
In our research, Actions for Environmental Justice Through Autonomous and Community-Based Technological Infrastructures, we point out that, far from facilitating our communications, corporate Internet monopolies sustain extractive business models based on surveillance and data accumulation. On the other hand, autonomous and community-based technological infrastructures enable online communication while respecting privacy and maintaining a commitment to environmental sustainability and social justice. Because they do not collect data on a massive scale, these infrastructures require significantly less energy. However, the question arises: What other actions, options, and strategies can be put into practice?
Aiming to answer this question, we carried out a series of interviews with small and medium-sized internet service providers, including independent, feminist, cooperative, and alternative projects. We talked to them about their needs, ideas, options, and strategies in the process of researching and developing non-extractive digital technologies. We found that these groups implement other sustainable practices (in addition to not collecting data on a massive scale), such as reusing and repairing hardware, using open-source software, and designing lightweight websites or static pages. However, they also face economic, technological, and cultural obstacles ranging from high costs and dependence on external data centers to planned obsolescence and hegemonic corporate narratives.
Despite this, we strive to continue charting paths toward sustainability through shared infrastructure, energy-efficient technologies, territorial decentralization, and feminist and community-based approaches.
Following the publication of the research report <https://sursiendo.org/2025/07/actions-for-environmental-justice-from-autonomous-and-community-based-technological-infrastructures//>, we met again with the participating groups to share our findings and follow up on potential courses of action that will allow us to move toward the environmental and social justice we seek.[1] What we share here is a summary of the main reflections that emerged from these conversations.
The groups agreed on the need to raise awareness among users of digital services about the space these services actually occupy and the material processes that sustain what we call the “virtual.” To address this, our strategy is based on making it clear that this is not magic or an ethereal cloud, but files, processes, and servers with concrete impacts both on the material world and on our daily lives. While groups that maintain autonomous infrastructures have been pointing out this reality for years, the expansion of data centers for artificial intelligence is now bringing this narrative to a broader audience.
There was a generalized recognition of the strain that current trends in artificial intelligence place on hardware, forcing constant and costly upgrades. Some groups mention a “cascade of scrapers/bots” that scour websites and servers for massive data extraction, which overloads infrastructure, consumes more energy, and accelerates obsolescence. Experiences are shared regarding attempts to mitigate these effects using tools that block bots; the paradox is that some of these tools also rely on AI systems.
While most collectives reuse hardware, they agree that obsolescence is accelerating and that some devices can become insecure if not properly maintained. Initiatives such as the Nodo TAU cooperative in Argentina, which recycles electronic waste, are mentioned, although they face difficulties in finding markets for recycled equipment such as servers.
The discussion revealed widespread skepticism toward “green data centers.” Environmental certifications (ISO, LEED, etc.) often mask problematic practices, such as the constant replacement of hardware or the use of hydroelectric power under the label of “clean energy,” which actually conceal the material consequences mentioned earlier. The lack of independent audits and transparency was highlighted, which turns these certifications into a greenwashing mechanism comparable to misleading “organic” food labels.
It was proposed to allocate a portion of the revenue generated by hosting services to subsidize free access for environmental organizations. At the same time, it is suggested that we engage directly with social and environmental movements that still rely on corporate services, inviting them to rethink their technology choices and strengthen collaborative networks.
There was a general consensus that proprietary software is absurd and contradictory to the goals of autonomy and sustainability. Using free software is thus viewed not merely as a technical choice, but as a political stance against the logic of capitalist intellectual property.
The discussion highlighted the contradiction between some NGOs’ demands for WordPress-based solutions and the possibility of building lighter websites. Strategies for offering alternatives that reduce environmental impact without forcing users to fundamentally change their workflows were discussed. Some groups highlighted experiences with plugins that convert dynamic sites into static ones, as well as the development of custom content management systems (CMS) that facilitate the use of more sustainable technologies.
Virtualization and containerization can improve the utilization of computing resources by allowing multiple workloads to run on a single server. Under certain conditions, this can delay the purchase of new equipment and help extend the useful life of hardware, although this is not an automatic outcome.
For participating groups, this practice also has a political dimension: sharing infrastructure and avoiding the duplication of resources challenges the logic of consumption and disposal promoted by the market. On the other hand, investing in collaborative infrastructure and shared activist servers could also increase vulnerability to targeted attacks.
This measure is unattainable for most groups due to its capital intensity, legal barriers, and high operating costs. Even so, there is room to explore this possibility on a smaller scale, for instance, by renting racks in data centers. The experience of collectives managing just a few racks highlights both the possibilities offered by virtualization and the limitations imposed by a lack of technological sovereignty, geopolitical asymmetries, and dependence on services shaped by colonial dynamics.
Several groups agreed that these sources, although “better” in terms of emissions, cause serious social and ecological harm, ranging from habitat destruction to the displacement of communities. It is pointed out that the emphasis on these energy sources often serves as a mechanism for greenwashing, which obscures the root of the problem: a global model of technological consumption that perpetuates inequalities and extractivism.
There was a discussion about the importance of incorporating circular economy principles to reintroduce technological waste into other production processes. Some participants proposed exploring measures of carbon emissions linked to web access without holding individual users accountable, thereby highlighting the responsibility of the corporations and extractive chains that sustain the infrastructure:
We are estimating carbon emissions based on access logs (website visits) using the sustainable web calculation, but the applications we’ve seen focus on individual consumption (i.e., your website visit is equivalent to a hamburger, or planting a tree). We’d like to compare these estimates with the emissions generated by mining, bombings, etc., to emphasize that the problem lies elsewhere. (Sutty, pad post, 2025)
Some groups highlighted the importance of creating more physical community spaces for training and collective experimentation, recognizing that without them it is difficult to sustain shared learning processes and experiment with technologies. Although it is noted that these spaces are limited by available resources, it is also emphasized that their existence is key to strengthening the transmission of knowledge and the social fabric upon which our infrastructures are built.
The conversation also addressed the need to integrate gender and social justice perspectives throughout our processes. Technological sustainability cannot be understood in isolation from struggles for social justice. It also involves developing collective financing mechanisms, sharing knowledge, and addressing the oppressive dynamics that shape the everyday lives of communities. Beyond symbolic gestures of “inclusion,” this conversation opened up the possibility of building more concrete alliances with feminist and decolonial movements, while also raising questions about how these perspectives and alliances can contribute to making our infrastructures more sustainable.
Another interesting contribution came from groups that questioned the need for servers to be constantly available: “By definition, a server is a machine that is always available. We need to consider whether the services we host really all need to be 100% available: Could we offer services that can be shut down on demand?” This thought-provoking question opens the door to imagining intermittent services (email, cloud storage, etc.) that are activated only when needed, reducing energy consumption and challenging the capitalist logic of constant connectivity.
Finally, the groups recognized that much of this process involves transformations in values and collective imaginaries, a process that is “slow and difficult, but at the same time the most powerful.” Building more sustainable and environmentally friendly infrastructures is, therefore, not only a technical task but also a cultural and political one: it requires challenging narratives, collaborating, sharing knowledge, transforming habits, and sustaining processes of resistance across different timeframes and scales.
[1] We would like to thank the groups that participated in this conversation: Sutty, Koumbit, MariaLab, MayFirst, Maddix, Cloud68, Riseup y GreenNet.