# Nvidia Explores Repurposing German Nuclear Sites for AI Data Centers Nvidia and PreussenElektra evaluate decommissioned nuclear facilities for future AI computing infrastructure development. - **Source:** [Latest from TechRadar](https://www.techradar.com/feeds.xml) - **Published:** 2026-10-29 - **Topic:** [Business](https://brifing.si/c/business) - **Original article:** - **On Brifing:** https://brifing.si/s/78de33ea-629a-4ca2-af57-2985819a2095 ## Strategic Site Repurposing Initiative Nvidia and PreussenElektra GmbH signed a preliminary agreement to investigate converting retired German nuclear power plants into AI data centers. This initiative focuses on utilizing existing industrial infrastructure rather than reviving nuclear fission. The project specifically targets decommissioned sites to support Germany's growing demand for high-performance computing and AI factory capacity. ## Infrastructure Advantage Analysis Former atomic stations possess unique assets including established grid connections, industrial zone approvals, and water access for cooling systems. Nvidia aims to leverage these existing foundations to potentially lower construction costs and accelerate development timelines. These sites offer ready-made industrial environments that are otherwise difficult to secure for large-scale data center projects. ## Scope of Technical Assessment The exploratory study will evaluate whether these specific locations meet the rigorous operational requirements of modern AI factories. Teams will examine grid capacity, available land, and existing infrastructure to determine feasibility. This assessment is critical for understanding if these sites can support the intense power and cooling demands of high-density computing clusters. ## Exclusion of Nuclear Power Nvidia explicitly ruled out restarting any idle nuclear reactors during this conversion process. The collaboration focuses solely on repurposing the physical land and grid links left behind after the final German reactors closed in April 2023. This approach avoids the political complexities associated with nuclear energy production within the current German regulatory framework. ## PreussenElektra Site Oversight PreussenElektra currently manages the closure and dismantling of eight atomic stations across Germany. While the specific locations under review remain undisclosed, these eight sites are the primary focus of the investigation. The company intends to conduct a sound assessment to determine how these legacy industrial assets can serve future-oriented computing needs. ## Unproven Economic Viability The project remains in an exploratory phase without disclosed funding figures, sanctioned designs, or binding building contracts. While reusing sites appears efficient on paper, neither company has confirmed that existing grid capacity matches the massive power appetite of modern AI hardware. Consequently, projected savings in time and capital remain purely speculative at this stage. ## Regulatory and Social Hurdles Conversion plans must still navigate complex approval processes for dismantled facilities and local community acceptance. The announcement does not indicate whether regional authorities or local residents have been consulted regarding these potential AI hubs. Addressing these non-technical barriers is essential before any physical transformation of the decommissioned sites can realistically begin in Germany. ## Future of AI Infrastructure This partnership highlights the industry's search for creative solutions to the unprecedented expansion of AI infrastructure. By targeting decommissioned industrial sites, Nvidia seeks to bypass the scarcity of suitable land and power connections. The success of this model depends on balancing technical requirements with the logistical realities of transforming legacy energy sites. ## Sources covering this story - https://www.techradar.com/pro/nvidia-wants-to-turn-old-nuclear-power-plants-into-huge-ai-data-centers-but-says-it-draws-the-line-at-turning-them-back-on — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._