Research Diary: Following Meta’s Data Centre in Talavera de la Reina
As part of Rooted Clouds, our effort tracing the planetary justice impacts of AI data centres, this diary follows my case study on Meta's planned data centre campus in Talavera de la Reina (Spain), documenting what I found when trying to assess its potential environmental, societal, economic and health impacts.
This phase of the research consisted in filling a structured spreadsheet developed for this project, based on our analytical framework and the AI Supply Chain Impact Framework. The spreadsheet requires input across multiple categories: land use, water extraction, energy demand, grid interaction, regulatory approvals, political framing, and local responses.
It is important to underline that the data centre campus in Talavera de la Reina has not yet been constructed. However, the project is currently in its final administrative phase, specifically undergoing the land reparcelling process, with the stated objective of initiating construction works in the near term. While its impacts are therefore still projected rather than materialised, the scale of the proposed installation allows for a reasoned assessment of its likely implications.
Why Talavera de la Reina?
Choosing where to begin felt important. Among the various data centre projects I could have followed, Talavera de la Reina stood out for a specific reason: it sits at the intersection of several pressures that make it analytically significant and, I would argue, urgently worth examining.
Large-scale data centre developments have increasingly raised concerns in regions facing water stress, energy transition pressures, and uneven territorial development. In contexts such as central Spain, where drought conditions are intensifying and rural regions face economic vulnerability, the siting of hyperscale data centres introduces questions about resource allocation, distributive justice, and long-term sustainability.
Talavera de la Reina sits precisely within this context: a historically underinvested region, ecologically pressured, and now the proposed home for one of the largest data centre campuses in Spain. It is within this broader frame that the case becomes analytically significant.
Beginning the document search: Environmental Impact Assessments and planning approvals
My starting point was the Environmental Impact Assessment (EIA) submitted as part of the project’s approval process. The EIA, published through Castilla-La Mancha’s regional environmental authority, is one of the primary regulatory instruments governing large infrastructure projects in Spain.
The environmental assessment confirms several key characteristics of the project:
The data centre campus will occupy approximately 191 hectares of land on the outskirts of Talavera de la Reina.
The projected annual water consumption exceeds 500 million litres, sourced primarily from the Tagus River basin.
The facility will rely on large-scale electricity supply, with reported installed capacity estimates of approximately 248 megawatts, placing it among the largest planned data centre installations in Spain.
However, accessing and interpreting this information required navigating multiple separate documents. The EIA was not available as a single consolidated report, but rather as a collection of technical annexes, environmental studies, and administrative resolutions published across regional government portals.
This fragmentation made it difficult to reconstruct a comprehensive picture of the project’s environmental footprint.
For example, water consumption figures were presented in isolation, without comparison to regional water stress indicators, drought projections, or competing water demands. Similarly, while cooling systems were described in technical terms, the assessment did not clearly contextualise their operational implications under scenarios of prolonged drought, an increasingly relevant condition in central Spain.
The Proyecto de Singular Interés: Fast-tracking strategic infrastructure
The project was approved in 2023 under Castilla-La Mancha’s Proyecto de Singular Interés (PSI) framework. PSI designation allows regional authorities to accelerate planning approval for infrastructure considered strategically important for economic development.
Reviewing the PSI documentation provided insight into how the project was framed politically.
The PSI dossier emphasises several anticipated benefits:
Regional economic investment
Technological development
Job creation during construction and operation
These arguments align with broader political narratives positioning data centres as essential infrastructure for digital transformation and economic competitiveness.
While examining official planning documents, I also began tracing how the project was discussed within public political discourse.
Meta’s investment in Talavera emerged during a period of intense policy focus on digital infrastructure across Europe, framed under the European Union’s digital and green transition agenda. Public announcements by regional and national officials repeatedly described the project using familiar language: sustainability, innovation, and digital leadership.
This framing mirrors broader EU-level narratives positioning digital infrastructure as essential to economic resilience, technological sovereignty, and climate transition.
At the regional level, the Government of Castilla-La Mancha presented the Talavera data centre as a major economic opportunity, particularly for a region historically affected by industrial decline and rural depopulation.
“We are talking about hundreds of jobs. The most important thing is that there are jobs, that there are companies, because that generates wealth and that generates taxes,” said Emiliano García-Page, president of Castilla-La Mancha, in 2025.
These announcements emphasised investment and modernization, while environmental considerations were largely addressed through reference to regulatory compliance rather than detailed public explanation.
Media coverage and institutional communication
To complement official documentation, I reviewed national and regional media coverage, including reporting in outlets such as El País, Business Insider, and regional Castilla-La Mancha newspapers, as mentioned before.
Here, I encountered something notable. Local and national media have already reported many of the concerns that the spreadsheet seeks to identify:
Water extraction levels
Energy demand in a region facing grid transition challenges
Questions around employment promises versus automation realities
Tensions between rural hosting territories and global digital markets
In other words, the warnings and potential threats are not absent from the public sphere.
However, what appears largely absent is a substantive public institutional response beyond references to regulatory compliance and mitigation measures outlined in the EIA.
Government communications consistently emphasise that the project meets environmental standards and will operate sustainably. Yet I did not find evidence of expanded public debate, additional hydrological modelling made accessible, or significant revision of project parameters in response to these warnings.
Encountering local responses: territorial perspectives
As I continued researching the project, I became aware of the emergence of local and national civil society responses, including the collective Tu nube seca mi río (“Your cloud dries my river”), which currently appears to be one of the most visible and active movements in Spain addressing the expansion of data centres, and increasingly engaging at the European level.
The collective was formed directly in response to data centre developments in Spain, including the Talavera project. Their concerns focus on water use, territorial justice, and the uneven distribution of infrastructural impacts between urban technology users and rural host regions.
Following their work helped situate the Talavera case within a broader national and global context. Data centre expansion is not occurring randomly; rather, it follows identifiable patterns. These projects are often located in regions where land and water are available, and where political incentives align with infrastructure investment, such as Castilla-La Mancha or Aragón in Spain.
According to the collective’s latest report, many of these projects are being developed in what they describe as “sacrifice zones,” which “tend to be among the most vulnerable places within their respective countries, and are typically where these large-scale projects are located, projects that damage nature, local history, and communities.”
Research challenges: reconstructing infrastructure through fragments
One of the most significant challenges in this research phase was not the absence of information, but its dispersion.
Relevant information was spread across:
Environmental assessment annexes
Planning and PSI approval documents
Governmental communications
Press coverage and corporate announcements
Each source provided partial insight, but none offered a complete, accessible and simple overview of the project’s material and environmental implications.
In addition to fragmentation, the highly technical nature of the documentation posed a further challenge. Environmental impact assessments rely on specialised terminology, hydrological modelling references, engineering specifications, and regulatory jargon that require interpretation. While such technical language is standard in infrastructure planning, it creates an additional barrier for public scrutiny and comparative analysis.
In this sense, the research process itself became a form of infrastructure mapping, not only charting land, water, and energy flows, but also navigating the bureaucratic and technical layers that shape how these flows are documented and made visible.
Although the data center in Talavera de la Reina has not yet been built, one can glimpse the complexity involved in a project of this kind: bureaucratic entanglements, language that is complex and difficult to digest for anyone interested, grandiose promises in a region ravaged by neglect, and a community that hopes that progress will solve its problems.
Next steps
The next step in this research will be to develop Talavera de la Reina into a full case study. For this purpose, I will conduct interviews with residents and community members.
These interviews will allow the spreadsheet-based impact analysis to be complemented by lived experience, territorial knowledge, and governance perspectives that are not captured in formal documentation.
The goal is to understand:
How the project is perceived locally
Whether affected communities feel adequately informed and consulted
How water allocation and territorial development are discussed at the municipal level
What accountability mechanisms are visible, or absent, from the ground
We are excited about what is coming next, and this case study is just one piece of a wider effort. Rooted Clouds is mapping AI data centres not only by their location, but by their planetary justice footprints, covering water use, energy sources, land occupation, and the communities living alongside these infrastructures. If you want to follow the program, read more about it, or get involved, you can learn more about Rooted Clouds on the project page.
