Data Center Cooperatives

From P2P Foundation Wiki
Jump to navigation Jump to search


Examples

Case Study: GAD EG, Germany

= "GAD e.G. (Gesellschaft für automatische Datenverarbeitung), a German cooperative that operated various data centers hosting infrastructure for a network of cooperative banks in Germany (Volksbanken und Raiffeisenbanken) between 1963 and 2015." [1]

* Report: COOPERATIVE DATA CENTERS: LESSONS FROM GAD EG. By Tara Merk. Institute for the Cooperative Digital Economy / Platform Cooperativism Consortium, 2026.

URL = https://ia600404.us.archive.org/32/items/merk-icde-2026/Merk_ICDE_2026.pdf

BlockchainGov:

"Dr. Tara Merk’s new report, published by the Institute for the Cooperative Digital Economy / Platform Cooperativism Consortium, examines the German cooperative GAD eG as a case study in democratic digital infrastructure. Its central insight is that power in AI is not located in data or code alone, but also in the material and institutional conditions that sustain them: buildings, servers, energy contracts, and ownership structures. The case shows that cooperative data centres can succeed when members share risk, gain efficiencies, and benefit from legal and financial advantages rooted in a trust-based culture. Yet the model remains vulnerable to rising costs, market consolidation, and demutualization pressures."

Abstract:

"Recently, various scholars have called attention to the monopolizing dynamics and harmful environmental and social externalities, which increasing big tech data center ownership causes, and which AI exacerbates. Existing research argues that alternative ownership and governance models for the computational resources housed in data centers are key to reducing our reliance on a small number of large tech companies and creating more ethical and accountable AI. Taking this analysis as my starting point, in this paper I explore cooperative data centers as an alternative to corporately owned and governed data centers. Specifically, I ask: under which conditions do cooperative data centers emerge and dissolve? I investigate my research question by conducting a historical case study of GAD (Gesellschaft für automatisierte Datenverarbeitung) eG, a German cooperative that operated various data centers hosting data processing infrastructure and services for a network of cooperative banks in Germany (Volksbanken und Raiffeisenbanken) between 1963 and 2015.

I identify four conditions enabling GAD’s emergence:

(1) collective efficiency gains,

(2) risk mitigation through joint ownership,

(3) legal and financial advantages of the cooperative form, and

(4) cultural embeddedness within a cooperative ecosystem.

Conversely, I find that the eventual dissolution of GAD was primarily due to competitive consolidation pressures and practical difficulties in maintaining cooperative legal structures in mergers. Reflecting on these findings, the paper discusses their implications for cooperative AI data centers, arguing that such models may offer viable alternatives to hyperscaler dependence, particularly for established firms in sectors likely to benefit substantially from AI, though their success hinges on stable financing, clear efficiency gains, and effective governance."


Introduction by the author:

"In this paper I aim to explore cooperative data centers. Specifically, I ask: under which conditions do cooperative data centers emerge and dissolve? To answer this question, I conduct a historical case study of GAD e.G. (Gesellschaft für automatische Datenverarbeitung), a German cooperative that operated various data centers hosting infrastructure for a network of cooperative banks in Germany (Volksbanken und Raiffeisenbanken) between 1963 and 2015. GAD eG makes for a compelling case study, as it operated successfully for over 50 years, scaling both in terms of physical locations and revenue, and adapting to technological innovation. In this sense, it is an interesting case to investigate conditions which encourage and sustain cooperative data centers. In 2015, GAD eG merged with Fiducia AG, a non-cooperative competitor, who served the same group of cooperative banks albeit in a different part of Germany. Since then, the new organization has continued its operations without a cooperative legal entity. By analyzing the factors that led GAD to decide to give up its cooperative status, this case study can also help to understand conditions that deter data centers from operating as cooperatives. Drawing on my insights into enablers and deterrents identified from the case study, I then discuss how they might translate to the case of cooperative data centers in the AI industry today. The paper is structured as follows. First, I describe data centers and cooperatives in more detail to make the hypothetical case for cooperative (AI) data centers (section 2). Next, I outline the historical case study methodology used in this paper (section 3). I then construct the case study of GAD e.G. (section 4) and report on the main findings (section 5). After discussing my findings in the context of the AI data center industry (section 6), section 7 concludes."

(https://ia600404.us.archive.org/32/items/merk-icde-2026/Merk_ICDE_2026.pdf)