Hostname: page-component-74656f9884-rngvg Total loading time: 0 Render date: 2026-11-05T22:26:05.916Z Has data issue: false hasContentIssue false
  • English
  • Français

Agro-Algorithmic Stockpiles, Food Security and the CAP: Europe’s New Governance of Uncertainty in Agricultural Markets

Published online by Cambridge University Press:  07 September 2026

Antonioluigi Costato Open the ORCID record for Antonioluigi Costato [Opens in a new window]
Antonioluigi Costato*
Affiliation:
Independent Scholar, Rovigo, Italy
*
Email: alcostato@grandimolini.it
Rights & Permissions [Opens in a new window]

Abstract

This article introduces the concept of agro-algorithmic stockpiles as an analytical category for understanding an emerging form of governance in EU agricultural markets, based on the interplay of food security, data infrastructures and algorithmic coordination. Building on the historical role of public stockpiling under Article 39 TFEU and the Common Agricultural Policy (CAP), it argues that the market-stabilisation function is shifting from the management of physical reserves to the regulation of data-driven predictive systems. The analysis is situated within the framework of European risk regulation, focusing on the interaction between the Data Act and the AI Act. While the Data Act establishes the legal infrastructure for access to and non-opportunistic use of agricultural data, the exclusion of agrifood systems from the scope of high-risk artificial intelligence systems results in a regulatory deficit in the management of systemic risks, including algorithmic monocultures, pro-cyclical dynamics and the concentration of decision-making power. By framing food security as a domain of critical infrastructure governance, the article concludes that agro-algorithmic stockpiles constitute a paradigmatic test case for the EU’s risk-based approach to artificial intelligence regulation and for a reassessment of the stabilising function of the CAP in the digital era.

Keywords

AI Actagro-algorithmic stockpilesData Actfood securitysystemic risk

Information

Type
Articles
Information
European Journal of Risk Regulation , First View , pp. 1 - 14
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2026. Published by Cambridge University Press

I. Introduction: agricultural stockpiling in the second half of the twentieth century

The marketing of agricultural products has historically represented a problematic issue for rural communities, traditionally characterised by a position of weakness vis-à-vis commercial operators active in the supply chain. In the agricultural sector, risk assumes a distinctive dimension, defined in agrarian legal scholarship as biological risk, namely, the uncertainty intrinsic to the life cycles of plants and animals, which may be compromised by climatic, pathological or accidental factors capable of significantly affecting the economic outcomes of production.Footnote 1

Precisely because of this structural feature of agricultural activity, instruments designed to manage production insecurity have been developed since the earliest times, anticipating in embryonic forms what is now conceptualised as food security.Footnote 2 As early as Ancient Egypt, for example, state organisations established systems for the collection and storage of surplus cereals intended to cope with famines or fluctuations in harvests, entrusting public granaries with the function of stabilising food supplies.Footnote 3

Comparable logics of public stock management emerged, in different forms, in the cereal policies developed in Europe during the interwar period. In Italy, mechanisms such as compulsory stockpiling and quota-based stockpiling were introduced;Footnote 4 in France, public organisations for the regulation of the wheat market were established, such as the Office du blé and, subsequently, the O.N.I.C., Footnote 5 whereas in Germany, control over the agricultural sector was brought within the corporate structure of the Reichsnährstand. Footnote 6

Despite the different institutional configurations, these experiences shared a common feature: reliance on public management of cereal stocks through the collection and physical storage of surpluses in silos and warehouses, enabling the State to intervene in market dynamics by redistributing foodstuffs during periods of scarcity and thereby contributing to price stabilisation and the protection of food security.

II. From agricultural stockpiling to food security governance: the CAP as a European stabilisation infrastructure

As the economist Thomas Piketty observes, the construction of more stable and equitable economic arrangements depends less on the denunciation of inequalities than on the capacity to design public institutions capable of governing uncertainty;Footnote 7 from this perspective, the common agricultural policy can be read as one of the most significant European attempts at the institutional construction of stability in agricultural markets.

Agriculture itself, with its distinctive economic and social logics, has played a decisive role in fostering the development of agricultural law as a distinct legal discipline. In pursuing the protection of the primary sector, agricultural law has historically introduced limits and corrective mechanisms to the classical logic of laissez-faire, thus emerging as a kind of juridical-economic laboratory in which institutional solutions aimed at promoting a greater balance among the different sectors of the economy have long been tested.Footnote 8

In the aftermath of the Second World War, the food crisis that struck the continent made clear the need to ensure stability in agricultural markets and security of supply, thereby fostering the emergence of the European Economic Community and, in particular, of the Common Agricultural Policy (CAP).Footnote 9

From its inception, the CAP was conceived as an instrument for stabilising agricultural markets through mechanisms of public intervention, including intervention prices and the purchase of surpluses, which gave rise to a system of stock management rooted in the traditional logics of agricultural stockpiling. These instruments represent one of the ways through which public institutions have historically sought to govern the structural uncertainty of agricultural markets.Footnote 10

Today, however, this framework operates within a profoundly transformed context: the growing volatility of commodities, geopolitical tensions, and the emergence of digital technologies are progressively reshaping agricultural markets from predominantly global spaces into systems increasingly shaped by geopolitical logics,Footnote 11 marking a gradual overcome of the trade liberalisation paradigm that emerged after the Marrakesh Treaty and dominated from the 1990s through the first decades of the twenty-first century.Footnote 12

In this new context, the stabilising function historically performed by the CAP tends to acquire an additional meaning: no longer merely as an instrument for the economic regulation of agricultural markets but also as an institutional infrastructure through which the European Union may contribute to the management of the systemic risks that characterise the contemporary food system.Footnote 13 This calls for an examination of how the traditional logics of market stabilisation, historically based on material instruments such as the stockpiling of surpluses and the management of reserves, may evolve toward new forms of food security governance, in which the management of information, the coordination of value chains, and the use of digital technologies play an increasingly prominent role.

It is precisely in light of this transformation that it becomes appropriate to examine more closely the shift from a predominantly global agricultural market to one increasingly influenced by geopolitical logics, as well as the implications that this change entails for agricultural law and for the governance of agri-food value chains.

III. From the global agricultural market to the geopolitical agricultural market: a conceptual premise

Over the past few decades, analyses of agricultural markets have been predominantly framed through the lens of economic globalisation dynamics. From this perspective, international trade in agricultural commodities was progressively incorporated, within the framework of the World Trade Organization, into a multilateral system based on the reduction of trade barriers, the integration of value chains and the growing interdependence of national production systems. Agricultural markets thus came to be conceived as predominantly global spaces in which commodity prices were formed on an international scale and public policies increasingly sought to reduce trade distortions and traditional instruments of state intervention. Within this framework, the European Union’s Common Agricultural Policy likewise progressively adapted its instruments, shifting from a model based on guaranteed prices and the management of surpluses to systems more strongly oriented toward the market and compatible with multilateral trade rules.Footnote 14

In recent years, however, this interpretative paradigm has shown increasing signs of crisis. Growing commodity price volatility,Footnote 15 geopolitical tensions, supply chain fragility and the renewed prominence of food security are progressively reshaping the functioning of agricultural markets. Moreover, the diffusion of digital technologies and data-collection and data-analysis infrastructures along agri-food value chains introduces new forms of economic power and new modes for governing strategic information.

In this context, the agricultural market increasingly tends to be less configured as a neutral space regulated exclusively by global economic dynamics and more as a domain in which the security of supply, control over value chains and the management of information acquire an eminently geopolitical dimension.Footnote 16 The expression “geopolitical agricultural market” thus refers to a context in which the functioning of agricultural commodity markets no longer depends solely on the interaction between supply and demand at the international level but is increasingly shaped by factors such as competition among states for the control of strategic resources, the governance of data along agri-food value chains, and the use of digital technologies as instruments of coordination and economic power.Footnote 17

This shift in paradigm is not without implications for agricultural law. The growing entanglement of agricultural markets, food security, digital infrastructures and geopolitical competition has led to new legal issues concerning the governance of value chains, the management of agricultural data, market stabilisation and the role of public institutions in ensuring a balance between economic efficiency, security of supply, and the protection of collective interests.

IV. Agro-algorithmic stockpiling and data governance in agricultural value chains

The transformations outlined above prompt reflection on the evolution of the instruments through which public institutions have historically sought to govern uncertainty in agricultural markets. In the traditional model, developed primarily during the second half of the twentieth century, market stabilisation largely relied on material instruments and was based on the physical management of agricultural surpluses through systems of accumulation and storage of commodities. From this perspective, cereal stockpiling, supported by mechanisms of public price intervention and stock-management policies, represents a tool aimed at offsetting fluctuations in harvests and ensuring continuity in the food supply.

However, in the contemporary context, the growing availability of agricultural data,Footnote 18 the use of satellite monitoring technologies and the development of predictive models based on artificial intelligence are progressively reshaping the ways in which these objectives can be pursued.Footnote 19 Alongside the management of physical stocks, an increasingly significant dimension is emerging related to the collection, integration, and analysis of information concerning agricultural production, logistical flows, and the dynamics of commodity markets.

It is precisely within this framework that the concept of agro-algorithmic stockpiling can be introduced,Footnote 20 an expression intended to describe an emerging institutional function in which the stabilisation of agricultural markets is achieved not only through the management of physical stocks but also through systems for the collection, integration and predictive analysis of data, capable of supporting both public and private decision-making in the management of food-related risks. This notion is characterised by the coexistence of five elements: (i) a public-interest objective aimed at market stabilisation and the protection of food security; (ii) the integration of heterogeneous data along the entire value chain; (iii) a predictive algorithmic component; (iv) a functional linkage between the governance of information and the governance of physical stocks; and (v) a multi-actor governance structure subject to public oversight.Footnote 21

The terminology adopted here is deliberate. The term “stockpile” renders the historical institution of the heap: the centralised pooling and storage of grain. This practice is as old as organised agriculture, the state granaries of ancient Egypt are an early example, but it took a specific form in the early twentieth century, when physical reserves were concentrated in the hands of a few public or para-public actors. The continuity of the term is intentional: the shift from grain stockpiles to stockpiles of algorithmically processed information does not dissolve the underlying risk of concentration but reconfigures it, displacing it from the control of physical reserves to the control of informational infrastructures by a limited number of actors, notably the large firms of the agro-digital sector. What the term holds in view of, across this transformation, is the question of who controls the stockpile. The paradigm is therefore not decentralised as such: decentralisation is a contested outcome that the European regulatory framework, as the following sections argue, may either foster or frustrate. The qualifier “algorithmic,” rather than “data,” is equally deliberate. Raw data, taken in isolation, are inert; their strategic value, and above all the systemic risk they generate, arise from the large-scale algorithmic processing that converts fragmented information into the capacity to anticipate and steer markets. To speak of “data stockpiling” would understate the point, shifting attention to the substrate rather than to the level at which power and risk concentrate, the level at which, not coincidentally, the two regulations examined below operate: the Data Act on access to and circulation of data, and the AI Act on its processing. The term is used here in the broad sense established in the literature on algorithmic governance and algorithmic regulation, which encompasses contemporary artificial intelligence systems, rather than in the narrow sense of a deterministic set of instructions.

It therefore differs from precision-farming platforms, which are oriented toward the micro-economic optimisation of individual producers, from the administrative information systems of the CAP (such as the SIAN and IACS),Footnote 22 which operate according to an ex post compliance logic; from the European data spaces, which constitute an infrastructural precondition but not a governance function; and from international market-intelligence systems (such as the AMIS, FAO-GIEWS, and JRC-MARS),Footnote 23 which perform monitoring and early-warning functions without integrating information management with stock coordination or incorporating real-time predictive models.

From this perspective, the issue of the availability and management of agricultural data is becoming increasingly relevant not only from a technological standpoint but also from an institutional standpoint. The collection and processing of information related to agricultural production, logistical flows and the dynamics of commodity markets raise questions concerning the modalities of access to, sharing of, and control over such data along agri-food value chains. In this sense, the development of agricultural information systems capable of supporting market stabilisation requires inclusive forms of data governanceFootnote 24 capable of preventing concentrations of informational power and ensuring that the benefits derived from the use of digital technologies are distributed across the entire production value chain.Footnote 25

If agro-algorithmic stockpiling can be understood as an instrument through which the traditional function of stabilising agricultural markets extends from the management of physical commodities to the governance of strategic information, then reflecting on the legal framework governing the collection, circulation, and use of data along agri-food value chains becomes inevitable. In a context in which information assumes the role of a critical infrastructure for food security, issues concerning data ownership, conditions of access and sharing, prohibitions on opportunistic use, and the transparency of algorithmic systems no longer constitute ancillary aspects but rather structural components of the governance of agricultural markets. Stockpiling is no longer merely a place for the accumulation of commodities but rather a function of informational coordination of risk, which must be supported by a regulatory architecture capable of ensuring balance, interoperability, and public oversight of strategic information.

V. From cereal stockpiling to informational stockpiling: the Data Act as the legal architecture of food security

Within the emerging context of agro-algorithmic stockpiling (understood, as noted earlier, as hybrid infrastructures in which the stabilisation of agricultural markets results from the integration of physical stocks and advanced information systems), it becomes essential to reflect on the regulatory framework governing the collection, circulation, and use of data along agri-food value chains. Unlike in the traditional model, which is based on the management of commodities and on public intervention in prices,Footnote 26 in the current scenario, the legal dimension of data is considered a structural element of food security.Footnote 27 This gives rise to the need to analyse the European framework governing the protection, sharing and use of information to assess whether such an architecture is indeed capable of supporting forms of governance able to reconcile efficiency, transparency, and equity across the entire value chain.

Within this framework, Regulation (EU) 2023/2854 (the Data Act)Footnote 28 does not start from the regulation of artificial intelligence as such but rather from the definition of the conditions governing access to, circulation, and use of data, which constitute an indispensable prerequisite for any algorithmic application. In this regard, Recital 5 establishes that data generated through the use of a connected product or related service primarily accrue to the user, affirming their centrality and promoting transparency and the protection of rights.Footnote 29 This framework is accompanied by an articulated system of obligations and rights of access to data. At the level of design, Article 3(1) requires that connected products be designed such that the data generated by their use, including the metadata necessary for their proper interpretation, are, by default, easily, securely, and free of charge accessible to the user in a machine‑readable format. At the informational level, Article 3(2) obliges the seller to inform the user, prior to the conclusion of the contract, of the nature, volume, and frequency of the data that the product will generate, as well as of any capacity to generate data continuously and in real time. At the remedial level, where direct access by default is not technically achievable, Article 4(1) grants the user the right to obtain the data from the data holder without undue delay, free of charge, or with the same quality as that enjoyed by the latter.

This regulatory framework intervenes in the traditional informational asymmetries of agricultural value chains, which are often dominated by machinery manufacturers and digital platforms and open the way to a more distributed and less concentrated form of data governance.Footnote 30 The redistribution of access thus becomes a prerequisite for framing food security as a public good, removing it from purely proprietary logics. The implications for agro-algorithmic stockpiling are evident: the ability of farmers, processing firms and consortia to access and share heterogeneous data makes it possible to aggregate such data and transform them into indicators, predictive models and early-warning systems. In this sense, Article 5 constitutes the legal cornerstone that enables such informational cooperation along the value chain.

The Data Act also introduces an explicit prohibition on the opportunistic use of user-generated data (Recital 27; Arts. 4(13) and 5(6)). Data holders may not exploit information on yields, production methods or economic conditions to obtain bargaining advantages to the detriment of farmers, for instance, by steering purchase prices on the basis of privileged data. This represents a direct response to the risk of informational value extraction by dominant actors in digital agriculture.

Consistent with this approach, the Regulation also intervenes at the contractual level, qualifying as abusive those clauses that limit or prevent access to, use of, or sharing of data (Art. 13; Recitals 58–62). In particular, clauses unilaterally imposed by the device manufacturer are deemed abusive, thereby setting a boundary against “take-it-or-leave-it” practices, which are typical of contracts for the supply of precision-farming machinery and services and are historically characterised by data-exclusivity constraints and forms of lock-in that are difficult to reverse.Footnote 31

The interplay between the prohibition of opportunistic use and the nullity of abusive clauses thus delineates a framework consistent with the logic of agro-algorithmic stockpiling as an instrument for the collective stabilisation of value: the data generated by farmers must serve farmers rather than feeding informational rents for third parties. This gives rise to the conditions for genuinely open, interoperable and non-discriminatory platforms in which data circulate according to shared rules rather than the convenience of the strongest actor.

Added to this is the obligation of interoperability of systems and formats. The Data Act does not merely guarantee access to data but requires that data be technically reusable and transferable, accompanied by the information necessary for their interpretation. The aim is to avoid closed ecosystems and prevent new forms of technological lockdown. For agro-algorithmic stockpiling, this entails the availability of an open infrastructure capable of integrating data streams from sensors, logistics, satellite observation and stock-monitoring systems, thereby enabling predictive models and early-warning mechanisms.

Interoperability thus becomes a legal condition, in addition to a technical condition, of informational stockpiling.

The Data Act allows public authorities to request data held by private actors, which is indispensable for addressing systemic crises: supply shocks, logistical disruptions or tensions that threaten food security. This instrument reconstructs, in digital form, a modern public capacity for intervention analogous to twentieth-century stockpiling mechanisms: no longer control over physical commodities but regulated access to strategic information. This represents the point of conjunction between old and new stabilisation tools.

The regulation further introduces an obligation of accessibility by design for connected products: machinery must be designed in such a way as to generate data that are accessible, exportable and shareable, with documented interfaces and without technical barriers to portability. For digital agriculture, this constitutes a genuine paradigm shift: data governance is no longer an optional contractual arrangement but rather a requirement embedded in hardware itself. In this way, the Data Act dismantles the proprietary logics that have long characterised agricultural machinery and lays the technical and legal foundations for a truly open ecosystem in which data can flow along the value chain without artificial constraints.

The Data Act thus allows the transformation of stockpiling to be understood not as a mere digitalisation of stock-management practices but as a shift toward a governance model based on information as a critical infrastructure. Rules on data ownership and access, prohibitions on opportunistic use, mandatory interoperability, the possibility of mobilising data in cases of exceptional need, and accessibility by design together form a regulatory framework that aims not only to allocate rights but also to reshape the balance of power along the agri-food value chain.

From this perspective, agro-algorithmic stockpiling emerges as the space in which these principles find concrete application: a system in which information is no longer captured by a limited number of technological actors but circulates according to shared rules, becoming a stabilising factor for markets alongside, and beyond, physical storage. Data governance thus becomes the new grammar of food security, marking the transition from a regime centred on commodities to one based on the collective capacity to produce, share and govern strategic knowledge.

The practical effect of these provisions, however, will depend on their implementation, which is still at an early stage: the Data Act has applied only since September 2025, and much of its operational impact remains to be tested. In concrete terms, the recitals and provisions examined above should enable a farmer or a producers’ consortium to obtain, from the manufacturer of a connected machine, the data generated by its use and to aggregate such data, across holdings and across the value chain, into shared indicators, predictive models, and early warning tools that were previously precluded by the exclusive control of manufacturers and platforms. The prohibition of opportunistic use and the nullity of abusive clauses should, in turn, prevent dominant actors from converting this information into a bargaining advantage. However, several conditions determine whether these effects materialise. The technical standards of interoperability on which accessibility “by design” depends are not yet fully defined; the cost and complexity of exercising access rights may continue to weigh disproportionately on smaller actors, and the scope of the public authorities’ power to request data in cases of exceptional need remains to be clarified in practice. The recitals of the Data Act, in other words, articulate a clear redistributive rationale, but their capacity to reshape the informational balance of agri-food value chains will be measured at the level of enforcement rather than of legislative design: a caveat that is particularly apposite given the recency of the act.Footnote 32 However, ensuring access to and circulation of data is only one side of the problem: as the next section argues, the algorithmic processing of that same data introduces a further, systemic dimension of risk that the data-availability framework alone cannot address.

VI. The systemic risk of the algorithmic use of data in informational stockpiling

If the Data Act builds the legal infrastructure for data access and circulation along agri-food value chains, the algorithmic use of such information opens up a different and additional dimension of risk, no longer linked to the mere availability of data but to its transformation into a criterion for decision-making and for the coordination of market behaviour.Footnote 33 When predictive models and artificial intelligence systems are employed to steer production choices, stock allocation, emergency management, or stabilisation interventions, the risk is not limited to isolated errors or individual discrimination but extends to systemic effects capable of amplifying cyclical dynamics, generating pro-cyclical feedback loops, and concentrating decision-making power in the hands of those who control the models.Footnote 34 In this sense, the algorithmic use of data within agro-algorithmic stockpiling may transform agricultural uncertainty into an automated, opaque and scarcely contestable decision-making process, affecting the overall balance of markets and food security. At this level, the notion of risk, as conceptualised by the Regulation (EU) 2024/1689 (AI Act), becomes relevant not as a mere abstract protection of fundamental rights but as a tool for containing the systemic risk arising from the automation of economic coordination functions.Footnote 35

These amplification dynamics warrant closer examination, since they distinguish the systemic risk of agro-algorithmic stockpiling from the localised risks addressed by existing data and consumer protection laws. A first mechanism operates through correlation: when a plurality of operators along the value chain rely on similar predictive models, their decisions tend to align, producing an artificial synchronisation of behaviour, an “algorithmic monoculture,” that increases systemic fragility even in the absence of any explicit coordination. A second mechanism is procyclicality: predictive systems trained on past market dynamics may reinforce rather than dampen phases of expansion or contraction, transmitting and magnifying shocks across the chain through self-reinforcing feedback loops. A third mechanism concerns the concentration of decision-making power: where the automated activation of stabilisation or emergency measures depends on models controlled by a limited number of actors, the assumptions embedded in those models become both decisive and opaque, displacing contestation from the political arena to the inaccessible layer of algorithmic design. Taken together, these mechanisms show that the relevant risk is not the malfunctioning of an individual system but the capacity of widely adopted predictive infrastructures to reshape, in a correlated and self-reinforcing manner, the collective behaviour of agri-food markets: precisely the kind of large-scale, cross-cutting effect that a risk-based regulatory framework should be expected to address.

It should be noted, however, that Regulation (EU) 2024/1689 does not list agriculture or agri-food value chains among the high-risk areas enumerated in Annex III, thereby leaving agri-food predictive systems outside the compliance, transparency, and human-oversight requirements of Chapter III. This exclusion is difficult to reconcile with the choices already made elsewhere in Union law. The CER Directive expressly classifies the production, processing and distribution of food among the sectors among the critical sectors listed, recognising the agri-food chain as critical infrastructure; however, the AI Act does not extend to that same chain the safeguards it reserves for AI systems used in critical infrastructure under point 2 of Annex III. The result is a divergence of risk assessments within the same legal order: a sector deemed critical enough to fall under the resilience regime of the CER Directive is not deemed critical enough to bring the AI systems that govern its dynamics within the high-risk regime of the AI Act. Systematic coherence between the two instruments would therefore warrant treating agro-algorithmic stockpiling (as an informational infrastructure governing the stability of a critical sector) as falling within the “critical infrastructure” area of Annex III, with the consequent application of the high-risk regime. The mechanism contemplated by Article 7, which empowers the Commission to amend Annex III by delegated act, can operate only within the areas already listed therein; it could thus serve to clarify the inclusion of agri-food stabilisation systems under the critical-infrastructure heading but not to create a new agricultural area ex novo. In the absence of such an intervention, the systemic risk identified above remains, paradoxically, the least regulated precisely where the stakes for food security are highest.Footnote 36

VII. Beyond neoliberalism: the CAP as an infrastructure of sovereignty in the era of agro-algorithmic stockpiling

According to the sociologist Wolfgang Streeck, democratic capitalism, understood as the institutional arrangement based on the compromise between market and democracy that characterised the post-war period, entered a phase of progressive disintegration with the rise of late neoliberalism from the 1970s onwards. The previous model, sustained by full employment, the expansion of the welfare state, and redistributive mechanisms, presupposed a state capable of governing markets and mediating social conflicts through political instruments endowed with broad legitimacy. This capacity weakened with the decline of mass parties and with the progressive removal of economic decision-making from the sphere of democratic mediation, thereby reducing state autonomy in the governance of market processes.Footnote 37

Within this trajectory, the Common Agricultural Policy represents a particularly instructive observatory. In fact, the CAP emerged as one of the most fully developed forms of such compromise, grounded in direct public intervention aimed at stabilising agricultural markets, protecting farmers’ incomes, and ensuring security of supply. Within this framework, the regulation of agricultural markets (through instruments such as intervention prices, surplus management, and stockpiling) presupposes a clear assumption of public responsibility in the management of economic uncertainty and systemic risk.Footnote 38

The reforms introduced from the 1990s onwards, within the broader context of market liberalisation and global trade integration, progressively scaled back these instruments, steering the CAP toward forms of support that are decoupled, conditional and formally neutral with respect to market functioning.Footnote 39 However, the structural transformation of contemporary agricultural markets (marked by commodity volatility, geopolitical tensions and the centrality of informational infrastructures) renders an understanding of the CAP as a mere market-accompanying policy increasingly unsustainable.

It is within this space of tension that the function of agro-algorithmic stockpiling is situated. Rather than signalling a return to twentieth-century dirigisme, it configures a new mode of exercising economic sovereignty in which the stabilisation of agricultural markets no longer relies solely on the control of commodities but increasingly on the governance of strategic information, data and decision-making models that guide production, storage and resource allocation. In this sense, the CAP can be read as the institutional infrastructure through which the European Union seeks to reconstruct, in new forms, a public capacity to govern uncertainty, reactivating, beyond the neoliberal paradigm, its original function of mediating between the market, food security and collective interests.

The trajectory outlined above allows for a more general proposition to be formulated: food sovereignty in the twenty-first century is no longer measured in tonnes of stored grain but in the public capacity to produce, govern and redistribute strategic knowledge across agri-food value chains.

From this perspective, agro-algorithmic stockpiling represents the contemporary form through which such capacity becomes institutionally materialised. Just as twentieth-century cereal stockpiles translated public responsibility for security of supply into physical structures, agro-algorithmic stockpiling translates public responsibility for the management of systemic uncertainty into informational architectures. The change is not one of function but of substance: no longer the control of commodities but the governance of data; no longer the reserve of grain but the reserve of knowledge.

In this sense, a CAP that aspires to govern agro-algorithmic stockpiling is neither the CAP of twentieth-century dirigisme nor that of neoliberal decoupling. Rather, it is a CAP that rediscovers its original function, mediating between market forces, food security and collective interests, through radically new instruments. It is a public policy that does not intervene in quantities but rather ensures that knowledge about stocks is produced, aggregated and made accessible; that does not fix prices but safeguards the conditions of transparency and informational symmetry within which prices are formed; that does not manage stocks but ensures that knowledge about stocks is not private privileged.

This reconfiguration directly engages the Streeckian tension between democracy and the market: if the crisis of democratic capitalism manifests itself in the progressive withdrawal of economic decisions from the sphere of political mediation, the governance of agro-algorithmic stockpiling represents, or may represent, an attempt to reverse that trajectory. Bringing back under public oversight, the decision-making models that guide the production, storage and allocation of food resources mean reaffirming that certain choices, those concerning the availability of food, cannot be entirely delegated to private algorithms, multinational platforms or the logic of financialised commodity markets. This means, ultimately, asserting that democracy still has a voice in agriculture.

The question remains open: whether the CAP will be able to embrace this function or whether agro-algorithmic stockpiling will remain a missed opportunity (a technical infrastructure without a political architecture) will depend to a large extent on the willingness of Member States and European institutions to recognise that the governance of agricultural data is not merely a technical matter but rather a question of power.

VIII. Conclusions: from commodity stockpiling to knowledge stockpiling: findings and perspectives

First. The notion of “agro-algorithmic stockpiling,” proposed here in a descriptive sense, has been shown to constitute an autonomous analytical category, distinct from the existing concepts with which it might be confused and irreducible to any of them. Its analytical value lies in bringing together, within a single institutional function, the public-interest objective of market stabilisation, the predictive algorithmic processing of heterogeneous data and a multi-actor governance structure subject to public oversight: a combination that none of the existing categories captures.

Second. The European normative architecture already provides a structured legal framework, albeit not specifically designed for the agricultural sector, within which this function may be situated. The Data Act (Regulation (EU) 2023/2854) constitutes its infrastructural precondition by defining rules on access, portability, interoperability and the prohibition of opportunistic use of data generated along the value chain; the AI Act (Regulation (EU) 2024/1689) safeguards its systemic-risk dimension by intervening in the algorithmic transformation of data into criteria of decision-making and economic coordination. The interaction between the two regulations (the former operating at the level of data availability and circulation and the latter at the level of predictive processing) is a normative coupling potentially capable of supporting the governance of agro-algorithmic stockpiling grounded in transparency, non-discrimination and public oversight. However, neither the Data Act nor the AI Act contains sector-specific provisions explicitly calibrated to agri-food value chains, leaving significant room for further regulatory adjustment. In particular, the AI Act does not list agri-food value chains among the high-risk areas of Annex III, even though the CER Directive already treats the food chain as critical infrastructure; systematic coherence would warrant bringing agro-algorithmic stockpiling within the “critical infrastructure” heading of that Annex, since here the predictive system itself governs the stability of the protected sector.

Third. Common agricultural policy now stands at an institutional crossroads. The transformation of agricultural markets, from predominantly global spaces to systems shaped by geopolitical logics, structural informational asymmetries and the concentration of data in the hands of a limited number of technological providers, calls for a reconsideration of the stabilising function historically exercised by the CAP. From this perspective, agro-algorithmic stockpiling represents the contemporary form through which public responsibility for food security may be re-actualised: no longer through control over commodities but through the governance of strategic information. In the twenty-first century, food sovereignty was measured as the public capacity to produce, govern and redistribute knowledge across agri-food value chains.

It remains to be assessed (and this constitutes both the stated limitation of the present work and its opening toward future research) whether and to what extent European institutions will be able to recognise agricultural data governance not as a merely technical issue but as a matter of institutional architecture and, ultimately, of power. The answer to this question depends largely on the capacity of the next CAP reform to explicitly integrate the informational and algorithmic dimension among its intervention instruments, bridging the gap between the normative infrastructure already in place (the Data Act, the Data Governance Act and European data spaces) and its translation into an agricultural policy that is fully aware of the centrality of data as a public good of food security.

Competing interests

The author declares no competing interests.

AI disclosure

AI-assisted tools (Microsoft Copilot) were used for language revision and translation support during the preparation of this manuscript. The author reviewed and edited all AI-generated output and takes full responsibility for the content of this article.

References

1 M Alabrese, Riflessioni sul tema del rischio nel diritto agrario (Firenze, Edizioni ETS 2009) p 78 ff.

2 On the topic of food security, reference may be made to P Borghi, I Canfora, A Di Lauro and L Russo, Trattato di diritto alimentare italiano e dell’Unione europea (Prima ed. Ed., Milano, Giuffrè 2021) p 8 ff.

3 A M Roth, Egyptian Phyles in the Old Kingdom, The Evolution of a System of Social Organization (Chicago-Illinois, The Oriental Institute of Studies of Chicago 1991) <https://isac.uchicago.edu/sites/default/files/uploads/shared/docs/saoc48.pdf> (last accessed 23 May 2026).

4 See L Costato, L’evoluzione dell’intervento pubblico nel mercato del grano (in Italia e negli altri paesi della MEC) (Milan, Giuffrè 1968) p 11 ff., with reference to Law No. 720 of 30 May 1932 (Art. 4), which allowed silos and collective warehouses to be entrusted also to voluntary consortia of farmers.

5 On the first French public intervention in the wheat market, introduced by the Law of 1 December 1929 (which sought to stabilise prices by reducing the taux de blutage and storing part of the harvest to withdraw supply from the market), see L Costato, L’evoluzione dell’intervento pubblico nel mercato del grano (in Italia e negli altri paesi della MEC), cit., p 36 ff.

6 On the regulation of the German cereal market under National Socialism, see the Reichsnährstand, established by the Law of 13 September 1933 as a public-law corporation bringing together all actors in the production, processing and marketing of agricultural products. Cf. L Costato, L’evoluzione dell’intervento pubblico nel mercato del grano (in Italia e negli altri paesi della MEC), cit., p 49 ff.

7 T Piketty, Una breve storia dell’uguaglianza (Milano, La Nave di Teseo 2021) p 20.

8 See F Albisinni, La sicurezza alimentare in Europa e nelle regole WTO (Accademia dei Georgofili) p 976 ff., where the author observes that agricultural law and food law have, over time, become a “privileged laboratory” for the development and testing of legal institutions and regulatory models, available at: <https://www.georgofili.net/articoli/la-sicurezza-alimentare-in-europa-e-nelle-regole-wto/3542> (last accessed 23 May 2026).

9 For further discussion, see F Sotte, La politica agricola europea, Storia e analisi (Florence, Firenze University Press 2023) p 13 ff.

10 The objectives of the CAP, as defined by Art. 39 of the Treaty on the Functioning of the European Union (TFEU), included market stabilisation and security of supplies. On this basis, intervention prices and mechanisms for the purchase of surpluses were introduced, giving rise to systems of agricultural stock management. See P Ackrill, The Common Agricultural Policy (Sheffield Academic Press 2000); A Matthews, The European Union’s Common Agricultural Policy and the Developing World (Edward Elgar 2017).

11 Agricultural markets are increasingly exposed to geopolitical factors, climate shocks and trade tensions. The Russian–Ukrainian conflict and the growing diffusion of food export restrictions have made the vulnerability of global supply chains increasingly evident. See FAO, The State of Agricultural Commodity Markets, Rome, latest ed.; OECD–FAO, Agricultural Outlook, OECD Publishing, Paris, latest ed.; IPES-Food, Guiding Action for Sustainable Food Systems (2022).

12 The Marrakesh Treaty (1994), through the Agreement on Agriculture, introduced a multilateral framework aimed at reducing trade-distorting subsidies and barriers to agricultural trade, thereby fostering a high degree of market integration. This arrangement is now being challenged by geopolitical tensions, the fragility of supply chains and the emergence of new forms of data governance within agri-food value chains.

13 See European Commission, The Common Agricultural Policy: Supporting Farmers and Rural Communities (Brussels) <https://agriculture.ec.europa.eu/common-agricultural-policy/cap-overview_en> (last accessed 23 May 2026).

14 Starting with the MacSharry reform (1992), the European Union progressively reduced support mechanisms based on guaranteed prices and market intervention, introducing more market-oriented instruments compatible with the multilateral framework established by the Marrakesh Treaty (1994). See European Commission, The Common Agricultural Policy at a Glance <https://agriculture.ec.europa.eu/common-agricultural-policy/cap-overview/cap-glance_en> (last accessed 23 May 2026).

15 Significant episodes of volatility occurred during the 2007–2008 food price crisis and following the Russian invasion of Ukraine in 2022, which triggered sharp fluctuations in the prices of cereals, fertilisers and energy, thereby highlighting the geopolitical vulnerability of global value chains. See FAO, The State of Agricultural Commodity Markets; World Bank, Food Security Update, 2022.

16 See J Clapp, Food (Polity Press 2020); S Strange, States and Markets (Bloomsbury Academic 2015); FAO, The State of Food Security and Nutrition in the World, Rome, various editions.

17 On the growing relevance of the geopolitical dimension of food markets and agricultural value chains, see TA Wise, Eating Tomorrow: Agribusiness, Family Farmers, and the Battle for the Future of Food (New Press 2019); as well as the Food and Agriculture Organization, The State of Agricultural Commodity Markets, Rome, various editions, which highlights how food security, agricultural trade and geopolitical tensions are increasingly interconnected in the governance of global agri-food markets.

18 On the growing economic centrality of data within contemporary production systems, see S Zuboff, Il capitalismo della sorveglianza (Rome, Luiss University Press 2023) p 365 ff., where the author analyses the strategic role of data accumulation and processing in the new models of economic power characterising digital economies.

19 The use of satellite data and predictive analytics systems for monitoring agricultural production is now widespread. At the European level, the Copernicus programme and the European Commission’s MARS system make use of satellite data and forecasting models to monitor crop conditions and anticipate harvest trends. See European Commission, MARS – Monitoring Agricultural Resources; ESA, Copernicus Programme: Earth Observation for Agriculture.

20 On the concentration of agricultural data within private platforms and the ethical implications for producers, see E Jakku et al., “If They Do Not Tell Us What They Do With It, Why Would We Trust Them?” (2019) NJAS – Wageningen Journal of Life Sciences; S van der Burg, MJ Bogaardt and S Wolfert, Ethics of Smart Farming, ibid., 2019; see also the EU Code of Conduct on Agricultural Data Sharing (2018).

21 The expression “agro-algorithmic stockpiling” is used here in a descriptive sense to denote the evolution of traditional instruments for the management of agricultural stocks toward monitoring and forecasting systems based on data and digital technologies. Although the notion has not yet been consolidated within legal scholarship, it fits within the broader debate on agricultural digitalisation and the governance of agri-food data. On the role of digital systems for data collection and analysis in the sector, see G Cirillo, “Transizione digitale e disciplina dei dati nella filiera agroalimentare” (2025) XIX(1) Rivista di diritto alimentare; A Costato, “Agricoltura, alimentazione e intelligenza artificiale: opportunità e sfide per una governance inclusiva” ibid., no. 4, 2025. On the Agricultural Market Information System (AMIS), established by the G20 in 2011 following the 2007–2008 food price crisis, see <https://www.amisoutlook.org> (last accessed 23 May 2026).

22 SIAN (Sistema Informativo Agricolo Nazionale) is the Italian information system managing CAP-related agricultural data and payments; IACS (Integrated Administration and Control System) is the EU-wide framework for the management and control of CAP payments under which Member States operate their national systems.

23 AMIS (Agricultural Market Information System) is discussed supra at note 21; GIEWS (Global Information and Early Warning System on Food and Agriculture) is the FAO system monitoring global food supply and demand; JRC-MARS (Monitoring Agricultural Resources) is the European Commission Joint Research Centre’s crop-monitoring and yield-forecasting system supporting the CAP.

24 On the concept of inclusive economic institutions, see D Acemoglu and JA Robinson, Perché le nazioni falliscono (Milan, Il Saggiatore 2013) p 100, where the authors argue that inclusive economic institutions are grounded in the recognition of secure property rights and in the possibility for a broad and cross-cutting segment of society to participate in economic opportunities, thereby avoiding the concentration of resources in the hands of narrow elites.

25 The risk of informational concentration is analysed by van der Burg, Bogaardt, Wolfert, Ethics of smart farming, cit., pp 4–6, who distinguish between the digital divide (lack of access to technologies) and the big data divide (the asymmetry between those who collect and interpret data and those who are excluded from it), highlighting the use of agricultural data for speculative operations on commodity markets or for input price discrimination (ibid., p. 4, with references to Sykuta, 2016, and Kamilaris et al., 2017). On copyright claims by machinery manufacturers aimed at preventing farmers’ access to the software of their own machines, see ibid., p 6, note 3 (the John Deere case).

26 In this regard, see L Costato, L’intervento pubblico nel mercato del grano in Italia e negli altri paesi della MEC, cit., already referred to in the first part of the present work.

27 On the role of data governance in contemporary economic systems, see European Commission, A European Strategy for Data, COM(2020) 66 final; Regulation (EU) 2023/2854 of 13 December 2023, on harmonised rules on fair access to and use of data (Data Act). With specific reference to the agricultural sector, see J Wolfert, L Ge, C Verdouw and M J Bogaardt, “Big Data in Smart Farming – A Review” (2017) 153 Agricultural Systems 69 ff., which argues that the main challenges of agricultural digitalisation do not concern technological aspects as such, but rather the governance of the data value chain, including issues of availability, access, quality, responsibility, ownership, and privacy.

28 On the regulatory framework governing the circulation of data generated by connected products and related services, see Regulation (EU) 2023/2854 of the European Parliament and of the Council of 13 December 2023 on harmonised rules on fair access to and use of data (Data Act) OJ L, 2023/2854, 22.12.2023, which lays down rules on the conditions for access to, sharing, and use of data, and operates upstream with respect to the regulation of artificial intelligence systems. More generally, see European Commission, A European Strategy for Data, COM(2020) 66 final.

29 On the limits of a purely infrastructural approach to agricultural data sharing, van der Burg, Bogaardt, Wolfert, Ethics of Smart Farming, cit., pp 8–9, observe that the codes of conduct developed thus far (in Europe, the United States and Australia) “have not yet been informed by (…) existing and developing views on what ‘good’ smart farming means, what ethical requirements are concerning the epistemology of data interpretation platforms, or what stakeholders in a network need to trust each other so that they can exchange data and knowledge” (ibid., p 9), thereby confirming the inadequacy of mere data-opening instruments in the absence of an institutional function of predictive coordination.

30 This principle breaks the informational asymmetry of agricultural value chains dominated by machinery manufacturers or digital platforms. As highlighted in the literature, the collection of agricultural data by large technology providers – including multinational input suppliers and digital platforms – generates imbalances in value distribution along the value chain, to the detriment of farmers who generate the data but do not control its use (J. McFadden et al., in OECD Food, Agriculture and Fisheries Paper No. 176, The Digitalization of Agriculture: A Literature Review and Emerging Policy Issues, OECD, April 2022, pp 7–8, 36).

31 Cf. Council Directive 93/13/EEC of 5 April 1993 on unfair terms in consumer contracts, OJ L 95, 21.4.1993, p. 29, Art. 3.

32 The Data Act entered into force on 11 January 2024 and applies, for most provisions, from 12 September 2025; the data-access-by-design obligation (Art. 3(1)) applies to products placed on the market after 12 September 2026, and the regime on unfair B2B contractual terms extends to certain preexisting contracts from 12 September 2027. See European Commission, Data Act, <https://digital-strategy.ec.europa.eu/en/policies/data-act> (last accessed 23 May 2026).

33 In Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act), OJ L, 2024/1689, 12.7.2024, the risk-based approach (Arts. 5–6; Recital 26) is grounded in the awareness that the use of AI systems in predictive or decision-making functions may generate cumulative and systemic effects on resource allocation and on the coordination of market behavior. Recitals 5, 47 and 58 clarify that certain uses of AI may amplify existing dynamics, produce feedback loops, and lead to concentrations of decision-making power, with impacts that go beyond the individual legal relationship. The relevant risk does not concern the mere availability of data, but rather its algorithmic transformation into a criterion of economic coordination, capable of affecting market equilibria and, in the agri-food sector, the security of supply. On the risk-based approach more generally, whose stated aim is to ensure the protection of fundamental rights against the risks posed by such systems without unduly hindering innovation and the uptake of AI technologies, see U Ruffolo (ed.), AI Act. La regolamentazione europea dell’intelligenza artificiale (Rome, Luiss University Press 2025) p 156 ff.

34 On these mechanisms see, on algorithmic monoculture, J Kleinberg and M Raghavan, “Algorithmic Monoculture and Social Welfare“ (2021) 118(22) Proceedings of the National Academy of Sciences; on procyclical feedback loops, H L Lee, V Padmanabhan and S Whang, “The Bullwhip Effect in Supply Chains” (1997) Harvard Business Review and OECD, Supply Chain Resilience (2022); on the opacity of the underlying assumptions, S Wachter, B Mittelstadt and C Russell, “Why a Right to Explanation of Automated Decision-Making Does Not Exist in the General Data Protection Regulation” (2017) 31(2) Harvard Journal of Law & Technology. See also Arts. 9–13 of the AI Act.

35 On the issue of systemic risk associated with the use of artificial intelligence in agri-food value chains and, more generally, with the algorithmic transformation of decision-making processes in the agricultural sector, the following work is recommended: A. Costato, Agricoltura, Alimentazione e Intelligenza Artificiale: Opportunità e sfide per una governance inclusiva, cit.

36 Annex III, point 2, of Regulation (EU) 2024/1689 covers AI systems used in the management and operation of critical infrastructure; agriculture and agri-food value chains are not listed as a separate area. Directive (EU) 2022/2557 (CER Directive), OJ L 333, 27.12.2022, includes “production, processing and distribution of food” among critical sectors listed in its Annex (Annex, point 11). On the amendment of Annex III by delegated act within the existing areas, see Art. 7 of the AI Act. The coherence argument advanced here does not conflate sectoral criticality with the high-risk character of AI systems: it rests on the fact that, in agro-algorithmic stockpiling, the predictive output itself governs the activation of the stabilisation measures on which the continuity of the critical infrastructure depends, so that the two coincide in this specific case.

37 This is the core thesis advanced by W Streeck, Globalismo e democrazia. L’economia politica del tardo neoliberismo (Milan, Feltrinelli 2024) esp. pp 23 ff.: the decline of mass parties and the growing asymmetry between the Staatsvolk (citizens-voters) and the Marktvolk (investors-creditors) remove fundamental allocative decisions from democratic mediation, reducing the State to an executor of market compatibilities. The distinction between Staatsvolk and Marktvolk had already been articulated in Id., Tempo guadagnato. La crisi rinviata del capitalismo democratico (Milan, Feltrinelli 2015) ch II, with specific reference to the European sovereign debt crisis; in Globalismo e democrazia, this distinction is extended to the algorithmic governance of global markets.

38 In this regard, see A S Milward, The European Rescue of the Nation-State (London, Routledge 1999) p 178 ff., who highlights how Community policies, including the Common Agricultural Policy, are embedded within the broader framework of the compromise between economic integration and the strengthening of European States’ capacity to ensure social and economic stability in the post-war period.

39 On the gradual shift from price support and surplus management to market-oriented instruments and decoupled payments, see A Cunha and A Swinbank, An Inside View of the CAP Reform Process: Explaining the MacSharry, Agenda 2000, and Fischler Reforms (Oxford, Oxford University Press 2011) <https://doi.org/10.1093/acprof:oso/9780199591572.001.0001> (last accessed 23 May 2026); on the compatibility of post-1990s CAP reforms with trade liberalisation, see JS Niemi, Implications of Trade Liberalization and Domestic Reforms on EU Agricultural Markets (2009) 12(4) International Food and Agribusiness Management Review 1. <https://ageconsearch.umn.edu/record/92554/?v=pdf> (last accessed 23 May 2026).