1 Introduction

Over the last decade, digital participation platforms have become embedded in very different parts of political life (Deseriis, 2021). In Taiwan, the non-governmental organization (NGO) “g0v” (government zero) has deployed vTaiwan, a process that combines online and offline procedures to involve citizens, experts, civil society actors, and public authorities in the discussion of national policy issues; g0v notably implemented this process through consensus-building tools, such as the open source platform Pol.is (Moats & Tseng, 2023). In Italy, the Five Star Movement (FSM), a political party founded in 2009 that combines anti-establishment politics with a strong emphasis on digital participation and has been part of government from 2018 to 2022, turned its platform Rousseau into a central instrument of intra-party decision-making, using it to organize candidate selection and ratify crucial political choices (Pianini & Omicini, 2019). At the municipal level, New York, London, and Paris made digital participation a flagship component of participatory budgeting (Calabrese et al., 2020; Doustaly, 2019), while Berlin developed meinBerlin as a central portal intended to amalgamate state- and district-level participation procedures into a single interface (Pruin, 2022). India’s Smart Cities Mission has also accelerated the platformization of urban governance through service portals, command-and-control centers, and data-driven municipal infrastructures (Parkar et al., 2023). While these cases may differ in institutional setting, political ideology, and ambition, they all show a process of platformization of political participation (Santini & Carvalho, 2019a). In all these cases, and many more around the world, digital tools have been promoted as a way to involve citizens more closely in public life, to compensate for declining trust in parties and institutions, and, ultimately, to renew democratic legitimacy (Ercan & Gagnon, 2014).

Historically, digital democracy emerged in the mid-1990s alongside the diffusion of the internet as a mass communication medium (Berg & Hofmann, 2021). The first online tools for political participation, such as deliberative forums, began to emerge in the mid-1990s (Coleman & Norris, 2005; Päivärinta & Sæbø, 2006). Early scholarship was often marked by optimism. Many authors stressed the possibility that digital tools could lower the costs of participation, facilitate direct interactions between citizens and representatives, and broaden the deliberative public sphere (Grossman, 1995; Rodota, 1999; van de Donk et al., 1995). Nevertheless, the critical literature on online and offline participation has repeatedly shown that most of the experiences have limited transformative power, whether because participation remains weakly binding or because publics are unevenly represented (Blondiaux, 2022; Douay, 2016; Gerbaudo, 2019; Santini & Carvalho, 2019b). Anyhow, the intensification of political digitalization in the last 15 years has renewed the relevance of these debates, legitimately. The growth of digital parties, the expansion of civic tech companies, and the diffusion of online platforms in public administration have indeed made digital participation more consequential than during the earlier period of experimental forums and local consultations (Barberà et al., 2021; Greffet & Wojcik, 2008; Randma-Liiv & Lember, 2022).

The literature has already addressed several important dimensions of this phenomenon. A first body of work focused on the emergence, organization, and ideological character of political parties making intensive use of digital technologies (Deseriis, 2020; Franzosi et al., 2015; Lefebvre, 2018; Theviot, 2016). A second examined platform design and its effects on deliberation, participation formats, and political expression (Bravo et al., 2019; Does & Bos, 2021; Kies & Wojcik, 2010; Moats & Tseng, 2023). A third analyzed how digital participation may be instrumentalized for purposes other than genuine co-decision, including communication, agenda control, or the management of public discontent (Gourgues, 2023; Lioy et al., 2019; Mabi, 2019). However, one fundamental question remains surprisingly underexplored: What makes participatory platforms viable from an infrastructural point of view, and what political effects follow from different sociotechnical arrangements?

Scholarship on commercial platforms has already shown that these digital artefacts can be analyzed through many different lenses. Media and communication studies have examined datafication as a key algorithmic, economic, and political dynamic (van Dijck et al., 2018). Science and Technology Studies (STS) have shown how platforms increasingly operate as essential infrastructures of everyday life (Plantin et al., 2018; Plantin & Punathambekar, 2019) – a research stream backed by another that focuses on the material foundations of platforms and the internet, such as marine cables and data centers (Carnino & Marquet, 2018; Cellard & Marquet, 2023). Critical political economy has analyzed the industrial models, extractive logics, and geopolitical concentration of the Silicon Valley platforms (Durand, 2024; Srnicek, 2017; Zuboff, 2019), as well as the digital labor on which they rely (Casilli, 2025). By contrast, research on online political participation has, to the best of my knowledge, never followed these paths. Participatory platforms are increasingly studied through their content and interfaces – the “front-end,” one might say – but almost never through their “back-end:” cloud hosting contracts, maintenance work, governance models for digital infrastructure, and their entanglement with political visions and effects.

This article proposes addressing such blind spot by bringing online participation studies into dialogue with infrastructure research (Bowker & Star, 1999; Epstein et al., 2016; Larkin, 2013; Slota & Bowker, 2017). From this perspective, platforms are not a neutral substrate, but a site of a power struggle that enables and constrains political action. Applied to digital democracy, this STS approach shifts attention toward questions that are typically treated as secondary: where platforms are hosted, who owns or administers their servers, what kinds of scalability they can sustain, how failures are managed, how visible or opaque these arrangements remain to users and publics, and how these back-end issues influence politics. The guiding question of this paper is, therefore, twofold: What kinds of infrastructural arrangements sustain participatory platforms and with what political effects?

The first working hypothesis is that there are at least three models with different tradeoffs. Privately managed infrastructures might offer reliability and scaling flexibility, but they may tend to concentrate control and privilege economic efficiency over democratic accountability (Gerbaudo, 2019; Srnicek, 2017). Public infrastructures may appear more compatible with transparency and digital sovereignty, yet they can remain centralized, under-resourced, and occasionally compatible with surveillance-oriented logics (Nocetti, 2021; Petit, 2020). Commons-based or public-commons hybrid arrangements may hold the strongest democratic promise as they distribute stewardship more horizontally, but they may face organizational and economic fragilities (Dulong de Rosnay & Stalder, 2020; Fuchs, 2021). The second hypothesis is that the political effects of platform material arrangements remain mostly invisible, but can have tremendous impacts when they arise. The idea of this research stems from my observation of media and citizen criticism directed at the FSM in Italy, whenever its Rousseau platform crashed on the days of one-off votes because of excessive traffic (Fatto Quotidiano, 2017). Such dysfunctions, among other issues, diminished the tool’s credibility and appeal, thereby decreasing participation rates in the digital processes it performed (Gerbaudo, 2019; Mosca, 2020).

By investigating these issues, the paper contributes to a denaturalization and re-politicization of the infrastructural layer of digital participation. Analyses and evaluations of platforms should not stop at organization and ideologies, participation rates, interface design, or deliberative quality, but should also include the hidden infrastructural conditions that make these platforms possible. The remainder of the study proceeds as follows. Section 2 presents the materials and methods. Sections 3 and 4 analyze the private and public infrastructural solutions observed in the field. Section 5, the final analytical chapter, examines a largely absent but normatively important alternative: digital commons arrangements as a possible horizon for participatory infrastructures. Section 6 returns to the broader implications of these choices and calls for an increasing integration of infrastructure studies into online participation studies.

2 Materials and Methods

In this section, I first analyze the materials used in this article, then the theoretical framework and the research approach, and finally the three analytical dimensions through which the sociotechnical arrangements found on the terrain have been questioned.

2.1 Materials

The empirical basis for this study consists of 40 qualitative interviews (approximately one hour on average) conducted between 2019 and 2025 with individuals directly involved in the design, deployment, management, or promotion of online participatory tools. The interviews were conducted in face-to-face settings in Paris, Barcelona, Berlin, and Copenhagen, or through videoconference. The sample included a wide range of organizational positions – software developers, platform designers, participation consultants, public servants, civic tech activists, and NGO directors – across four continents. The interviews were semi-structured, allowing for open-ended responses while ensuring coverage of key themes, among which infrastructural issues were systematically featured. The interviews covered actors located in Barcelona, Boston, Brasília, Brussels, Buenos Aires, Copenhagen, Geneva, Graz, Helsinki, Mexico City, Milan, Monterrey, Paris, Taipei, and Yaoundé. The platforms discussed included widely recognized civic tech initiatives, such as Decidim, DemocraciaOS, Rousseau, and vTaiwan, each embedded in different institutional and political contexts. The organizations that employed my interviewees included municipal and regional governments, public innovation agencies, political parties, civil society groups, and private companies operating in the civic tech sector. These interviews constitute the empirical foundation of the small typology developed in the article, and are occasionally quoted to illustrate specific points: Table 1 presents the interviews utilized in the article and mentions their role, platform or civic tech enterprise affiliation, and the date of the interview.

Table 1: List in alphabetical order of the interviews utilized in the article

Interviewee	Role	Platform or enterprise	Date
Roberto Andrade	UX designer	Go Vocal	25 April 2025
Florencia Caffarone	Executive director	Democracia en Red	6 February 2025
Quentin Champenois	Developer	Open Source Politics	11 November 2023
Agustín Frizzera	Board president	Democracia en Red	26 March 2025
Alejandra González	Former Director of Open Government as public servant at Mexico City and Codeando México consultant	Decidim	14 October 2022
Cyril Lage	CEO	Cap Collectif	22 December 2023
Arnau Monterde	Deputy Researcher at the Barcelona City Council and Head of Digital Innovation	Decidim	31 October 2019
Andrés Pereira de Lucena	Computer scientist and Decidim’s system administrator	Decidim	3 March 2025
Enrica Sabatini	Co-president	Rousseau Foundation	3 September 2020
Audrey Tang	Former Minister of Digital Affairs of the government of Taiwan	vTaiwan	6 March 2025
Cui Jia Wei	Coordinator of the activist community	vTaiwan	14 February 2025

The case selection followed a theoretically driven and contrastive logic rather than a representative one. The aim was not to provide a statistical overview of all participatory platforms in the world, but to compare a set of empirically rich cases that vary along different dimensions. The selected cases nevertheless remained comparable as they all belonged to the same family of objects, namely digital platforms used to organize political or public participation, consultation, or co-decision. The comparative logic is thus one of analytical variation: By contrasting cases that differ, I sought to identify trade-offs and begin building a typology of infrastructural models in digital democracy. The epistemic interest of this comparison lies less in measuring the frequency of each model than in clarifying the political consequences of different sociotechnical arrangements. In the iterative and abductive approach that gave rise to the typology (Charmaz, 2008), I began mapping across cases the repertoire of arrangements used by participation promoters to host and manage their platforms. I then consolidated this mapping into a typology of recurring configurations. The ordering principle used to present the typology follows a gradient from the least to the most democratic and sovereign configurations. Here, “democratic” refers to issues of participative governance, transparency, and accountability of platforms, while “sovereign” refers to issues of control over the production and maintenance of the infrastructure.

2.2 General Framework

This article builds on a previously developed interdisciplinary framework that integrates political sociology, digital sociology, and infrastructure studies in order to examine the sociopolitical, software, and hardware dimensions of online participative platforms (Li Vigni, 2026)1. Political sociology helps unpack the institutional arrangements and philosophies that shape each platform’s governance, through a comparative analysis of organizational structures, framing discourses, and the roles of parties, public institutions, and civil society in their development and oversight (Deseriis, 2020). Digital sociology provides tools to examine user interfaces and deliberative framings by analyzing platform design and use cases to understand how these digital tools stage participation (Badouard, 2014). Finally, infrastructure studies direct attention to the material and technical underpinnings of these platforms by tracing server locations, hosting choices, and governance (Musiani et al., 2016).

2.3 Research Approach

Infrastructure studies emerged within STS in the 1980s and 1990s, notably through the work of such historians as Thomas Hughes on large technical systems (Hughes, 1983), and sociologists such as Susan Leigh Star and Geoffrey Bowker, who emphasized the everyday, often invisible role of material settings and standards in shaping social and technical orders (Bowker & Star, 1999). These works sought to re-politicize infrastructure by analyzing how it enables and constrains human activity through complex sociotechnical processes – thus challenging its presumed neutrality (Edwards et al., 2009; Winner, 1980). In these works, infrastructures are viewed not just as material support, but as a site of power struggles. Often relegated to the background until exposed by breakdowns, they encompass the material objects, organizational procedures, and symbolic standards that underpin a wide array of human activity. Anthropologist Brian Larkin (2013) offers a clear and general definition of them:

Infrastructures are built networks that facilitate the flow of goods, people, or ideas and allow for their exchange over space. As physical forms, they shape the nature of a network, the speed and direction of its movement, its temporalities, and its vulnerability to breakdown. (p. 328)

Infrastructures are, he argued, multidimensional – “an amalgam of technical, administrative, and financial techniques” (Larkin, 2013, p. 330) – and characterized by their invisibility in use: “We often see computers not cables, light not electricity, taps and water but not pipes and sewers” (p. 329). These observations resonate with the field’s foundational insight that infrastructures become visible primarily at the moment of failure, and that they are defined by the interrelated dynamics between their material form and social use. 2 In this view, infrastructures are characterized by fragility and endless maintenance (Denis & Pontille, 2015; Domínguez Rubio et al., 2025). Despite this body of work, digital democracy research has largely ignored infrastructure studies. Applying an infrastructural lens to online political participation, this paper explores how infrastructural invisibility is produced, what normative decisions are embedded in infrastructural design, and how this can influence the legitimacy of online participation.

As “infrastructure” can refer to a wide range of technical layers and institutional arrangements, this article narrows its focus to the server-side of participatory platforms. The object of analysis is thus the set of conditions that make a political digital tool operable in practice: hosting solutions, cloud or in-house servers, orchestration strategies, maintenance work, security routines, and staffing. To avoid treating infrastructure as a diffuse background, the analysis distinguishes three conceptual dimensions that are separable for analytical purposes, but intertwined in practice.

2.4 Three Analytical Dimensions

The first dimension is infrastructural ownership and control, that is, the governance of the server layer. This dimension asks who effectively possesses the infrastructure on which participation relies: a private company, a public administration, a civic tech company intermediary, or a commons-oriented collective arrangement. This dimension is analytically central because control over hosting and maintenance shapes the degree of digital sovereignty and the possibility of modifying or discontinuing a platform (Li Vigni, 2026). Here, the question is not simply whether a tool is participatory at the interface level, but whether the backend arrangements themselves are visible and discussable by actors beyond a narrow technical circle. This issue becomes especially visible when one contrasts more open or closed infrastructural cultures. On the one hand, Decidim has institutionalized spaces where techincal and organizational choices can be publicly discussed. The open assemblies of the MetaDecidim community are held during DecidimFest (among other moments) and organized by the association together with members of the coordinating and product teams. They bring together developers, public officials, civic-tech providers, researchers, and other participants to discuss such issues as sustainability, partnerships, management of the association, digital architecture, democratic quality, participation models, and the prioritization of the development roadmap. In this sense, MetaDecidim renders parts of backend governance visible and discussable. By contrast, both Cap Collectif and the FSM have been criticized for limiting this kind of scrutiny in their respective platforms. In the case of the Great National Debate, French critics pointed out that Cap Collectif’s proprietary software made the code and some key processing operations inaccessible to citizens, thereby weakening transparency claims (Knaebel, 2019); FSM’s Rousseau was similarly criticized for the opacity of its code and the concentration of technical control around the enterprise of one of the founders’ (Biondo & Canestrari, 2019).

The second dimension is infrastructural capacity, understood as the material and human ability to keep the platform running. This includes server robustness, scaling capacity, security, repair work, monitoring, and the availability of qualified personnel able to intervene when traffic spikes, bugs, or attacks occur. This dimension is indispensable because participatory legitimacy also depends on whether the platform works under (extra)ordinary conditions. Such point can be illustrated with an example drawn from the DecidimFest 2022 held, as every year, in Barcelona. This was the seventh annual gathering of the international community built around the Decidim participatory platform born in the Catalan capital. One workshop stood out for what it revealed about the hidden material foundations of digital democracy. The session was dedicated to Decidim.Vis, a tool designed to visualize user interactions on Decidim through network graphs. After some discussions on the potential new functions, the conversation quickly shifted to an entirely different terrain: server capacities, memory overload, bandwidth constraints, and the significant fragility of computational processes. Visualizing user interaction networks is not simply a matter of generating elegant graphs, but, first and foremost, a question of managing API calls and hardware limitations. This became strikingly clear when, to answer the question of the workshop animator, the developer of Decidim.Vis explained the recurrent challenges he faced in deploying the tool on real instances of the platform:

Workshop animator: “I have a question: Did you have problems when downloading and using the data in terms of capacity and broadband? Did you crash a Decidim instance [...]?”

Decidim.Vis developer: “Yeah, sometimes it happens. The problem is when you have to download a lot of information from the API and, when it crashes, you have to reload. Another problem is when you are processing interactions between users. At this moment, we don’t have any matching power [to ensure that service]” (DecidimFest workshop recording, 2022).

For a website or platform, a crash is a failure that causes the service to cease functioning, usually because the system can no longer handle a technical problem, such as an overload, hacker attack, or software error. As the developer demonstrated a graph with just 200 users – already enough to overwhelm the system’s memory – the workshop facilitator reacted with surprise and captured the capacity dimension efficaciously:

Workshop animator: “That’s exponential! That’s factorial n! So basically, if you launch this for Barcelona, you will probably be dead before the computation finishes!”

Decidim.Vis developer: “Exactly” (DecidimFest workshop recording, 2022).

Another example illustrates the capacity of infrastructures, focusing on human and financial resources. It comes from an interview with the UX designer at Go Vocal, a civic tech company based in Belgium that develops modular participatory platforms for cities and governments across Europe and North and South America. When his enterprise was contracted by the UK government to support the consultation process Change NHS: A health service fit for the future, the platform unexpectedly attracted millions of visits. The surge in participation produced immediate infrastructural stress, and the Go Vocal team had to mobilize intensively:

… democracy is not a profitable industry, so you don’t have millions to make sure of expenses. And in the end, large companies that operate around the world, like Twitter/X, [Meta, etc.] also have OTs. 3 […] It’s linked to crazy capitalist dynamics. It takes a lot of money to maintain that level of credibility and stability. […] The NHS participation went viral because people on Twitter were sharing jokes and […] ideas that had been uploaded to our platform. And suddenly we had a traffic peak […] I mean, there were three days that we didn’t sleep. We were basically solving problems, reducing the load, finding ways to optimize, and so on. And in the end, we saved [the platform]. It was slow, but didn’t crash. But because we had the resources to do it and because we had people who could dedicate themselves to it. (Roberto Andrade, personal communication, April 2025)

The third analytical dimension concerns the political effects of infrastructural governance and capacity. Software and hardware decisions shape how the platform’s organization and legitimacy are perceived, especially when technical problems disrupt its ordinary functioning. In the cases of Cap Collectif and the FSM, the absence of auditable open-source codes weakened the public credibility of the platforms and, by extension, that of their promoters. Under mounting criticism, both Cap Collectif and the FSM eventually moved toward greater openness by making the platform code available in the former case 4 and, in the latter, only the app’s code (De Carolis, 2019). More broadly, breakdowns and slowdowns can reduce participation, undermine trust in outcomes, and damage the public credibility of the processes that depend on these tools. An Italian newspaper article on the online primary to select the FSM’s candidate for prime minister reported criticism from users:

“After wasting a lot of time, I give up.” “Impossible to vote, I’m asking for an extension.” “I’ve been trying for three hours, now I’m fed up.” The troubled FSM primary for the selection of its candidate for prime minister continues to provoke debate. The electronic polls opened on the Rousseau platform at 10 a.m. and were initially scheduled to close at 7 p.m. this evening. Technical problems, however, led the staff to extend the voting period first until 11 p.m., and then again from 8 a.m. to 12 p.m. on Friday, 22 September: this had never happened before in the history of the Five Star Movement. (Fatto Quotidiano, 2017)

In this excerpt, users’ testimonies convey strong frustration, which likely contributed to the gradual decline in the platform’s participation rates (Gerbaudo, 2019). The same article also reported the remarks of one of the FSM primary’s rival candidates, himself an IT entrepreneur, who advanced two hypotheses to explain Rousseau’s malfunctioning:

“I’m trying to access Rousseau, but the server is extremely slow. I would recommend something more powerful to Casaleggio Associati, which I could provide myself…,” candidate Vincenzo Cicchetti, an IT entrepreneur, told the Adnkronos news agency. […] “There are two possibilities,” he argued, “either the platform is under a hacker attack, or the server is unable to handle the volume of access.” (Fatto Quotidiano, 2017)

The selection of these three dimensions is both theoretical and empirical. Theoretically, the literature has consistently shown that infrastructures should be approached through questions of ownership, dimensions, and social consequences. Empirically, these same issues recurred throughout the interviews: Actors repeatedly spoke about who controlled servers, how platforms could or could not scale, and how moments of failure could affect the credibility of the participative processes. The three dimensions retained here are thus not arbitrary categories, but recurring problems both in academic production and in the fieldwork. These dimensions should also be distinguished from the three models discussed in the next empirical sections. “Private,” “public,” and “public-common” infrastructures are not analytical dimensions in themselves, but ideal-typical configurations produced by different combinations of the three above-described axes. Table 2 summarizes this analytical grid.

Table 2: Analytical dimensions used to study participatory platform infrastructures

Types	Governance	Capacity	Political effects	Examples
Private	Concentrated and provider-led	More or less scalable and performance-oriented	Does not prevent breakdowns when kept in-house or weakly orchestrated; crises often push actors toward stronger cloud and orchestration solutions	Rousseau; Cap Collectif; Open Source Politics; Decidim Barcelona
Public	Sovereign but centralized	Weak material resources and human skills	Can strengthen digital sovereignty, but remains limited by funding, staffing, and bureaucratic inertia	Democracia en Red; vTaiwan; Decidim Mexico City
Commons	Shared, transparent, participatory	Collectively funded and governed servers	Strongest democratic promise, but absent in digital democracy for lack of funding	None

3 Cheap and Reliable. Private Servers or the Digital Dependence

Among the various infrastructural arrangements observed within digital democracy, the most widespread and politically significant is the reliance on private solutions. While the term might evoke corporate ownership in a narrow sense, it encompasses a variety of configurations. This section distinguishes between two levels of such private solutions: first, platforms managed directly by their promoters; second, those entrusted to external civic tech providers that act as infrastructural intermediaries for multiple clients. 5 In both cases, the hosting service is mostly externalized to one or more large cloud companies.

3.1 Self-managed Platforms

At the first level of privatization, server infrastructures are managed directly by the platform’s promoters – be they political parties, public administrations, or civic organizations. These “self-managed” platforms do not outsource provision to third parties, but rather develop and govern the infrastructure themselves, albeit with varying degrees of transparency, capacity, and reliance on commercial services. Two cases – Rousseau in Italy and Decidim in Barcelona – illustrate the technical choices and political implications of this model.

Rousseau, the platform used by the FSM during its time in power in Italy, was conceived as a central pillar of the party’s organizational model (Deseriis, 2017; Gerbaudo, 2019). Promoted as a tool of “direct democracy,” Rousseau allowed members to vote on key decisions and participate in shaping the party’s legislative agenda (Li Vigni, 2022; Pianini & Omicini, 2019). However, its promise of citizen empowerment was undermined by a highly centralized infrastructure (Dotta, 2021). Here, “centralized” refers to the concentration of effective control over hosting and maintenance in one organization or a narrow group of actors. The platform’s source code was never made public (unlike its mobile “app” version). Its architecture relied on a combination of open-source technologies owned by private entities, such as Keycloak 6 and GraphQL, 7 and paid services from corporations and foundations, like Amazon Web Services (AWS) 8 and the Cloud Native Computing Foundation, which provides Kubernetes. 9 However, before this hybrid assemblage – implemented in 2019 (six years after the platform’s launch) through hiring a new system administrator with experience in Silicon Valley – Rousseau was hosted in internal servers and, as discussed in Section 2, faced recurrent technical failures: Voting days were frequently marked by platform crashes due to excessive traffic (De Carolis and Zanca, 2018). This compromised both the quantity of participation and the perceived credibility of outcomes (Biondo & Canestrari, 2018; Mosca, 2020). Episodes of bandwidth congestion and system instability prompted many users to abandon the process mid-vote, fueling criticism and disengagement. According to one of the platform promoters, Rousseau’s later adoption of autoscaling technology, inspired by such platforms as Airbnb or the UK government’s official website, speaks to a logic of infinite scalability to correct the flaws of the previous configuration:

One of the new developments [in 2019], a real IT innovation, is that the new platform we’ve released has an autoscaling system. That means the underlying technology is no longer a limitation with regard to the number of people who can actively participate. So, […] now we can replicate the software as many times as needed to accommodate as many people as we want. In fact, we’re using the same technology used by Airbnb and even the UK government – organizations that want infinite scalability. (Enrica Sabatini, personal communication, September 2020)

Financially, as is often the case, the cost of server usage was variable and depended on traffic and activity levels. Higher participation and cyberattacks (Loucaides, 2020) raise bandwidth consumption and resource usage, increasing expenses in turn. Following deep disagreements between the FSM’s elected officials and the Rousseau Foundation (which managed the platform), in April 2021, the party, led by lawyer Giuseppe Conte, severed its ties with the Rousseau promoters and began migrating user data and technical features to a new site 10 (Fatto Quotidiano, 2021).

The case of Decidim Barcelona presents a contrasting, though not unproblematic, example. Developed by the city administration under Mayor Ada Colau and inspired by such grassroots movements as the 15M (Indignados) and the PAH (Platform for People Affected by Mortgages), Decidim is a public-common project for participatory democracy. As exposed in Section 5, this means that it is funded by the Barcelona administration, but self-organized and piloted by the community developing and using the platform (Barandiaran et al., 2024). Its open-source code and modular architecture have enabled its replication and adaptation in over 30 countries (for a total of more than 400 instances), supported by a community of developers, institutions, and civic actors. In Barcelona, the platform is actively used for participatory budgeting, urban planning, and the co-creation of public policies (Borge et al., 2018; Deseriis, 2020). The administration’s commitment to transparency and deliberation is reflected in the platform’s design, which provides citizens with access to technical documentation, a traceable feedback system for proposals, and structured deliberative tools (Li Vigni, 2022). Yet even here, infrastructural dependencies complicate the picture. Despite its aspiration to become a digital common, Decidim’s hardware layer relies on a mix of commercial hosting services (Arnaud Monterde, personal communication, October 2019). While the codebase is collaboratively developed on GitHub 11 by public institutions and cooperatives – coordinated through the MetaDecidim community – the platform’s main instance in Barcelona is hosted on Heroku, 12 a proprietary platform chosen via public tender. Other services linked to the broader Decidim ecosystem are hosted on Hetzner 13 or benefit from free-tier offers by such providers as Google (via TechSoup), 14 Netlify (for documentation and the website), 15 and GitHub (for code repositories). According to one of the project’s maintainers, the use of these tools is dictated less by ideological choice than by practical constraints: With limited funding and only two developers employed by the Decidim Association, maintaining fully sovereign infrastructure remains out of reach (Andrés Pereira de Lucena, personal communication, March 2025). In other words, while the contract decisions are taken in the MetaDecidim association’s assemblies, digital sovereignty stays absent, since publicity or common property of clouds is unachievable for many administrations.

In both the Italian and Barcelonian cases, platform promoters rely on private solutions for hosting. Initially, Casaleggio Associati operated Rousseau through a proprietary and opaque infrastructure controlled solely by the company itself, rather than by the FSM. (“Opaque” refers to the absence of publicly auditable and discussable information about source code, hosting arrangements, certification procedures, or infrastructural decision-making.) Later, the platform was migrated to more conventional cloud services, but, in both configurations, the reliance on private infrastructure undermined the very trust Rousseau claimed to promote. By contrast, Decidim embraces an open-source ethos and promotes a more distributed governance model. Yet it too remains dependent on private infrastructure providers, due to a lack of public funding for human and server resources. Until their servers stay private, platforms should not be fully considered as digital commons. And yet, the problem of efficiency is a real one: Private cloud providers tend to be more cost-efficient because they combine economies of scale with advanced orchestration tools that dynamically allocate computing resources across large client portfolios, thereby reducing fixed costs, limiting underused capacity, and lowering maintenance expenses – advantages that self-managed and public institutions cannot easily match.

3.2 Civic Tech Companies as Infrastructural Intermediaries

Beyond the case of individual platforms directly managed by political organizations or administrations, a second type of private infrastructural solution is offered by civic tech companies that manage multiple participatory instances for a range of clients. These companies, such as Cap Collectif and Open Source Politics in France, offer both software solutions and infrastructural management, acting as intermediaries between public institutions and the technical architectures that make participation possible.

Founded by French entrepreneur Cyril Lage in 2014, Cap Collectif operates as a software-as-a-service provider for participatory democracy. Its clients range from municipalities and government ministries to private companies and universities. While the company currently insists on the open-source nature of its platform, it maintains full control over server deployment and software updates. The code became available in 2021, but the infrastructural configurations are clearly private: The servers have been hosted on AWS and Google in the past, and are now hosted on the French providers Scaleway and OVH. Interestingly, according to Lage, Cap Collectif’s clients expect flawless platform stability while simultaneously being “allergic to anything that comes from the United States” (Cyril Lage, personal communication, December 2023). Defining himself as a “pragmatist,” Lage explains that he tends to choose “whatever works best while complying with the law,” referring primarily to the GDPR and other digital regulations. As with Rousseau, the underlying logic here prioritizes scalability and performance over digital sovereignty and infrastructural democracy. Although Lage adapts to his clients’ preference for French or European providers, he nonetheless maintains that American companies continue to offer the highest quality services. Even if Cap Collectif’s platforms generally performed well and led to no major publicly reported crash, this emphasis on performance and service continuity did not shield the enterprise from public criticism. In its analysis of the platform used for the national Great Debate, Le Monde highlighted duplicated contributions, empty fields, and other irregularities in the processing of online input, pointing to broader limits in the robustness of the consultation device (Ferrer and Parienté, 2019). Even when a platform remains operational, technical and procedural weaknesses due to limited staff can still undermine the credibility of participation, as the next case illustrates.

Regarding cloud configuration, a similar – albeit more hybrid – approach is adopted by Open Source Politics (OSP), a Paris-based cooperative specializing in the implementation and customization of the Decidim platform for institutional clients. Unlike Cap Collectif, however, OSP works with fully open-source software and willingly relies on European-based solutions. In practice, OSP manages the infrastructure of almost 80 Decidim instances in France and elsewhere, using a Kubernetes cluster deployed via the French cloud provider Scaleway. The orchestration and monitoring of these instances are outsourced to IndieHosters, a small company specializing in managing open infrastructures. In this model, OSP handles the application layer ensuring that each Decidim instance functions properly, while IndieHosters manages the server-side orchestration to scale and secure the entire cluster. OSP adopted this solution after a campaign hosted for the French Senate. A viral petition promoted by environmental journalist and activist Hugo Clément triggered a traffic spike that overwhelmed the servers, causing the platform to crash. The backlash, for the Senate more than for OSP, was swift and damaging:

Before the Kubernetes migration, we were still on our system, where we had one set of machines per client, and in the Senate’s case, we knew there was a lot of traffic, so we had several machines behind it. […] Hugo Clément […] shared a petition […] there was a huge traffic spike all at once […] and the platform crashed. We had to restart it and everything. And Hugo Clément, at that point, tweeted something like […] “Simulacrum of Democracy.” So, he was essentially criticizing the Senate, whereas the real issue was the infrastructure – it wasn’t the client’s fault at all. […] In that moment, there wasn’t much we could do. We could add servers, but it takes a little bit of time and requires manual intervention. Now, thanks to Kubernetes, since it’s an orchestrator, it can detect these things and do automatically what we used to have to do manually. (Quentin Champenois, personal communication, November 2023)

It was the bad buzz launched by Hugo Clément that pushed OSP to adopt Kubernetes and partner with IndieHosters to build a more resilient setup. The platform’s stability and credibility thus depend on the coordinated maintenance of its entire sociotechnical system.

The cases analyzed in this section confirm the first hypothesis: Private infrastructures tend to centralize control, prioritize efficiency, and reduce transparency. Whether managed directly by political actors or mediated by civic tech companies, infrastructural choices are rarely debated publicly and often dictated by scalability and performance. Even when the software is open source and managed as a public-common, server ownership remains private and oriented toward service delivery rather than digital sovereignty.

4 Sovereign but Small: Public Servers or the Ambiguous Autonomy

Some participatory democracy platforms are embedded in public digital infrastructures. While these projects are often constrained by limited funding and human skills, as well as by fluctuating political will, they represent alternative configurations to private ones where online participation is partly supported by public servers and public servants.

One illustrative case is DemocraciaOS, developed by the Argentinian organization Democracia en Red. Originating from the 2013 campaign of the Partido de la Red – a political experiment where elected representatives pledged to follow the will of platform users (Deville & Hallé, 2014; Mancini, 2015) – the initiative evolved into a civic tech foundation with a strong focus on Latin America (Florencia Caffarone, personal communication, February 2025). While Democracia en Red no longer maintains an active international developer community like Decidim’s, all its tools remain open source and are accompanied by customized support to ensure institutional continuity. Initially hosted on Microsoft Azure 16 through an NGO grant, DemocraciaOS later migrated to DigitalOcean 17 to reduce costs. However, one of the interviewees explained that Democracia en Red generally encourages public actors to self-host platforms on their own servers. Examples are Consulta Pública and Leyes Abiertas – operated by the Argentine Government and the National Congress, respectively, and both aimed at crowdsourcing ideas for public action. While security and sovereignty are important considerations in Democracia en Red’s discourse, the primary motivation is different: By hosting the platform themselves, public institutions would be more likely to take ownership of it and to ensure the continuity of both the platform and the participatory process:

Interviewee: We always prefer that the servers belong to the government.

Investigator: Why?

Interviewee: Because that ensures greater commitment, greater ownership of the tool, and more chances that its use will continue over time, beyond any political administration, since it becomes something that is appropriated as belonging to the government. (Florencia Caffarone, personal communication, February 2025)

Another interviewee from Democracia en Red pointed out that only these larger public bodies (the government and the congress) agreed to run platforms on public infrastructure, such as ARSAT, the Argentine national server agency (Agustín Frizzera, personal communication, March 2025). In contrast, smaller municipalities, like San Lorenzo, with limited technical resources, went with private solutions. Considering it a mere cost-benefit issue, local governments indeed prefer to rely on private cloud providers simply because it is cheaper or because their internal IT departments lack the capacity to take charge of infrastructure. In the case of public-hosted platforms, a common trait is that the infrastructure is light: For instance, Consulta Pública has processed roughly 5,400 comments across 393 discussion topics since 2016, while Leyes Abiertas has, since 2018, gathered 120 laws, with only 20 having received between 20 and 200 comments. Hence, they both remain operational on a modest server configuration. However, one of the interviewees explained that when public institutions choose to self-host the digital tools on their own servers, Democracia en Red still includes the cost of platform management in the overall contract. This fee covers the technical assistance required, as public institutions often lack the expertise to properly configure and manage the platform. However, participation levels and the volume of online content remain so limited that Democracia en Red’s instances have never faced any real risk of crashing.

A second prominent example of public-based online participation is vTaiwan, a process using multiple platforms developed by the g0v civic tech community in Taiwan. Unlike DemocraciaOS, which emerged from a partisan initiative, vTaiwan arose from grassroots experimentation – namely the 2014 Sunflower Movement – which revealed a demand for greater citizen participation in Taiwan’s policymaking. While not legally binding, the process has facilitated discussions through sorting and consensus-building algorithms on 28 cases, with official figures stating that 80% resulted in some form of decisive government action. 18 vTaiwan’s infrastructure is hybrid: The domain and platform server (vTaiwan.tw) are managed by g0v contributors, while individual consultations often rely on Pol.is, a deliberation tool hosted by Taiwan’s Public Digital Innovation Space or by the nonprofit Computational Democracy in the U.S., which develops and provides Pol.is. This division of labor was enabled by Audrey Tang, the Taiwanese civic hacker and former Digital Minister closely associated with g0v. While initially sought for its flexibility, it also introduced friction. According to my interviewees from Taiwan, the collaboration between civic tech volunteers and public institutions within the online participatory process has occasionally been hindered by delays in data access or technical maintenance – particularly when governmental staff at the Ministry of Digital Affairs are slow to respond to requests (Cui Jia Wei and Audrey Tang, personal communications, February and March 2025). The political consequence of this is a loss of momentum for g0v in their public advocacy. A slowdown in data communication delays the activists from producing the reports through which they plea for their positions. Despite these frictions, the infrastructure has not suffered from crashes because, unlike Rousseau and similarly to Democracia en Red’s platforms, vTaiwan processes typically unfold over weeks rather than single-day vote surges.

A third case, drawn from a somewhat different context, reinforces the pattern identified in the previous examples. Unlike those hitherto discussed, where the server-side infrastructure remained under the control of civic tech organizations, the Decidim instance in Mexico City was managed by public servants, one of whom explained that:

In general, Mexico has this sentiment of wanting digital sovereignty. So, Mexico City has […] invested in data centers, and we hosted [the platform] there. […] using Decidim is very cheap, because it takes up a very small space on a server, and I did everything myself. […] I installed everything. And I maintained it. […] I had one person who gave me access to the servers. They are Linux servers – Linux Red Hat. […] Its specialty is that it has a lot of security, supposedly. So, I just worked with someone, but that person was an expert in servers, not in Ruby on Rails 19 or any of that. (Alejandra González, personal communication, October 2022)

This quotation leads me to propose three generalizations about the conditions of possibility of the public solution. First, both institutions and civic tech organizations – or at least the latter – must be aware of the issue of digital sovereignty. Second, public servers are sufficient only for limited use of participatory platforms. Third, if the infrastructure is not managed by public servants, it must be managed by civic tech organizations themselves.

Overall, the cases discussed in this section prompt us to reconsider the initial hypothesis. Against it, infrastructure in public solutions is not necessarily synonymous with state or administrative control; frictions between platform promoters and public institutions can arise both upstream and downstream of the participatory process, but for different reasons. Upstream, such frictions concern the conditions under which the platform is deployed, as in Argentina and Mexico, where misalignments may emerge between public institutions and platform promoters over the practical modalities of implementation; downstream, they concern the follow-up to the participatory process itself, as in Taiwan, where administrative delays can slow its political and media uptake. As shown in the next section, underfunding remains the crucial issue leading to the rarity of public or common solutions.

5 The Missing but Ideal Solution: Digital Commons

In a more normative and speculative tone, this section examines why server infrastructures for digital democracy platforms can and should be reclaimed as commons, and why they are not. As underlined by several social scientists, the digital infrastructures that host our social, political, and economic life are increasingly governed by logics of rent extraction and control, rather than democratic participation and public property (Durand, 2024; Srnicek, 2017; Varoufakis, 2023; Zuboff, 2019). As a European based in Europe, I focus here on a specifically continental perspective. Recent analyses of Europe’s digital sovereignty agenda have underscored how limited the EU’s capacities remain (Floridi, 2020; Nocetti, 2021; Pohle & Thiel, 2020a; Roberts et al., 2021). As certain authors have shown, many of the policies and funding meant to bolster EU sovereignty fall short of enabling real infrastructural independence. Overall, Europe’s ambition is hampered by the absence of public alternatives with real scale due to fragmented political will and underinvestment.

Instead, server infrastructures could and should be brought into the realm of the digital commons (Dulong de Rosnay & Stalder, 2020; Fuchs, 2021), or at least into the public sphere (Pickard & Berman, 2019), or into a combination of both (Mendez de Andes Aldama, 2024). Scholars have defined digital commons as open and inclusive resources governed through shared, self-organized institutions that coordinate their reproduction beyond the market and the state: “Commons are resources that are collectively owned and that are reproduced in a cooperative manner” (Fuchs, 2021, p. 17). Making digital infrastructure public means it is operated by functionaries and funded through taxes; their use is free, but its governance is less democratic than commons that exclude citizens from decision-making. A hybrid configuration is the digital public-commons (Bollier & Helfrich, 2019), which is an arrangement whereby public institutions and commons-based communities enter into long-term partnerships to meet policy needs in which “the state provides legal, administrative and economic support to the commons, while recognizing their capacity to manage community projects” (Pera & Bussu, 2024, p. 165). In essence, the state warrants the material conditions for the project, while communities self-govern to continue it. In this third configuration, state support is envisaged to correct the structural fragility of cooperatives within a capitalistic society. As argued by Srnicek (2017) for Silicon Valley’s flagships: “all the traditional problems of co-ops (e.g., the necessity of self-exploitation under capitalist social relations) are made even worse by the monopolistic nature of platforms, the dominance of network effects, and the vast resources behind these companies” (p. 127). While the problem of capitalistic settings and the solution of commons are well-identified, the problem is that theoreticians of digital commons tend to exclude the hardware from the picture:

The digital commons are a subset of the commons, where the resources are data, information, culture, and knowledge, which are created and/or maintained online. […] Unlike tangible commons (such as urban gardens, forests or meadows), the digital commons (such as free software or Wikipedia) are not affected by overuse or material exclusivity. (Dulong de Rosnay & Stalder, 2020)

It is, therefore, time to direct scholarly and public attention to the material underpinnings of participatory platforms, so that their infrastructures can also be imagined and organized as digital commons. If this crucial dimension of online democracy remains absent from debate, reflection on its infrastructural commoning will struggle to advance. While digital sovereignty – “the capacity for digital self-determination by states, companies or individuals” (Pohle & Thiel, 2020b) – is an objective stated by an increasing number of governments, enterprises, activists, and citizens in Europe and around the world, servers, microchips, and cables are too rarely in the spotlight.

Yet, recent studies have shown that examples of an internet managed publicly or cooperatively outside markets logics do exist and are reproductible if the conditions are reunited. For example, writer Ben Tarnoff (2022) called for a de-privatization and democratization of the internet by highlighting U.S. municipal or grassroots alternatives to the dominant, profit-driven model. A standout example is Chattanooga, Tennessee, where the municipally owned Electric Power Board launched the country’s fastest broadband service, known as “the Gig,” delivering 1 gigabit per second through a fiber-to-the-home infrastructure. Funded by municipal bonds and federal grants, this initiative not only bypassed commercial internet service providers but also prioritized affordability by offering discounted rates to low-income families, thus addressing the digital divide issue (van Dijk, 2020). While the Gig is under municipal ownership, other similar initiatives (in North Dakota, for example) are:

operated by rural electric and telephone cooperatives that trace their origins to the New Deal, when the Roosevelt administration seeded them with federal loans as part of the same rural electrification campaign that saw the creation of [the Tennessee Valley Authority, a federal corporation created to electrify the region]. Such cooperatives are owned and controlled by their members, who are also the users of their services. (Tarnoff, 2022, p. 45)

In a context in which Europe remains structurally dependent in digital matters – largely due to its fragmented market, insufficient investment, political divisions, and the overwhelming power of American and Chinese platform firms (Nocetti, 2021) – the digital common-public solution would strengthen democratic oversight and enhance resilience against platform capitalism by combining public support with collective governance. This matters all the more as projects such as Gaia-X, despite their sovereignty-oriented rhetoric, have been criticized both for their modest ambitions and for the inclusion of major non-European cloud actors.

Then, why does even Decidim Barcelona rely on commercial cloud hosting? When asked, the system administrator pointed to a key structural limitation (Andrés Pereira de Lucena, personal communication, March 2025). Existing server commons, such as SinDominio, 20 RiseUp, 21 Autistici/Inventati, 22 and Pangea, 23 are primarily designed for individual activists or lightweight hosting needs, not for the robust, secure, and large-scale deployments required by widely used democratic platforms. While these initiatives represent interesting experiments in building non-commercial, privacy-respecting, collectively governed alternatives, their scope and funding remain modest. They rely largely on volunteer labor and donations – structures difficult to scale up for municipal or national-level platforms. Moreover, as emphasized by my interviewee, hosting Decidim on in-house or public servers would demand additional technical staff – a need that current funding levels cannot meet (Andrés Pereira de Lucena, personal communication, March 2025). As another leading figure in Barcelona’s digital innovation team put it: “The priority is not to be perfect; the priority is to deploy democracy in the city” (Arnaud Monterde, personal communication, October 2019).

The conclusion is thus clear. Confronted with the dominance of private cloud providers, even the most radical proponents of digital democracy must pragmatically prioritize service delivery over infrastructural ideals. If states or the EU genuinely aim to achieve digital sovereignty, they must commit far greater resources than they currently do. As long as austerity policies constrain public investment in digital commons, true digital sovereignty will remain an unattainable ideal.

6 Discussion and Conclusion

The literature on the materiality of digital infrastructure is growing, but I begin this discussion by letting an interviewee illustrate it firsthand within the realm of digital democracy. When motivating the choice of cloud providers, such as AWS, to host their platform, an interviewee in charge of Rousseau depicted this infrastructure in vivid terms: “The investments needed to ensure the security and reliability of the servers would be too great unless you rely on someone who manages them, who maintains armed perimeter guards around the servers, and everything else that’s necessary” (Enrica Sabatini, personal communication, September 2020). However, few practitioners in the digital participation field reflect on their infrastructural dependencies explicitly and publicly.

The previous sections have shown that server infrastructure choices are far from neutral in the deployment of participatory platforms. Rather, they embody political and economic trade-offs that affect who controls participation, how resilient platforms are, and what political effects they uphold. Combining insights from infrastructure studies and digital democracy research, this article has analyzed three types of arrangement – private, public, and common – each with its own pros and cons. Private solutions offer scalable, high-performance services at relatively low cost, but reproduce dependencies on large commercial cloud providers. Public solutions, although more aligned with principles of digital sovereignty, face limitations due to a lack of funding and skills, and are typically viable only amid modest participation. As for common-based solutions, they are largely absent from online political participation due to a lack of investment or because extant community-based solutions are not adapted to the resource demands of large platforms. The promoters of these tools must thus constantly navigate a tension between ideals of openness and autonomy (when present) and constraints of funding, staffing, and political will. Infrastructure is a site where power is exercised through procurement decisions and hosting arrangements. In this sense, I build upon and contribute to what Musiani et al. (2016) termed the “turn to infrastructure” in internet governance studies. This analytical shift invites us to attend not only to front-end processes, such as agenda-setting, voting, and deliberation, but also to the sociotechnical arrangements that underpin participatory platforms.

From a more normative stance, let us return to the introduction to problematize infrastructural invisibility. Servers typically become visible when a platform crashes or slows down under pressure. Most of the time, however, they remain hidden by design. This invisibility is occasionally due to the fact that participation flows are low or sufficiently distributed to avoid overloads. In other cases, infrastructures remain invisible because technical teams deliberately select stable, scalable private systems that minimize visible disruptions. Moreover, platform promoters and developers do not communicate about the infrastructural level, but generally handle it opaquely in small groups. When breakdowns do occur, they often prompt infrastructural adaptations or upgrades, restoring the illusion of seamlessness. Thus, infrastructural controversies are rare, and the invisibility of servers remains largely unchallenged. This invisibility, in turn, contributes to the depoliticization of infrastructure in a vicious cycle that weakens public debate about the political economy of participation. For these reasons, this article calls for a re-politicization of digital infrastructures – not just among academics, but also within civic tech communities, public institutions, and participatory platform promoters. We cannot fully democratize the front-end without doing the same for the back-end. This means questioning where our platforms live, who owns their servers, how infrastructural decisions are made, and how platforms are materially sustained.

Despite the breadth of empirical material and geographic diversity of the interviewees, this research is not without limitations. First, the analysis was based on a qualitative sample of participation professionals and developers rather than a systematic census of all platforms, which may have biased the findings toward more visible cases. Second, infrastructural arrangements were often reconstructed through interviews and publicly available documentation rather than through direct access to backend systems, which may have obscured some technical details or informal arrangements. Third, while this article foregrounded server underpinnings, it did not fully analyze adjacent infrastructural issues, such as working practices and division of labor or economic and legal aspects. Future research could expand in these directions by combining ethnographic, forensic, and technical methodologies.

To conclude, three openings can be offered. First, further inquiry could examine how infrastructure affects user experience and trust in participatory processes, especially when infrastructural failures occur. Second, comparative research across different national regulatory frameworks may help clarify how data treatment laws condition infrastructural decisions. Third, scholars could explore how emerging technologies, such as generative AI or blockchain, interact with infrastructural choices in participatory platforms. Finally, more work is needed on the design and viability of infrastructure commons: Which governance models, funding strategies, and technical innovations might enable public or cooperative hosting solutions at scale? In a context of digital sovereignty conflicts, these questions are all the more important.

References

Anand, N. (2017). Hydraulic city: Water and the infrastructures of citizenship in Mumbai. Duke University Press Books.

Badouard, R. (2014). La mise en technologie des projets politiques. Une approche « orientée design » de la participation en ligne. Participations, 8(1), 31–54. https://doi.org/10.3917/parti.008.0031

Baptista, I. (2019). Electricity services always in the making: Informality and the work of infrastructure maintenance and repair in an African city. Urban Studies, 56(3), 510–525. https://doi.org/10.1177/0042098018776921

Barandiaran, X. E., Calleja-López, A., Monterde, A., & Romero, C. (2024). Decidim: A brief overview. In X. E. Barandiaran, A. Calleja-López, A. Monterde, & C. Romero (Eds.), Decidim, a technopolitical network for participatory democracy: Philosophy, practice and autonomy of a collective platform in the age of digital intelligence (pp. 1–33). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-50784-7_1

Barberà, O., Sandri, G., Correa, P., & Rodríguez-Teruel, J. (2021). Digital parties. The challenges of online organisation and participation. Springer.

Barry, A. (2020). The material politics of infrastructure. In S. Maasen, S. Dickel, & C. Schneider (Eds.), TechnoScienceSociety: Technological reconfigurations of science and society (pp. 91–109). Springer International Publishing. https://doi.org/10.1007/978-3-030-43965-1_6

Berg, S., & Hofmann, J. (2021). Digital democracy (SSRN Scholarly Paper No. 3997151). Social Science Research Network. https://papers.ssrn.com/abstract=3997151

Biondo, N., & Canestrari, M. (2018). Supernova: I segreti, le bugie e i tradimenti del MoVimento 5 Stelle: storia vera di una nuova casta che si pretendeva anticasta. Ponte alle Grazie.

Biondo, N., & Canestrari, M. (2019). Il sistema Casaleggio. Partito, soldi, relazioni: Ecco il piano per manomettere la democrazia. Ponte alle Grazie.

Blondiaux, L. (2022). Le participatif en actes: Quel avenir pour l’injonction à la participation? Questions de communication, 41(1), 73–86. https://doi.org/10.4000/questionsdecommunication.28823

Bollier, D., & Helfrich, S. (2019). Free, fair, and alive: The insurgent power of the commons. New Society Publishers.

Borge, R., Balcells, J., Padró-Solanet, A., Batlle, A., Orte, A., & Serra, R. (2018). La participación política a través de la plataforma Decidim: Análisis de 11 municipios catalanes. IX Congreso Internacional en Gobierno, Administración y Políticas Públicas GIGAPP, Madrid.

Bowker, G. C., & Star, S. L. (1999). Sorting things out: Classification and its consequences. MIT Press.

Bravo, R. B., Balcells, J., & Padró-Solanet, A. (2019). A model for the analysis of online citizen deliberation: Barcelona case study. International Journal of Communication, 13, 25.

Calabrese, T., Williams, D., & Gupta, A. (2020). Does participatory budgeting alter public spending? Evidence from New York City. Administration & Society, 52(9), 1382–1409. https://doi.org/10.1177/0095399720912548

Carnino, G., & Marquet, C. (2018). Les datacenters enfoncent le cloud: Enjeux politiques et impacts environnementaux d’internet. Zilsel, 3(1), 19–62. Cairn.info. https://doi.org/10.3917/zil.003.0019

Casilli, A. A. (2025). Waiting for robots: The hired hands of automation. University of Chicago Press.

Cellard, L., & Marquet, C. (2023). Frictions sous-marines. Revue d’anthropologie des connaissances, 17(4), Article 4. https://doi.org/10.4000/rac.31070

Charmaz, K. (2008). Grounded theory as an emergent method. In S. N. Hesse-Biber & P. Leavy (Eds.), Handbook of emergent methods (pp. 155–170). The Guilford Press.

Coleman, S., & Norris, D. F. (2005). A new agenda for e-democracy (SSRN Scholarly Paper No. 1325255). Social Science Research Network. https://doi.org/10.2139/ssrn.1325255

De Carolis, L. (2019). M5s, Sabatini: “Apriremo Rousseau a tutti con una app sulle votazioni”. Il Fatto Quotidiano. https://www.ilfattoquotidiano.it/in-edicola/articoli/2019/09/19/apriremo-rousseau-a-tutti-con-una-app-sulle-votazioni/5462499/

De Carolis, L., & Zanca, P. (2018). SOS Rousseau: Casaleggio si affida a un tecnico. Il Fatto Quotidiano. https://www.ilfattoquotidiano.it/in-edicola/articoli/2018/04/19/sos-rousseau-casaleggio-si-affida-a-un-tecnico/4301821/

DecidimFest. (2022, September 29). Decidim.Vis: A viewer of participatory processes in instances of Decidim. DecidimFest 2022 program [Conference program]. https://meta.decidim.org/conferences/DecidimFest22/program/1698.

Denis, J., & Pontille, D. (2015). Material ordering and the care of things. Science, Technology, & Human Values, 40(3), 338–367. https://doi.org/10.1177/0162243914553129

Deseriis, M. (2017). Direct parliamentarianism: An analysis of the political values embedded in Rousseau, the “operating system” of the Five Star Movement. JeDEM - eJournal of eDemocracy and Open Government, 9(2), Article 2. https://doi.org/10.29379/jedem.v9i2.471

Deseriis, M. (2020). Two variants of the digital party: The platform party and the networked party. Partecipazione e Conflitto, 13(1), Article 1. https://doi.org/10.1285/i20356609v13i1p896

Deseriis, M. (2021). Rethinking the digital democratic affordance and its impact on political representation: Toward a new framework. New Media & Society, 23(8), 2452–2473. https://doi.org/10.1177/1461444820929678

Deville, V., & Hallé, C. (2014). How to make a net party. Partido de la Red, a case study. https://github.com/DemocracyEarth/party/blob/master/HOW_TO_MAKE_A_NET_PARTY.md

Does, R. van der, & Bos, D. (2021). What can make online government platforms inclusive and deliberative? A reflection on online participatory budgeting in Duinoord, The Hague. Journal of Deliberative Democracy, 17(1). https://doi.org/10.16997/jdd.965

Domínguez Rubio, F., Denis, J., & Pontille, D. (Eds.). (2025). Fragilities. Essays on the politics, ethics, and aesthetics of maintenance and repair. The MIT Press. https://mitpress.mit.edu/9780262550758/fragilities/

Dotta, G. (2021). Rousseau: server e soluzioni della piattaforma del M5S. Punto informatico. https://www.punto-informatico.it/rousseau-server-e-soluzioni-della-piattaforma-del-m5s/

Douay, N. (2016). La numérisation des dispositifs de participation de la Mairie de Paris. Le cas du budget participatif et de la plateforme “Madame la Maire, j’ai une idée!”. Netcom. Réseaux, Communication et Territoires, 30(3/4), 249–280.

Doustaly, C. (2019). Participatory budgeting and progressive cities: Are London and Paris listening to their own voices? In M. Douglass, R. Garbaye, & K. C. Ho (Eds.), The rise of progressive cities east and west (pp. 117–136). Springer. https://doi.org/10.1007/978-981-13-0209-1_7

Dulong de Rosnay, M., & Stalder, F. (2020). Digital commons. Internet Policy Review, 9(4), 1–22.

Durand, C. (2024). How Silicon Valley unleashed techno-feudalism: The making of the digital economy. Verso Books.

Edwards, P. N., Bowker, G. C., Jackson, S. J., & Williams, R. (2009). Introduction: An agenda for infrastructure studies. Journal of the Association for Information Systems, 10(5), 6.

Epstein, D., Katzenbach, C., & Musiani, F. (2016). Doing internet governance: Practices, controversies, infrastructures, and institutions. Internet Policy Review, 5(3), 1–14. https://doi.org/10.14763/2016.3.435

Ercan, S. A., & Gagnon, J.-P. (2014). The crisis of democracy. https://doi.org/10.3167/dt.2014.010201

Fatto Quotidiano (2017). M5s, primarie in rete: difficoltà di accesso al sito. Blog: “Alta affluenza”. Chiusura delle urne posticipata alle 12. Il Fatto Quotidiano. https://www.ilfattoquotidiano.it/2017/09/22/m5s-primarie-rete-difficolta-di-accesso-al-sito-blog-alta-affluenza-chiusura-delle-urne-posticipata-alle-12/3869877/

Fatto Quotidiano (2021). Rousseau si separa dal Movimento: “Non ha saldato i debiti, nostri dipendenti in cassa integrazione”. M5s: “Piattaforma non più neutrale, ne useremo un’altra”. Il Fatto Quotidiano. https://www.ilfattoquotidiano.it/2021/04/23/rousseau-si-separa-dal-movimento-non-ha-saldato-i-debiti-nostri-dipendenti-in-cassa-integrazione-m5s-piattaforma-non-piu-neutrale-ne-useremo-unaltra/6175322/

Ferrer, M., & Parienté, J. (2019). Doublons de contributions, participants frénétiques… les limites du grand débat en ligne. Le Monde. https://www.lemonde.fr/les-decodeurs/article/2019/04/08/doublons-contributeurs-frenetiques-les-limites-du-grand-debat-en-ligne_5447578_4355770.html

Floridi, L. (2020). The fight for digital sovereignty: What it is, and why it matters, especially for the EU. Philosophy & Technology, 33(3), 369–378. https://doi.org/10.1007/s13347-020-00423-6

Franzosi, P., Marone, F., & Salvati, E. (2015). Populism and Euroscepticism in the Italian Five Star Movement. The International Spectator, 50(2), 109–124. https://doi.org/10.1080/03932729.2015.1030273

Fuchs, C. (2021). The digital commons and the digital public sphere: How to advance digital democracy today. Westminster Papers in Communication and Culture, 16(1), Article 1. https://doi.org/10.16997/wpcc.917

Gerbaudo, P. (2019). The digital party: Political organisation and online democracy. Pluto Press.

Gourgues, G. (2023). Les faux-semblants de la participation. La Vie des idées. https://laviedesidees.fr/Les-faux-semblants-de-la-participation

Graham, S., & McFarlane, C. (Eds.). (2014). Infrastructural lives: Urban infrastructure in context. Routledge. https://doi.org/10.4324/9781315775098

Greffet, F., & Wojcik, S. (2008). Parler politique en ligne. Une revue des travaux français et anglo-saxons. Réseaux, 150(4), 19–50. Cairn.info.

Grossman, L. (1995). The electronic republic. Viking Press. https://www.google.fr/books/edition/The_Electronic_Republic/vaGHAAAAMAAJ?hl=fr&gbpv=0&bsq=the%20electronic%20republic%20grossman

Hughes, T. (1983). Networks of power: Electrification in Western society, 1880–1930. Johns Hopkins University Press.

Kies, R., & Wojcik, S. (2010). European web-deliberation: Lessons from the European citizens consultation. Online Deliberation: Proceedings of the Fourth International Conference-OD2010, 198–211. https://www.academia.edu/download/30840451/Proceedings_OD2010.pdf#page=206

Knaebel, R. (2019). Grand débat : pourquoi la plateforme de consultation en ligne est vivement critiquée pour son opacité. basta! https://basta.media/grand-debat-pourquoi-la-plateforme-de-consultation-en-ligne-est-vivement.

Larkin, B. (2013). The politics and poetics of infrastructure. Annual Review of Anthropology, 42, 327–343. https://doi.org/10.1146/annurev-anthro-092412-155522

Lefebvre, R. (2018). Vers un nouveau modèle partisan? Entre déclassement des partis de gouvernement et avènement des partis-mouvements. Cahiers de La Recherche Sur Les Droits Fondamentaux, (16), 21–30.

Li Vigni, F. (2022). “Partis plateforme” versus “partis en réseau.” Analyse comparative du design de plateforme et du débat en ligne dans Rousseau et Decidim. Réseaux, 236(6), 57–93. https://doi.org/10.3917/res.236.0057

Li Vigni, F. (2026). Vie et mort de Rousseau. Conditions de possibilité d’une plateforme politique participative [Manuscript in preparation]. Centre Internet et Société, CNRS.

Li Vigni, F. (2026). Online participation. A three-dimensional approach to study digital political platforms . Internet Policy Review, 15(2). https://doi.org/10.14763/2026.2.2093

Lioy, A., Del Valle, M. E., & Gottlieb, J. (2019). Platform politics: Party organisation in the digital age. Information Polity, 24(1), 41–58.

Loucaides, D. (2020). The party decides. Logic(s), 10. https://logicmag.io/security/the-party-decides/

Mabi, C. (2017). Citizen hacker. Books & Ideas. https://booksandideas.net/Citizen-Hacker

Mabi, C. (2019). La démocratie numérique au défi de la critique sociale en France. Le Mouvement Social, 268(3), 61–79. Cairn.info. https://doi.org/10.3917/lms.268.0061

Mancini, P. (2015). Why it is time to redesign our political system. European View, 14(1), 69–75. https://doi.org/10.1007/s12290-015-0343-9

Mendez de Andes Aldama, A. (2024). Becoming-common of the public. Municipalist lessons on urban commoning as alter-planning (Publication No. 35435) [Doctoral dissertation, University of Sheffield]. https://etheses.whiterose.ac.uk/id/eprint/35435/

Moats, D., & Tseng, Y.-S. (2023). Sorting a public? Using quali-quantitative methods to interrogate the role of algorithms in digital democracy platforms. Information, Communication & Society, 1–35. https://doi.org/10.1080/1369118X.2023.2230286

Mosca, L. (2020). Democratic vision and online participatory spaces in the Italian Movimento 5 Stelle. Acta Politica, 55(1), 1–18. https://doi.org/10.1057/s41269-018-0096-y

Musiani, F. (2022). Habilitation à diriger des recherches, volume 1: Synthèse du parcours scientifique; volume 2: Mémoire scientifique “Pièces d’Internet, jeux de pouvoir: Penser la gouvernance d’Internet à partir des infrastructures”; volume 3: Recueil d’articles (Publication No. tel-03855745) [Habilitation thesis, Institut d’études politiques de Paris – Sciences Po]. https://shs.hal.science/tel-03855745

Musiani, F., Cogburn, D. L., DeNardis, L., & Levinson, N. S. (2016). The turn to infrastructure in internet governance. Springer.

Nocetti, J. (2021). L’Europe reste-t-elle une “colonie numérique” des États-Unis? Politique étrangère, Automne(3), 51–63. https://doi.org/10.3917/pe.213.0051

Päivärinta, T., & Sæbø, Ø. (2006). Models of e-democracy. Communications of the Association for Information Systems, 17(1), 37.

Parkar, K., Zérah, M.-H., & Mittal, G. (2023). Platformisation, infrastructuring, and datafication. Economic & Political Weekly, 58(14), 53–60.

Pera, M., & Bussu, S. (2024). Towards democratisation of public administration: Public-commons partnerships in Barcelona. International Journal of the Commons, 18(1), 164–176.

Petit, P. (2020). “Everywhere surveillance”: Global surveillance regimes as techno-securitization. Science as Culture, 29(1), 30–56. https://doi.org/10.1080/09505431.2019.1586866

Pianini, D., & Omicini, A. (2019). Democratic process and digital platforms: An engineering perspective. In P. Contucci, A. Omicini, D. Pianini, & A. Sîrbu (Eds.), The future of digital democracy: An interdisciplinary approach (pp. 83–96). Springer International Publishing. https://doi.org/10.1007/978-3-030-05333-8_6

Pickard, V., & Berman, D. E. (2019). After net neutrality: A new deal for the digital age. Yale University Press.

Plantin, J.-C., Lagoze, C., Edwards, P. N., & Sandvig, C. (2018). Infrastructure studies meet platform studies in the age of Google and Facebook. New Media & Society, 20(1), 293–310.

Plantin, J.-C., & Punathambekar, A. (2019). Digital media infrastructures: Pipes, platforms, and politics. Media, Culture & Society, 41(2), 163–174.

Pohle, J., & Thiel, T. (2020a). Digital sovereignty (SSRN Scholarly Paper No. 4081180). Social Science Research Network. https://papers.ssrn.com/abstract=4081180

Pohle, J., & Thiel, T. (2020b). Digital sovereignty (SSRN Scholarly Paper No. 4081180). Social Science Research Network. https://papers.ssrn.com/abstract=4081180

Pruin, A. (2022). How organizational factors shape e-participation: Lessons from the German one-stop participation portal meinBerlin. In T. Randma-Liiv & V. Lember (Eds.), Engaging citizens in policy making: E-participation practices in Europe (pp. 209–224). Edward Elgar Publishing. https://www.elgaronline.com/edcollchap-oa/edcoll/9781800374355/9781800374355.00022.xmlRandma-Liiv, T., & Lember, V. (2022). Engaging citizens in policy making: E-participation practices in Europe. Edward Elgar Publishing.

Roberts, H., Cowls, J., Casolari, F., Morley, J., Taddeo, M., & Floridi, L. (2021). Safeguarding European values with digital sovereignty: An analysis of statements and policies. Internet Policy Review, 10(3). https://doi.org/10.14763/2021.3.1575

Rodota, S. (1999). La démocratie électronique. De nouveaux concepts et expériences politiques. Apogée.

Santini, R. M., & Carvalho, H. (2019a). Online platforms for citizen participation: Meta-synthesis and critical analysis of their social and political impacts (R. M. Santini & H. Carvalho, Trans.). Comunicação e Sociedade, (36), Article 36.

Santini, R. M., & Carvalho, H. (2019b). The rise of participatory despotism: A systematic review of online platforms for political engagement. Journal of Information, Communication and Ethics in Society, 17(4), 422–437. https://doi.org/10.1108/JICES-02-2019-0016

Slota, S. C., & Bowker, G. C. (2017). How infrastructures matter. The Handbook of Science and Technology Studies, 529–554.

Srnicek, N. (2017). Platform capitalism. John Wiley & Sons.

Tarnoff, B. (2022). Internet for the people: The fight for our digital future. Verso Books.

Theviot, A. (2016). Towards a standardization of campaign strategies dictated by the Obama “model”? The case of “American-style” canvassing during the 2012 French presidential election campaign. French Politics, 14(2), 158–177. https://doi.org/10.1057/fp.2016.7

van de Donk, W. B., Snellen, I. T. M., & Tops, P. W. (1995). Orwell in Athens: A perspective on informatization and democracy. IOS Press.

van Dijck, J., Poell, T., & Waal, M. de. (2018). The platform society: Public values in a connective world. Oxford University Press.

van Dijk, J. (2020). The digital divide. John Wiley & Sons.

Varoufakis, Y. (2023). Technofeudalism: What killed capitalism. The Boadley Head.

Winner, L. (1980). Do artifacts have politics? Daedalus, 121–136.

Zuboff, S. (2019). The age of surveillance capitalism: The fight for a human future at the new frontier of power. Profile Books.

Date received: 5 February 2026

Date accepted: 11 May 2026

---

1. 1 https://policyreview.info/articles/analysis/online-participation-digital-political-platforms

2. 2 The idea that infrastructures become visible only in moments of failure, while foundational in infrastructure studies, has been increasingly questioned. Several scholars have argued that invisibility is not an inherent property but a context-dependent phenomenon, more typical of Global North countries (Anand, 2017; Baptista, 2019; Barry, 2020; Graham & McFarlane, 2014). Rather than a strict binarity between visibility in breakdown and invisibility in normal use, recent scholarship has highlighted degrees of visibility shifting attention toward the notion of taken-for-grantedness: how infrastructures become naturalized and unquestioned in everyday life (Musiani, 2022).

3. 3 Operations Tickets (OTs) are support or incident tickets created to track and manage operational issues, such as server crashes, scaling problems, updates, maintenance, or bugs in cloud services.

4. 4 https://www.cap-collectif.com/post/cap-collectif-passe-en-open-source.

5. 5 Civic tech refers to technologies designed to enhance citizen engagement, government transparency, and democratic participation. The term is also frequently used to describe the ecosystem of actors (e.g., start-ups, NGOs, public institutions, or hacker collectives) that design, promote, or implement these tools (Mabi, 2017).

6. 6 A service owned by the multinational Red Hat that is used to secure user authentication.

7. 7 A language created by Facebook, later transferred to the Linux Foundation, used to manage user queries in a distributed execution environment.

8. 8 One of the most widely used cloud services in the world, enabling data storage and analysis.

9. 9 A service for scalability and containerization.

10. 10 https://www.movimento5stelle.eu/.

11. 11 A web-based platform for version control and collaborative software development, offering Git repository hosting, issue tracking, and project management tools.

12. 12 An American company providing multiple services, including scalability, data analysis, and mapping tools.

13. 13 A German company specializing in web hosting and cloud services.

14. 14 A nonprofit organization that provides technology resources, discounted software, and digital services to civil society organizations and nonprofits worldwide.

15. 15 A cloud computing company that offers services for deploying and hosting web applications, providing continuous deployment and an optimized content delivery network.

16. 16 A cloud computing platform and service provided by Microsoft, offering a wide range of solutions, including virtual machines, AI services, and database management.

17. 17 A U.S.-based cloud infrastructure provider that offers scalable computing resources, including virtual private servers, and managed databases.

18. 18 https://info.vtaiwan.tw/.

19. 19 The programming language (Ruby) and web application framework (Rails) used to develop Decidim.

20. 20 https://sindominio.net/quienes_somos/.

21. 21 https://riseup.net/es/about-us.

22. 22 https://www.autistici.org/about.

23. 23 https://laweb.pangea.org/es/quienes-somos/.