Contesting the Cloud
Infrastructure, Power, and Democratic Alternatives in AI
In March 2026, drone strikes damaged Amazon Web Services data centers in the United Arab Emirates and Bahrain, disrupting cloud services and making plain the fact that digital infrastructure is physical, strategic, and exposed. What becomes visible, when reading these introductory essays by 2026/2027 ICDE fellows, is a struggle over who builds, governs, and ultimately benefits from digital infrastructure. This cohort brings an interdisciplinary lens to questions of AI governance, labor and platform power, finance and scaling, cooperative infrastructure, and digital political economy.
Hidden Labor, Visible Harm
At one end of the AI landscape is the human cost that until recently, has been remains largely hidden. In The Silent Toll of AI Safety, Kauna Malgwi brings us face to face with the lived reality of content moderation. Her account cuts through abstraction. She describes the embodied labor– psychologically taxing, unevenly distributed, and largely outsourced to African workers. “For moderators,” she writes, “this manifests in several distinct ways. Hyper-vigilance sets in first: when you spend your day looking for threats online, you begin to see them everywhere offline. A crowded mall is no longer a place to shop, but a site of potential violence. Emotional numbing follows. To survive the shift, the mind builds a wall. You stop feeling the horror of the images, but that wall has no door. In time, you also stop feeling joy, intimacy, or empathy in your personal life. Then come the intrusive thoughts. The images do not stay at the office. They return in dreams, during family dinners, or in quiet moments of reflection.” Malgwi insists that analysis and recognition of this problem is not enough. Without enforceable protections, legal accountability, and structural change in the form of cooperative ownership, the system will continue to externalize harm onto those who are least able to fight it.
From there, we can ask what it would mean to redesign that system from the ground up?
Couriers as Employees
Several contributions point toward ownership as a decisive lever. Mansib Rahman’s work on last-mile delivery shows that even in a sector defined by logistical complexity and thin margins, multistakeholder cooperatives can be made to work. Restaurants and couriers are both members, couriers are directly employed, and investment is structured to provide capital without ceding control.
This concern with infrastructure ownership surfaces across geographies. In Indonesia, Iryandi Masputra describes a digital food system fractured into “too many islands,” where small farmers are left navigating disconnected platforms and weak coordination. Technology alone does not solve this fragmentation. What is required is cooperative infrastructure that grows through trust, adapts to local conditions, and values relationships over scale.
A similar grounding is visible in Shaunette Ferguson’s work in the Caribbean. NeedsMap is not simply a platform; it links communities and diasporas in moments of crisis. Here, digital tools are meaningful only insofar as they are embedded in social systems that can verify, respond, and sustain mutual aid. The technical architecture follows the social architecture, not the other way around.
Cooperative Attempts at Governing the Algorithm
At the same time, in the discussion around AI, governance remains a central tension. As Dorleta Urrutia-Onate asks: who decides when AI decides? As a member of Mondragon and a PhD student, her intervention points to the opacity of algorithmic systems and the risk that decision-making authority becomes sealed within black boxes, outside of even large cooperatives. Cooperatives, in this context, are not just alternative business models; they are potential sites for contestability. Tools like explainable AI are not merely technical add-ons. For Urrutia-Onate, they are preconditions for democratic control.
This raises a further question. Jeremias Meyer asks if large, complex cooperatives can maintain democratic vitality at scale, and whether AI , particularly large language models, might help address declining participation by surfacing collective knowledge and supporting deliberation. It is a tentative proposition. AI could either reinforce centralization or enable new forms of participation, depending on how it is governed.
Throughout, the essays resist the temptation to frame “ethical AI” as sufficient. Malu Villela argues that without supportive ecosystems such as finance, infrastructure, policy, alternative AI initiatives will remain fragile, easily absorbed or outcompeted by dominant corporate systems. Ethics, in isolation, cannot counteract structural asymmetries. Villela calls for an alignment of institutions that allows non-extractive models to endure.
Even in more speculative domains, such as blockchain, the same logic applies. Joshua Dávila suggests that crypto technologies might serve as coordination infrastructure—expanding governance possibilities, enabling new financing models, and supporting cross-border collaboration for the “Crypto Left.” Dávila’s emphasis then, is not on blockchain technology itself but on its potential to be shaped toward cooperative ends.
This insistence on crypto practices grounded in communities, is explored by Phakin Nimmannorrawong’s reflection on Bread Cooperative. What makes it “feel real” is not innovation for its own sake, but the presence of trust, the absence of speculative frenzy, and the effort to materialize a different kind of economic imagination. He writes:
At the social layer, Bread Cooperative feels real because of two reasons. First, it grows out of the trust, relationships, shared norms, and collaborative technical practices already built within Crypto Leftist communities. Second, it takes seriously democratic governance and interpersonal relationship among members, whereby voice is not simply tokenized (reduced to tradable digital votes) and trust is not naively outsourced to technology.
Workplace Democracy
Finally, Melike Kaplan’s provocation, Work Shouldn’t Be a Dictatorship, brings the conversation back to the quotidian. The design of workplaces, the distribution of voice, the possibility of participation. Her “Future Jobs Lab” in Denmark points toward a participatory redesign of work itself, where democratic principles are not aspirational but operational.
If the bombing of data centers stripped away the illusion of immateriality of the AI economy, then the question is no longer whether digital systems will be governed and owned, but by whom.
Image credits:
Top image, construction of a data center in the Emirates; second image, delivery operations for the Radish Cooperative in Quebec; third image, Bread Chain pop-up event in Berlin, 2025. All images are licensed under Creative Commons Attribution–ShareAlike–NonCommercial.
An essay by Nana Mgbechikwere Nwachukwu is forthcoming and will be added to this collection once it becomes available.