Contents
In July 2025, Laís Martins and Francisco Amorim of Intercept Brasil revealed that a single data center planned by TikTok in the Brazilian municipality of Caucaia would consume more electricity each day than 2.2 million Brazilians combined. Behind the euphemisms of ‘innovation ecosystems’ and ‘green infrastructure,’ the report laid bare the anatomy of a new energy metabolism – one that quietly reorganizes entire territories to serve distributed computation and platform scalability.
Far from marking a rupture with fossil fuels, the project exemplifies how AI and digital infrastructures are now co-produced through the same colonial grammars of accumulation, invisibilization, and consent-by-omission. In Caucaia, the future is now and it’s continuity in new clothes, as we watch a new ‘architecture of progress’ built atop silence, intensification, and abstraction. As COP30 nears, the real question is not how to decarbonize computation, but how to decolonize the infrastructures that make it possible.
This essay is an adapted preview of a forthcoming strategic report developed in partnership between the Brazilian think tank Plataforma CIPÓ and the Green Screen Coalition. As the world approaches COP30 (the 30th Conference of the Parties to the UNFCCC) in November 2025, the entanglement between AI, energy infrastructures, and climate breakdown becomes impossible to ignore.
The ongoing shift in the global geopolitical landscape – marked by the rise of BRICS+ (1), the erosion of multilateralism as we knew it, and growing demands from the Global Majority – presents a decisive opportunity to reframe the international debate on technology and the future of energy. Climate change is not simply a crisis to be managed; it’s a structural manifestation of an extractive and unequal socioeconomic order.
In this context, the reorganization of the global economy’s metabolism – in both its material and financial dimensions – emerges as a central terrain of contestation. This preview distills some core reflections from the forthcoming report, mapping pathways to ensure that an innovative industrial policy and the energy transition become spaces for justice, autonomy, and collective reimagining – not new frontiers of domination.
AI and the transformation of the energy matrix: avoiding new dependencies
AI has been widely recognized as a vector for the ‘energy transition’ and the reorganization of production chains, but its implementation reflects dynamics that amplify structural inequalities and consolidate new forms of technological dependency.
The pursuit of computational scalability and the concentrated control of digital infrastructures reinforce productive asymmetries, relegating countries of the Global South to the condition of indispensable raw material suppliers, while being politically marginalized in decision-making processes. Take satellite data, for example. Monitoring deforestation in the Amazon region involves petabytes of high-resolution imagery. But turning that raw data into usable climate predictions or land-use classifications requires massive computational capacity: processing millions of images, running complex machine learning models, and storing constantly updating geospatial information.
For digitalization to contribute to equitable development trajectories, it becomes fundamental to expand multilateral governance spaces and ensure the active participation of impacted communities. The formulation of regulatory frameworks and industrial policies must be guided by principles of climate justice and technological sovereignty, ensuring that the advancement of AI is aligned with a model that redistributes innovation, democratizes access to knowledge, and promotes productive autonomy.
The regime of conservation and equivalence, instituted in the 19th century, links thermodynamic physics and political economy by consolidating a grammar capable of converting heterogeneous forms such as energy, labor, and value. A new conceptual architecture emerged at the intersection of steam engines and wage contracts, thermodynamic equations and capital flows. This regime did more than describe how energy is conserved across systems – it produced the conditions through which surplus labor could be rendered productive, heat could be translated into economic growth, and social life could be reorganized as a series of energetically efficient exchanges. By rendering transformation itself as commensurable, it enabled the consolidation of infrastructures where extraction, productivity, and obsolescence could circulate under a single logic of accumulation. What might appear as a neutral system of measurement is in fact a metaphysical device – one that continues to govern how life, matter, and territory are managed under planetary computation.
We see this logic persist in contemporary data infrastructures, where efficiency is not only a technical goal but a governing ideal. In a remote valley where lithium is extracted, solar panels shimmer under the sun while automated drones map terrain for future extraction. The energy expended in this choreography – from the sensors to the satellites to the neural networks that optimize logistics – is rarely framed as labor or ecological cost. Instead, it is rendered invisible by a system that treats all inputs as interchangeable: electrons, calories, human hours, storage bandwidth. This is the quiet continuity of the Anthropocene, where technological modernity extends the regime of equivalence into new terrains, converting the vitality of ecosystems and the exhaustion of workers into abstract units of computation.
This device establishes an ontology that authorizes the circulation and stabilization of material and symbolic flows, while also founding the legitimacy of the technical and economic operations that shape capitalist modernity.
Isabelle Stengers and Éric Alliez (1988) (2) show how this grammar does not emerge merely as a calculation tool or scientific principle, but is configured as a metaphysical technology that defines the conditions and possibilities of material processes and social relations. Equivalence, as a performative regime, guides mineral extraction, energy generation, and the expansion of technological infrastructures, composing a logic that stabilizes and reinforces extractive arrangements and establishes continuity and obsolescence as the inevitable future of sociotechnical transformations.This regime offers the framework within which the idea of ‘energy transition’ acquires consistency and diffusion, functioning as a discursive operator that organizes social expectations and guides public policies.
Jean-Baptiste Fressoz, by historicizing the emergence of this language in the context of the oil crises of the 1970s, reveals how ‘transition’ is articulated with the technocratic management of risks associated with scarcity and the instability of energy markets.
His historical analysis of energy dynamics since the 19th century shows that the process conventionally described as a ‘transition’ never operated as a sequential and progressive substitution of sources, but rather as a cumulative process, in which new sources do not supplant but overlap and reinforce the previous ones, producing continuous and exponential material expansion. The narrative of an orderly energy succession – coal, oil, gas, electricity – obscures, according to Fressoz, the simultaneity of these matrices and the active role of institutional and narrative devices in maintaining their persistence.
Climate change and sociotechnical frictions: recursive colonialism and the architecture of innovation
The climate crisis is not only a consequence of environmental degradation but also of the material architectures that sustain extractive development. The climate crisis, as a material expression of colonial-capitalist modernity, manifests not only in patterns of expropriation and environmental degradation but also in forms of knowledge that sustain the production of infrastructures and the management of planetary life.
Denise Ferreira da Silva (3) invites us to shift the framing: The emergency we face is not an isolated episode but a recursive structure. In her view, coloniality persists not only in what is extracted, but in the very way systems of knowledge isolate, segment, and manage life. This demands a deeper interrogation of how energy, infrastructure, and AI function as governance tools that both operationalize and conceal the underlying logics of dispossession.
Recursive colonialism in the governance of AI systems manifests in the reproduction of control patterns that operate under the logic of militarization, through surveillance of everything and everyone, the intensive plundering of territories and biodiversity, and systemic opacity in different components of this architecture.
These dynamics do not operate in isolation but persist across layers of infrastructure, from the extraction of minerals to the design of algorithms.
Hyper-centralized computational infrastructures concentrate power in a few states and major economic actors and also crystallize asymmetries in the production and direction of innovation. The result is a self-reinforcing model where bias and structural violence are not aberrations but outcomes of systems optimized for economic scalability and control. This architecture replicates historic inequalities even as it claims to address them through digital solutions.
Rather than aligning AI with climate justice, current models risk deepening ecological degradation and social inequality.
In this context, the instrumentalization of AI to maximize algorithmic efficiency and economic scalability tends to obscure the relationship between digitalization and socio-environmental impacts, disregarding its interdependence with ecological limits and the structural inequalities of the global economy.
Global governance of AI cannot be built on fragmented regulatory tweaks; it must confront the systems that reproduce asymmetry. Only from this perspective is it possible to prevent the digital transition from reproducing the patterns of extraction and subordination that have historically accompanied industrialization cycles. Reimagining this architecture means developing policies that go beyond ethical guidelines toward redistributive strategies that recognize data, energy, and innovation as sites of struggle over sovereignty, representation, and the material conditions of life.
Toward COP30: AI governance and climate commitments
The upcoming COP30 marks a critical moment for aligning AI governance and real climate action based on rights and racial justice. The First Global Stocktake of the United Nations Framework Convention on Climate Change (UNFCCC Global Stocktake) and the negotiations of COP30 demonstrate that the reorganization of the economy’s metabolism, in both its material and financial dimensions, is an unavoidable axis of the energy transition and digital transformation.
As the demand for computational capacity grows, the entanglement of digital infrastructure and raw-material supply chains becomes more visible.
This scenario demands a redesign of industrial policies and climate financing mechanisms, so that digitalization does not amplify structural asymmetries and geopolitical vulnerabilities, but becomes a vector of productive reconfiguration and redistribution of technological capacities.
COP30 thus emerges as not just a climate forum, but as a forum for contesting how digital futures are shaped, who defines their priorities, and who benefits from them.
The principle of common but differentiated responsibilities (CBDR), enshrined in the climate regime, provides a relevant reference for the governance of emerging technologies. Rebuilding this principle means aligning energy transitions with technological innovation under climate limits. Applying the CBDR principle to AI governance could involve reimagining the architecture of responsibility in light of technological asymmetries and historical ecological debt.
Rather than assuming a level playing field, this approach would acknowledge differentiated access to infrastructure, data, and energy, and could inform mechanisms such as equitable allocation of computational resources, financing for sovereign digital systems, or participatory standards-setting processes. Just as CBDR has shaped expectations around climate finance and capacity-building, it could offer a framework through which states with greater historical emissions and technological concentration contribute more substantively to the conditions under which AI systems are developed, governed, and distributed.
The diversification of the energy matrix, the strengthening of distributed computational infrastructures, and the adoption of open technological architectures are essential elements to mitigate environmental impacts and expand margins of autonomy.
For digitalization to be progressive, it must be part of a broader paradigm shift toward regenerative, equitable innovation rather than entrenching financial extraction.
This approach requires a political, legal, and economic framework capable of articulating technological innovation with strategies for value redistribution, considering that digital and energy infrastructures are not neutral.
Overcoming extractivism means building shared infrastructures for cooperation, respecting indigenous and local knowledge, and decentralizing governance. Rather than perpetuating innovation guided by hyperscale metrics and algorithmic control over bodies and territories, it is essential to build alternatives based on technical cooperation, distributed governance, and the appreciation of local and ancestral knowledge.
COP30 can become the stage where states and blocs like BRICS+ launch commitments not just to reduce emissions, but to re-map innovation systems. In doing so, the digital and climate agendas merge into a project that recognizes the planetary boundaries and the accumulative burden of communities long impacted by settler colonialism.
Footnotes
(1) BRICS+ refers to the expanded group of emerging economies originally formed by Brazil, Russia, India, China, and South Africa (BRICS), which now includes new members such as Egypt, Ethiopia, Iran, Saudi Arabia, and the United Arab Emirates. This broader coalition has begun positioning itself as a counterweight to Western-led climate and tech governance, proposing alternative visions for development, energy, and innovation rooted in South-South cooperation. Yet its internal contradictions are palpable. While the group articulates calls for multipolarity, climate finance reform, and digital sovereignty, its members remain deeply embedded in fossil fuel extraction and export logics.
(2) Alliez, Éric, and Isabelle Stengers. Contre-temps: Les pouvoirs de l’argent. Paris: Éditions Gérard Lebovici / Distribution Hachette, 1988.
(3) Black Feminist Tools, Critique, and Techno-poethics by Luciana Parisi and Denise Ferreira da Silva.
Lori Regattieri (they/she) is a public interest technologist focused on the intersection of a just energy transition, territorial rights, and emerging technologies. As a consultant in the philanthropic and nonprofit sector, they facilitate cross-sector collaboration and support South-South climate-justice coalitions alongside Indigenous Peoples, Afro-descendants, and local community movements.
Bridging political economy, sociotechnical systems, and design justice, Lori helps shape strategies for land rights, energy governance, and inclusive economies. A Green Screen Coalition member and former Mozilla Foundation Senior Fellow in Trustworthy AI, they hold a PhD in Communication and Culture from the Federal University of Rio de Janeiro. In 2024, the Media Ecology Association honored them with the Jacques Ellul Award for Outstanding Media Ecology Activism.