financiers and bankers have become increasingly involved in food systems, spurred in part by soaring food prices, greater demand for meat, a growing biomass market, and the prospect of risk-adjusted returns using alternative investment methods (Schmidt, 2016; Stephens, 2022). Targeted as new arenas for arbitraging and capital accumulation, the financialization of agrifood production and commodity markets has exacerbated corporate concentration and compromised food system resilience (Clapp and Isakson, 2018; Fairbairn, 2014).
Financialization refers to the infiltration of financial actors, instruments, and motives into economic sectors through which profit is increasingly made from financial channels rather than from tangible, productive activities (Epstein, 2005; Krippner, 2011). China’s hog-farming sector is no exception. It is deeply intertwined with the platform economy and the state-mandated effort of “industrial upgrading” (chanye shengji)—the techno-economic restructuring of traditional sectors via research and development (R&D) and technological innovation.
Smart farming presents a fruitful yet understudied ground for research in platform studies. 2 Platform studies have addressed several concerns. The technicity and evolutionary trajectory of platforms (Helmond et al., 2019), along with the paradigmatic shifts brought about by tech giants that reestablish how firms can generate profit (Karppi and Nieborg, 2020; Weigel, 2023; Zuboff, 2018), are among the key issues examined. Researchers of Chinese platforms have studied the politics of digital labor (JY Chen and Qiu, 2019; J Lin and de Kloet, 2019), the mobile app economy, and symbiotic state-corporate relations (de Kloet et al., 2019; Jia et al., 2022).
Media scholars have begun to scrutinize the datafication of agrifood systems. Recently, the notion of farm media has been introduced as a lens through which to analyze the representational, infrastructural, and elemental forms of mediation that relate to agriculture (Kish and Peters, 2023). A representational lens teases out the gendering and colonial legacies undergirding modern farm work (Morris and Evans, 2001; Pringle, 2023). An infrastructural approach attends to how the “digital revolution” (Bronson and Knezevic, 2016) has exacerbated structural inequalities among agrifood players (Fairbairn and Kish, 2023; Fraser, 2019) and affords a particular way to feel about and implement precision-driven agriculture (Carolan, 2023; Miles, 2023). The media elemental approach probes the “constituents of agriculture and its theoretical implications” (Kish and Peters, 2023: 1836), shedding light on the different rhythms, modes of translation, and scales of time experienced by various actors even within practices of commercial farming (Hathaway, 2022; Tsai et al., 2016). Existing critical scholarship about smart farming software tends to focus on North American and European contexts (Bronson, 2022; Wolfert et al., 2017). For China, Zhang (2023) has examined the notion of smart ecosystematic management through the case of container aquaculture. Wang (2020, 2023) has illuminated how AI and blockchain are deployed in the countryside for urban consumers, as well as the tensions between state-led agro-modernization and indigenous agencies.
Using Pig Networks as a case study, this article adds another media studies focal point to farm media: platformization. Specifically, the article critically and theoretically examines how the financialization of hog agriculture goes hand in hand with platformization, and how China’s sociopolitical environment discursively legitimizes these dual processes. By platformization, I draw on Poell et al.’s (2019: 5–6) definition of the process as “the penetration of the infrastructures, economic processes, and governmental frameworks of platforms into different economic sectors and spheres of life.” I use Pig Networks as a case study to examine the convergence of a traditional sector—namely, hog agriculture—with the platform economy and the derivatives market in China. To do this, this article asks: What role does Pig Networks play in facilitating data-driven capabilities and the financialization of China’s hog industry? What does this case reveal about the political dimensions of industrial upgrading in China and its ethical and sociotechnical implications—particularly when such upgrading is driven by platform-mediated governance?
I argue that financializing China’s hog agriculture is operationalized through platformizing the end-to-end lifespan of a hog-farming enterprise. The process constructs a feedback-loop system that embeds farming operations within the platform ecosystem, making them susceptible to data-driven, algorithm-aided control. The dual processes of platformization and financialization can lead to two key consequences. First, they create structural incentives for industry actors—particularly producers—to adopt platform-mediated services for enterprise data verification, algorithm-aided farm management, and loan cycle completion. These services and tools translate farming into discrete, data-driven modular components. Second, they normalize labels and valuation criteria that define optimal farming outcomes largely through the lenses of risk containment, financial performance, and market stability. This serves the political functions of industrial upgrading—in which industry-finance integration plays a prominent role—and promotes a biomechanical, quantitatively conclusive, and inherently partial understanding of farming and life.
In the following sections, I outline the article’s media-genealogy approach and data collection method; I elucidate how the 2018–2019 African swine flu (ASF) crisis in China prompted industrial upgrading and the financialization of the hog-farming sector; I show what platformization entails for hog farming; and I analyze how this process as a whole unfolds as a multilayered, value-aligned scaffolding.
Method
In this article, a media-theoretical critique is prioritized over a grounded empirical investigation. 3 I have analyzed relevant corporate reports, news articles, and television programs that documented the evolution of Pig Networks over nine years (2015 to 2024). All were publicly accessible online, and were drawn from Chinese-language websites. I collected them via a comprehensive manual search using targeted keywords, without the aid of data cleaning or organizing software.
To critically examine the discourse of smart farming in China, I have also collected government policies; journalistic accounts of relevant industry conferences and panels; and industry investigation reports, such as those from Ebrun, China’s influential news platform and consulting services provider on e-commerce and industrial digitalization. Ebrun exemplifies how major industry symposiums and mainstream media portray Pig Networks as providing technical solutions for financialization and innovative business-to-business models in hog agriculture.
The audiovisual materials I have referenced are primarily state-run television programs, including shows from China Central Television (CCTV) and local broadcasts such as SXTV News. They demonstrate how government-sponsored media provide discursive legitimacy to industrial upgrading in hog agriculture through Pig Networks.
A media-genealogical approach is employed in the article. This is a method of problematization that examines how media technology functions as a tool of governance, conditioning how specific power relations, knowledge systems, and subject formation are (temporarily) sustained. The approach is built on Foucault’s (1977) concept of genealogy. Media genealogy addresses the affordances and underlying logics (assumptions, values, and motivations) of a technical media apparatus. More importantly, it attends to “the clashes of power that resulted as multiple technologies were (counter)posed as potential solutions within a problematic field, thus tracing the emergence of a stabilized (socio)technical apparatus” (Monea and Packer, 2016: 3145–3146). This approach is especially well suited to this article, as it reveals how media technology enacts certain power dynamics and brings specific problems—such as government-recognized issues in the hog industry—to the forefront as targets for technical solutions. It allows one to see how Pig Networks enacts an infrastructural mechanism that governs new financial resource channels and aligns data-driven algorithmic capabilities with farming operations in the context of China. It also shows how the platformization process normalizes new ontologies and valuation criteria, such as standardized, quantified metrics used to assess the productive capacity and commercial viability of a hog farm, which become operational throughout the hog-farming cycle within the platform ecosystem.
Industrial upgrading
In China, food security is loosely analogous to assuring a sustainable supply of grain and pork. While pork used to be an occasional gastronomic luxury, China today accounts for half of the world’s pork production, at approximately 55.5 million metric tons per year—virtually all of which is consumed domestically (United States Department of Agriculture, 2023). This “pork miracle” emerged out of the state’s agro-modernization efforts and concerted fiscal policies since the reform era to consolidate a robust, domestic agribusiness sector in service of industrial production and capital (Schneider, 2017).
A recent outbreak of ASF pushed the world’s largest hog-farming sector to a crisis point. ASF was first reported in Liaoning Province in August 2018 and had eruptively spread across China by mid-2019. ASF is highly contagious, and it has an almost 100% fatality rate. It causes severe hemorrhaging in pigs, but the early symptoms (e.g., lameness, mild coughing) are hardly recognizable. MARA (quoted in Ma et al., 2021) statistics showed an abrupt decline in swine inventory—greater than 40% from a year earlier, in October 2019, leading to a wholesale pork price of more than double the pre-ASF level by November 2019. The total economic loss from the ASF outbreak was estimated at approximately 111.2 billion USD, amounting to 0.78% of China’s GDP in 2019 (You et al., 2021, 803–804). Losses were on several fronts: There were direct losses to the swine industry, such as reduced reproduction due to ASF-related deaths and the culling of breeding pigs; indirect losses across linked economic sectors, such as the processed meat industry, with significant impacts in eastern coastal, central, and southern China; and decreases in consumer surplus and government expenditures, such as compensation for affected farmers.
Small-to-medium-scale farmers produce 60% of China’s pigs and constitute the majority of the swine industry. Many of them, burdened by overdue debt, were on the brink of bankruptcy, and many experienced outright financial collapses. Struggling to regain confidence, many small-scale farmers left the sector in large numbers after this devastating ASF crisis (Ding and Wang, 2020). Their direct experience made them acutely aware of the inherent biosecurity risks and the roller-coaster-like pork cycle in the livestock industry.
The Chinese government was under enormous pressure to maintain a stable supply and market price of this food staple. Pork accounts for 62.7% of the Chinese residents’ animal protein intake, and the skyrocketing price, from 1.8USD to 5.3USD per kilogram, soon burgeoned into public discontent (Huaxia, 2019). Frozen pork was liberally imported. The government also released frozen pork from its strategic pork reserves. 4 However, these measures proved insufficient to remedy this severe, prolonged pork shortage. Amid the escalating ASF crisis and the risk of international spread, which could have jeopardized China’s trade reputation, the State Council (2019) issued a directive urging all ministries to take responsibility for restoring the swine industry. In the absence of an effective treatment or vaccine against ASF, neither of which exists to this day, multiple government levels called for industrial upgrading as a long-term resolution.
Industrial upgrading generally refers to the advancement of production standards and the structural change of various economic sectors (S Chen 2016). The industrial upgrading of China’s hog sector has been operationalized via platformization (see “Platformization,” below). There is no standardized definition of the term in China’s official media. Critical attention has focused on the R&D-driven shift toward automation and its impact on workers (Cheng et al., 2019; Huang and Sharif, 2017; Lei, 2021). Sociologist Ya-Wen Lei (2023: 22) has described China’s techno-development from the mid-2000s to the present as a time-compressed process that has been moving away from an economy defined by “labor-intensive, export-oriented manufacturing” toward “techno-state capitalism.” This new system, this “state-capital-talent-technology” symbiosis, is marked by “the rise of tech capital and an asymmetrically symbiotic relationship between tech capital and the state” (Lei, 2021: 6), The Chinese government has not wanted “wild blossoming”—the birds can fly, but only within a state-managed cage, lest they fly away.
Industrial upgrading has been layered into techno-developmentalist agendas such as the “Internet Plus” initiative, a top-down scheme to stimulate e-commerce growth and the integration of information and communication technologies into traditional sectors (State Council, 2015a). “Made in China 2025,” a national action plan, strived to move China away from the status of “world’s factory” to an innovation-driven powerhouse via subsidizing self-developed high-tech manufacturing in 10 strategic areas, including agricultural machinery (State Council, 2015b). China’s current five-year plan (2021–2025), its 14th, sets ambitious objectives for agricultural sectors: By 2025, smart farming would account for 27% of China’s total agrifood production, and 100 high-tech agri-innovation bases are planned to be established (MARA, 2022a). In keeping with the 2021 No. 1 Document, MARA (2021) endorsed “dragonhead enterprises” (officially accredited, leading firms of an industry) as the backbone of industrial upgrading. These enterprises, including Nxin, were to play indispensable roles in optimizing cross-regional supply-and-demand relations, enhancing production efficiency, driving an income increase for farmers, and (re)formatting rural sectors (MARA, 2021).
Financialization
Pig Networks 1.0 debuted in 2015. It offered free services on digital farm management to accumulate industry data (Liao, 2021). In 2016, as the Internet Plus initiative gained momentum, Yang Wang, then-vice premier of the state council, and Changbin Han, then-minister of MARA, visited Nxin and launched a state pig e-market, SPEM, to be China’s only nationwide, government-accredited, pig e-commerce marketplace (Liao, 2021). SPEM allows farmers to digitally list pigs, directly transact with buyers, and restock agricultural inputs. According to Nxin, SPEM helps remove value-extracting intermediaries that diminish small-and-medium-sized farmers’ bargaining power (Yu, 2016). Commercially bred pigs typically go to market at around six months of age, but (unlike manufactured merchandise) they can be sold at any point to maximize revenue. To guide farmers, SPEM publishes historical prices, market trend predictions, the number of pigs sold daily, and the estimated regional inventories. To date, more than 74 million pigs have been sold on SPEM, culminating in a market value of 100 billion yuan/13.73 billion USD (SPEM, 2021).
The sixth version of Pig Networks features a cloud-native, service-oriented digital architecture, comprising five core interconnected service components:
Zhuqiwang/Pig Firm: Data-driven farm management solutions comprising four standard versions, for: Smallholder farms, industrial farms, conglomerate groups, and backyard contract farmers. The annual service fee for industrial farms is around 120,000 yuan/17,000 USD. Zhujiaoyi/Pig Exchange/SPEM: An e-commerce marketplace for live hogs and agricultural inputs. Zhujinrong/Pig Finance: Financial services (e.g., loans, insurance). Zhuxiaozhi/Pig Intelligence: The core backstage system supporting cross-module reference and algorithm-aided front-end functionalities. Zhufuwu/Pig Services: Connecting users to self-developed and third-party services (e.g., online veterinary consultation).
5
China’s official media praised Nxin for trying to break the pork cycle. This periodical fluctuation of pork supply and prices is a long-standing industry dilemma in China (MARA, 2024). Occupying a special place in Chinese cuisine cultures, pork sales ascend seasonally, prompting some farmers to aggressively expand their production capacities. However, overproduction induces price plummets. 6 An episode from CCTV’s documentary series, The Power of Agriculture (Xu, 2018), associated the hog cycle with three structural deficiencies: The dominance of smallholder farmers (those with fewer than 500 heads) who always “rise up and disperse in a rush”; the separation of production and sales as farmers have to be connected by regional aggregators; and information gaps about swine inventories and market demand in each province, including where have been ASF outbreaks. The documentary stated that if more than half of China’s fertile sows could be covered by Nxin’s “specialized business ecosystem,” the pork cycle might be smoothed over, as the platform could facilitate the infiltration of financial instruments and the use of big data to guide decision making for various industry stakeholders (e.g., when to scale up or clear inventory for farmers). It also emphasized that institutional intervention such as price caps, or the coordination of pig sales across provinces, could be undertaken much more easily in an integrated platform environment than in traditional, hustle-and-bustle market networks.
Through the Chinese government’s eyes, one factor holding back farmers from reinvesting in breeding stocks during the 2018–2019 ASF crisis lays in their lack of access to financial capital. Indeed, the Chinese government has been promoting financial intermediation in the countryside since the 2000s, yet farmers’ access to formal finance—credit in particular—remains constrained (Kong and Loubere, 2021). Only a limited number of banks, including the Rural Credit Cooperatives, the Agricultural Bank of China, the Bank of China, and the Rural Postal Savings Bank of China, issue loans to rural households (L Lin et al., 2019). At the 2021 Hog Agriculture Expo, agroeconomist Wenge Fu claimed that “farmers do not make profit from raising pigs but risk control” (quoted in Ebrun, 2021). Here, Fu refers to risk not only in terms of managing biosecurity threats, such as ASF outbreaks, but also in navigating cash flow disruptions.
In a CCTV broadcast, Ying Yu (2016), the director of Nxin’s research institute, pinpointed finance as the actual driver behind hog breeding: The lack of a (rural) credit system hinders farmers from accessing social finance in a transparent manner. . . . Whoever can avail oneself of financial leverage is invincibly positioned in the industry competition.
This is where smart underwriting and blockchain come into play. Nxin has formed strategic partnerships with major agrifinance players (such as China United Insurance Group). It now provides technologized means to track production and funds in real time, automatically underwrites policies, and quantifies farmers’ creditworthiness using self-developed risk-scoring algorithms. Users can compare offers, sign agreements, submit claims, and receive or make loan payments via an all-in-one-place portal on Pig Networks. In December 2021, Nxin launched the Ten Billion Assistance initiative, which approved a total credit line of 10 billion yuan to help farmers resume hog breeding and invest in infrastructural upgrades (Yang, 2021).
Nxin’s credit scoring model uses four types of data: farm performance, financial status, personal credibility, and market conditions. Subscribed users of Pig Firm can import data from “enterprise big data”—a cloud-based database that stores, if not validates, Nxin’s farm partners’ production histories. 8 It is marketed as the “ISO9000 of the pig industry” that can “forensically imprint” the entire lifecycle of each pig (including its vaccination status, breed, provenance) on the blockchain infrastructure as evidence for business records and loss claims (Yu, 2016). Since synchronized records cannot be modified without redoing the proof-of-work, blockchain affords trustworthiness without necessarily trusting the entities on the ledger (Werbach, 2018). In other words, smart contracts help financial institutions to technologically discern fraudulence, so that collateral or a guarantor is no longer required. For a farmer’s personal credibility, Nxin uses their rating on Pig Exchange/SPEM (on such things as the accuracy of the listed information) and other platform activities for reference.
The encapsulation of the entire loan cycle into Pig Network’s algorithm-aided feedback-loop system makes the lending process into an instantiation of “smart” supply chain finance (Zhang, 2020). Often presented as a more affordable alternative to receivables financing, supply chain finance involves a multisided platform with enlisted financial institutions that settle suppliers’ invoices in advance of their maturity dates, based on the buyer’s credit rating. 9 Credit provision traditionally starts with credit granting and ends with either repayment or a claim on the collateral upon default. Instead, Nxin’s platform-mediated agrifinance model has big-data-driven surveillance, assessment, and optimization built into the very design. Funds are not lent out ahead of time, but rather released based on calculated production needs. Farmers use their line of credit to obtain necessary inputs on the platform, and repayment is automatically withdrawn when grown pigs are sold online. Data from the previous growing and loan cycle are used to inform future credit granting (Ebrun, 2021).
This mode of feedback-loop credit granting instantiates a recursive logic. Following pioneering cybernetician Gregory Bateson (1979), recursivity functions as a regulating framework wherein the informational output of an algorithmic process is fed back to the system and informs its latter operations. Such logic is accentuated by the efficacy and seeming objectivity of algorithms. Scholars have cautioned that historical and contemporary prejudices and discriminatory categories are ingrained in every stage of algorithmic decision making (see Campolo and Crawford, 2020; Chun, 2021; Eubanks, 2018). Such algorithmic credit underwriting progressively promotes an implicit understanding of sustainability in hog farming as an enterprise of risk management, for instance by preventing prolonged interruption of the pork supply. This loan-determination mechanism also defines who can stay in the industry, excluding those who fail to meet the criteria.
Industry-finance integration, in the case of Pig Networks, has been operationalized through platformization, including algorithmic credit underwriting and embedded supply-chain-based loan disbursement, which encapsulates the entire loan cycle within the platform ecosystem. However, before financial rules can function, farming operations must first be reorganized around the logics of platforms: farm conditions and activities must be synchronized on the platform and made legible to financiers—“translated” into generic tokens of calculation so that comparison and speculation can take place. The following section elucidates what platformization entails for the production side of a hog-farming enterprise.
Platformization
How does platformization unfold in an industry where not all processes can be fully digitized? Platformizing hog farming adopts a key design strategy of software engineering: modularity. In modular design, an application’s complex functionality is logically partitioned into modular components based on related functions, data links, or other criteria, to ease manageability and maintenance. Nxin’s president, Suwen Xue, stressed that modularity figures prominently in the design of Pig Networks. Xue noted that Nxin had taken countermeasures against the prolonged pork cycle and market slump since 2018. The company marketed its innovation as designed to compartmentalize the software product (Pig Firm) into lightweight, self-sufficient modules; and it launched “Nongxin box,” a mountable device preinstalled with Nxin’s self-developed IoT sensors and algorithmic models, which can, according to Xue, “instantly intelligize” existing farm equipment (e.g., cameras) using edge computing techniques (quoted in Ebrun, 2022b). Xue explained that with a well-established platform-as-a-service base, the “modularization of smart solutions” would allow users to flexibly pick and choose the “software-as-a-service (SaaS) building blocks” (quoted in Ebrun, 2022b). Now, for example, those more concerned with swine pandemics can invest in Nxin’s noncontact temperature measurement service. In traditional practice, a mercury thermometer is inserted to measure rectal temperature. The thermometer needs to remain in the pig’s rectum for a few minutes, an extremely time-consuming process that can cause cross-contamination and high stress for pigs, leading to inaccurate results. Nxin’s noninvasive temperature estimation uses infrared thermography. The pig’s internal body temperature is statistically inferred combining factors like the surface temperature and the humidity level of the microclimate in real time. An alert is triggered when an abnormal temperature rise is detected. Such an “integrated system of hardware, software, and smart solutions” becomes (more or less) tailor-made for the farm, materializing what Xue envisioned as a “digitized, distributed, and decentralized mode of farm management,” one that moves away from “having the manager to make all decisions to using data-driven intelligence to solve problems” (quoted in Ebrun, 2022b).
SXTV News (2021), a government-sponsored local news outlet, once featured a special report presenting Pig Networks as capable of transforming farming into a function-oriented, algorithmically regulated modular operation. It opened with a scene featuring Yibo Zhu, the technical director of Zhonghe pig farm in Xingping, checking the farm conditions on his smartphone. In this farm, 4000 feeder pigs are kept in neatly arranged pens. Each pen has a precision feeder dispensing a grain-based diet based on a prescribed growth curve. Panoramic cameras are mounted on top of the pens. Computer vision is used to estimate the body temperature and weight of each pig in a noninvasive manner. The farrowing house has a smart environmental control system. Air conditioning and ventilation are activated when the temperature is higher than 26 °C or the intensity of hazardous atmospheric elements (e.g., ammonia) exceeds a certain threshold. Zhu tells the reporter that after collaborating with Nxin for a year, the survival rate of weaned piglets has risen to 98%. Significantly, smart farming has also saved human resources: five farmers and a tech operator can now look after 4000 pigs, which was “utterly unthinkable before.”
This episode from SXTV News presents Pig Networks in a techno-solutionist light, portraying it as a platform that can almost effortlessly—or even miraculously—upgrade farm management. Its algorithm-aided modular components claim to manage tasks like disease detection and feed distribution, ensuring pigs reach market readiness at the most profitable time based on a big-data-informed growth curve. All data is accessible via a single dashboard—either through a phone interface or the farm’s central control room—allowing the farm owner to monitor key components of a pig farm in real time for continuous oversight and proactive measures. For integrated “smart” management, Pig Firm, the farm management interface of Pig Networks, offers a choice of eight core modular possibilities:
Daily logistics: Automatic inventory counting, body weight and temperature estimation, environmental control. Breeding management: Estrus calculation, reproductivity performance evaluation, hygiene procedure instruction. Vaccination: Vaccination reminder and proof. Production input management: Procurement and material loss calculation. Biosecurity control: Real-time monitoring of aggressive actions such as tail biting, piglet crushing. Cost management: Multilevel operational cost calculation, summary of utility consumption. Digital bookkeeping. Farm productivity calculation: Productivity-profitability analysis, human resource management.
More than automating certain procedures, compartmentalizing (and containerizing) complex farming operations into modular scenarios of algorithmic interventions constructs a quantified stratum that allows farming to be linked to formal analysis. 10 In the case of Pig Networks, piglets weaned per sow per year (hereafter PSY) is used as the most important benchmark indicating a farm’s (re)productive efficiency and its likelihood to achieve financial sustainability. 11 However, in actual practice, merely improving PSY does not necessarily guarantee greater profit margins. PSY is affected by multifaceted production factors; to name a few: Nonproductive days for sows (NPD), farrowing rate (FR), mating rate within seven days after weaning (MR7DW), and number of piglets born alive per litter (PBAL). Increasing MR7DW can shorten NPD (which lowers maintenance costs) and prolonging the lactation period may increase the proportion of estrus in sows within four to six days after weaning. However, sows with lengthened lactation lose body reserves, which may result in lower piglet survival rates (Guan et al., 2022). Moreover, the daily cost to run an industrial hog farm—the expenses for feed, bedding, labor, veterinary medicine, equipment maintenance, utilities, and advertising—is convoluted. By codifying farming into functional modular scenarios, the efficiency of each module can be internally optimized while also being holistically referenced via Nxin’s self-developed algorithms. Users of Pig Firm receive daily, weekly, monthly, and seasonal reports that connect every course of action to data-driven analytics at a fine-grained scale (e.g., cost per kilogram grown; Nxin, 2021).
In Nxin’s ecosystem, hog farms are also networked: they are being turned into what Nieborg and Helmond (2019) described as “platform instances”—i.e., stand-alone derivatives that provide a distinct view of the larger app ecosystem while facilitating connectivity among various end users and partners. At the same time, they contribute production data and test Nxin’s proprietarily licensed AI capabilities. Algorithm-aided modules (e.g., feed intake, biosecurity) are defined more by nonlinear iterative relations among functions and datafied tokens than by a kind of rigidly constrained procedural structure. As mentioned earlier, many production factors (e.g., MR7DW, FR, or PBAL) are correlated with PSY, but they do not form a kind of linear, causal relationship. Platformizing farming operations into standardized modular components is framed as a solution to understand and manage these relations. For example, to assess a hog farm’s biosecurity risk, the financier can access the relevant module (in this case, biosecurity control) to verify whether the farm is equipped with quarantined facilities and review real-time data to confirm the absence of abnormal temperature fluctuations or contact with potential virus carriers. In other words, modularization translates the complex operations of a farm—once distant from the financier’s oversight—into standardized functional components for data-driven overview and streamlined evaluation. Smartness does not unfold merely as a sequence of fixed, semiautomated steps to be executed, but as algorithm-aided farming scenarios that “embrace the ideal of an infinite range of experimental existences” (Halpern and Mitchell, 2023: 4).
In the discourse on industrial upgrading, driven by the 2018–2019 national pork shortage, smartness serves as an epistemology that rests upon the premise that risk (e.g., ASF, tight liquidity) conditions farming. Risk, in this discourse, can only be managed through building a more resilient infrastructure, one that constantly surveils and optimizes current operations while extending itself by means of a field of data-collecting agents across the platform network to achieve a big-data-informed intelligence.
Platformization, in this light, seeks to abstract and encapsulate formerly messy, multilayered farming operations into platform-mediated modular functionalities, rendering precision-nested calculation and algorithmic control compatible with the everyday practices of a farming enterprise. Abstraction, in this case, is powerful in the context of industrial upgrading—not necessarily because such knowledge is closer to the truth or fundamentally superior to human judgment (though it does automate certain aspects to some extent, inaccuracies and oversights can persist in these automated systems, and many farming tasks—such as mating—still require careful human labor). Rather, its power lies in providing a basis for rational, methodical actions and valuation mechanisms that align well with the data-driven coordination and multi-sided nature of platforms. It renders farms as platform instances within the broader Pig Networks ecosystem and translates farm conditions into legible, standardized data points that can be synchronized across the platform ledger for data-driven assessment (e.g., enabling bankers to verify a farm’s biosecurity level).
It is, perhaps just as importantly, an indispensable tool of persuasion. It is used, for instance, in marketing to construct a compelling rationale for financiers and governments (which subsidize platform development and farm owners seeking to upgrade their farms) on which mode of agriculture can potentially ensure market stability, food security, and a more transparent and systematic approach to loan granting and distribution. It can also be used to adjudicate between competing ideological agendas and worldmaking practices—for example, alternative agricultural models like regenerative agriculture, which resist modularization due to complex ecological relationships beyond quarantined factory settings that “do not compute.”
Industrial upgrading via platformization thus signifies the next level of such a parsimonious, biomechanical reductionism of life, as the vital forces of living organisms are now rendered transactable in the derivatives market. However, filtering agrifood production through algorithms and platform interfaces may lead to an impoverished understanding of what can be known and what is considered “worth knowing”—bracketing out not only what simply “does not compute,” but also what is deemed irrelevant or technically difficult to systematically capture in data-driven terms from the perspective of financialization. The following section elaborates on the deeper ethico-political implications of this transition—specifically, what a future agrifood system could look like if built “from the internet up,” and how platformization also performs an epistemological function by concretizing a value-alignment structure.
Abstraction
Platformizing farm operations seeks to integrate all resources within the hog agriculture value chain into a new business ecosystem. In contrast, the old system is discursively framed as being characterized by scattered smallholder farmers and messy intermediary agents. Business strategy expert Ron Adner (2017) contrasted two modes of business ecosystem constructs: ecosystem-as-affiliation, which highlights how actors are associated based on their network affiliations; and ecosystem-as-structure, which starts with a value proposition and considers the actor coordination necessary for its materialization (43–44). Pig Networks falls into the latter category, as the very design of its platform-mediated business ecosystem promises to:
connect industry resources, so that the state can better oversee swine inventories and supplies across regions; streamline credit underwriting and agricultural insurance programs, which work toward the state’s rural revitalization and financial inclusion agendas; and facilitate the implementation of a big-data-driven managerial philosophy in farming to fulfill the agriculture-specific goals of the 14th five-year plan.
Similar to the ways in which the invocation of targeting in precision seeding does not replace the “industrial logics of standardization” but rather functions as a technique to “intensify and totalize them” (Miles, 2023: 1853), industry-finance integration via platformization also unfolds as an act of valuation, one that renders a particular mode of farming financially sustainable, and validates its rightness as an outcome of reliability qua consistency and algorithmic control. Technically speaking, for algorithms to express an inferential faculty in the first place, they must already be replete with values, thresholds, and probability weightings, so that so-called optimal output may be deduced from a multiplicity of potential pathways (Amoore, 2020, 8). This correlative form of reasoning is intrinsic to the ethicopolitics of algorithms (75). Pig Networks works along these lines. Its instrumental robustness lies in the ability to create a multilayered feedback loop that recursively normalizes a set of labels and valuation criteria through which what counts as the optimal outcome in farming becomes that which best adheres to financial objectives and the steady growth of the larger industry.
In a speech given at the “Seeing the Value of Datafication” internet corporation symposium, Suwen Xue stated: The future of SaaS lies in industry thinking, not product thinking. … Vertical, industry-specialized SaaS must adopt industry thinking, provide services covering the entire value chain … and connect all sorts of industry resources. (quoted in Ebrun, 2022a)
What Li described as being abstract enough does not equal mere datafication; it is neither measuring nor computing per se that renders hog agriculture amenable to the platform business model. Rather, it is a belief in abstraction or the “axiomatic method” as capable of revealing the underlying skeleton of industry operations or other phenomenological correspondences. 12 The method epistemically grounds the reciprocal interchange of two possibilities: the methodical abstraction of models from real-life scenarios, and the population of objects and ways of doing modeled on those abstract deliberations. In this sense, the platformization of China’s hog industry unfolds as a multilayered process of value translation and alignment. Particular sets of indicators, such as PSY, which translate farm conditions into generic forms for financial calculation, are used in it as a particular kind of valuation mechanism. These are systemized throughout a platform ecosystem, wherein the embrace of smartness conflates intelligence with the capacities of big data analytics and predictive modeling.
The creation of platform-as-infrastructure configurations like Pig Networks systematically institutionalizes a data-driven managerial stance that associates the potentially disruptive part of life with pathology—and thus to be contained—while the part remaining productive is counted in a biomechanical manner. Such a biomechanistic view constrains the kinds of questions deemed valuable, the technology designed, and the conclusions drawn. Pigs are treated as small machines. Their actions are seen as mere automatic reflexes to external stimuli. They are perceived as not feeling pain or possessing rights, which renders moral questions irrelevant. The conflation of the lives of the pigs into indifferent numbers—marketed by Nxin as “data-driven end-to-end lifecycle management for each pig” (Kang, quoted in MARA, 2022b)—to be circulated throughout a blockchain infrastructure has markedly limited humans’ ability to notice and appreciate the particulars—the diverse agencies and lifeways—of various more-than-human beings and how they create worlds in and around themselves. 13
Data-driven algorithmic management, though anthropomorphized as “smart” agents and thereby rendered high-tech and trustworthy, deepens the exploitative nature of industrial agriculture. It fabricates the logics of competition (not to mention metabolic exploitation) into the rhetoric of financial inclusion and industrial upgrading. Smart farming tools, hence, are never solely an external appendage that can be freely added or removed; instead, they work as a governance mechanism and aperture instrument, conditioning how capital flows and what mode of farming is financially attainable. The constructed platform ecosystem for industry-wide smart interventions—encompassing the production, commerce, and finance dimensions of the industry chain—aims to remove frictions in data synchronization, value translation, and industry convergence. Yet ontological frictions and epistemic incommensurabilities are the productive conditions that animate interspecies encounters and the co-constitutive relationships between the land and its diverse humans and more-than-human dwellers (Haraway, 2016; Tsing, 2015).
Coda
Pig Networks is a prototype of China’s wider industrial upgrading projects—Sheep Networks, Poultry Networks, and Fish Networks are currently being prepared, in the words of China’s official media, to equip agriculture with “machine-intelligent wings” (MARA, 2022b). Using the case of Pig Networks, this article has shown how industrial upgrading and financialization in China’s hog agriculture go hand in hand with platformization. These processes are (partially) politically legitimized by the imperative to ensure a stable supply of this nationally essential protein source, while being operationalized through the creation of an all-encompassing platform ecosystem that not only integrates the industry value chain but also renders the end-to-end lifespan of a hog-farming enterprise into discrete, data-driven modular components—legible to algorithm-mediated financial speculation and enabling data-driven, top-down control. This establishes systemic conditions that encourage industry actors—particularly farmers—to adopt platform-mediated services for datafication and access to key resources such as credit capital. It also reduces farming to statistical considerations such as PSY or return-on-investment rates, which ultimately convey a utilitarian, consequentialist standpoint that is computational at its core.
Within and beyond its platform governance logic, Pig Networks also reminds us of the ethico-political dimensions of smart farming: how platformization and data farming return to the very biological. Beneath abstraction lies not only a drive for everlasting growth but also the ineluctably concrete realities of metabolic exploitation. Biocapital needs more-than-human vitalities to create value, but that dependence is methodically erased in quantitative reckoning, which reduces animals to “organismal micro-machines” that can conveniently convert grain to cheap protein.
Due to the media-theoretical lens and the scope limitations of this article, further directions remain to be critically explored on the deployment of digital capabilities in agriculture. For example, constructing a platform ecosystem as the default channel for credit-based capital circulation can potentially lead to corporate monopolization and more entrenched forms of sorting—separating those who can profitably carry credit from those who cannot—as wealthier producers who can afford the necessary hardware and support smart farming can secure a privileged position in the competition for loan support. In what ways does the Chinese context diverge from global counterparts? How do digitization and platform penetration affect workers on farms and along the pork supply chain? Equally crucially, should we doubt whether upgrade is the right word, and ask: What is a proper level of abstraction in agrifood production? Is alignment a useful tactic for design, or should we treat such superficial anthropocentrism with nuanced suspicion?
Footnotes
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.