PreussenElektra And NVIDIA Will Assess Germany’s Former Nuclear Plant Sites For AI Factories

Human Verified 👁 1 0

Share to: , and

A decommissioned nuclear cooling tower beside rows of glowing AI server containers at dusk — the former plant sites PreussenElektra and NVIDIA are assessing

PreussenElektra, the E.ON subsidiary dismantling Germany’s shut nuclear power stations, will work with NVIDIA to assess whether its former plant sites can host AI factories, the companies announced on October 21. The assessment is open-ended: it matches each site’s energy infrastructure, grid connection and available land against the technical and operational requirements of high-performance AI data centres, and no investment or site decisions are attached to it.

Engineers with tablets assessing a power station site with high-voltage pylons — PreussenElektra and NVIDIA's site review for AI factories

What The Assessment Covers

PreussenElektra brings its knowledge of the sites, their infrastructure and their regulatory conditions to the study. NVIDIA, as technology partner, will assess the requirements AI factories place on a site and the technical concepts such a site could support. On that basis, the two companies will work out at which sites, and under what conditions, a more concrete project would make sense.

The logic of the sites is straightforward. Power stations offer large industrial plots, powerful grid connections and established infrastructure — the three things data-centre developers across Europe are struggling to secure. The study is a first step to evaluate the technical, infrastructural, economic and regulatory prerequisites of possible projects.

“Our former power plant sites have valuable industrial infrastructure and offer perspectives for new, future-oriented uses,” said Guido Knott, chairman of PreussenElektra’s management board, who announced the planned cooperation. “Together with NVIDIA we want to thoroughly examine whether and how these conditions can be used for AI factories.” Markus Hacker, a senior regional director at NVIDIA, said AI factories depend on a reliable energy supply, suitable land and powerful infrastructure, and that former industrial sites offer ideal conditions for them.

Eight Shut Plants, One Grid Inheritance

PreussenElektra is responsible for the decommissioning and dismantling of eight nuclear power plants in Germany. Every plant is shut — Germany closed its last reactors in April 2023 — and each sits in a different phase of dismantling. As that work progresses, the company is creating the conditions for reuse concepts and the long-term development of the sites.

Regional reporting has singled out Grafenrheinfeld in Bavaria, one of the eight sites, as a focus of local attention. Bloomberg, which reported the cooperation on October 22, noted the operators are looking for ways to redevelop the sites while dismantling of the reactors continues — a process that runs for years and leaves the grid infrastructure in place throughout.

Why It Matters

PreussenElektra’s study matters because power, not chips, is now the binding constraint on Europe’s AI buildout — and because it shows where the industry is hunting for it. Projects are already dying on grid timelines: Microsoft walked away from Nscale’s Loughton data centre in the UK after grid power slipped toward the 2030s, a reminder that a site without a firm connection is worth little. A decommissioned nuclear plant inverts that problem. The connection exists, the land is industrial, and the owner is a utility that knows how to run both. Germany shut its reactors as a political choice, not because the grid assets lost value, and its utilities are now offering that inheritance to the AI industry — with NVIDIA, which is simultaneously signing chip-supply and capacity deals across the continent, positioning itself as the technical gatekeeper for what an AI-ready site must offer.

Background

Germany completed its nuclear phase-out on April 15, 2023, ending six decades of atomic power and leaving E.ON’s PreussenElektra with eight plants to dismantle. Since then the country has become one of Europe’s most contested AI-infrastructure markets: NVIDIA deepened its German industrial partnerships this year, including a national industrial AI partnership, and locked in six chip-supply partners with $1.1 billion for new US test plants as the supply chain stretches. Microsoft’s exit from the Loughton project showed the cost of getting the power side wrong — which is precisely the problem a former power station does not have.

Source: PreussenElektra announcement, Bloomberg via Hart Energy

Share to: , and

Faheem R. Khattak
Founder & Senior Editor

Senior automation engineer and AI marketing journo covering LLMs, algorithms, analytics, core updates, Bing search, and technical SEO.

✓ 10+ Years Covering Google Search & AI

0 Comments

Leave a Reply

Optional Take Tag:

🛡️ Spam protection: Honeypot + Time-Trap Active

This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Add as Preferred Source on Google Follow on Google News