Test report DSG-3317 · Rev B · tested October 30, 2026
AI Datacenter InfrastructureDevice under test
Nvidia and PreussenElektra to Assess Shuttered Nuclear Sites as AI Data Centers
Nvidia and PreussenElektra signed a preliminary deal on October 21 to study converting Germany's decommissioned nuclear sites into AI data centers — reusing grid, land and cooling infrastructure, not reactors.
- Read
- 3 min
- Words
- 502
- Node
- 28nm
- Operator
- Grace Kim
Spec summary
- Nvidia and PreussenElektra announced a preliminary partnership on October 21 to assess former nuclear sites as AI data centers.
- Germany shut its last nuclear plants in April 2023 and is dismantling about 30 nationwide; PEL handles eight of them.
- No investment amounts, approved plans, or construction agreements exist; the project is strictly a feasibility study.
- The plan involves no nuclear power — it targets existing grid links, land, water access and industrial infrastructure.
Nvidia and German energy firm PreussenElektra GmbH (PEL) signed a preliminary partnership on October 21 to assess whether decommissioned nuclear power plant sites in Germany can be converted into high-performance AI data centers. No money is committed, no plans approved, and — the companies stress — no nuclear power is involved.
The deal is exploratory, strictly a feasibility study. PEL, a subsidiary responsible for dismantling eight of Germany's roughly 30 nuclear plants currently being torn down, will examine whether its shuttered sites meet the operational requirements of what Nvidia calls "AI factories."
Why old nuclear sites?
The logic is infrastructure, not energy. Germany shut down its last operational nuclear plants in April 2023, following the Fukushima disaster, and is now dismantling about 30 reactors nationwide. What those sites leave behind is exactly what new data centers struggle to find.
A from-scratch build needs well-located land, industrial-zone approvals, proximity to water for cooling, and high-voltage grid connections. Nuclear plants already have all four. Repurposing them, Nvidia hopes, could significantly cut cost and development timelines at a moment when the AI infrastructure buildout is running at unprecedented scale.
"Our former power plant sites have valuable industrial infrastructure and offer opportunities for new, future-oriented uses," said Guido Knott, Chairman of PEL's Management Board. "Together with NVIDIA, we intend to conduct a sound assessment of whether and how these conditions can be leveraged for AI factories."
Knott framed the effort as pairing responsible site reuse with new economic opportunities for the surrounding regions.
What does this change for Germany's AI buildout?
Right now, nothing is final. There are no investment amounts, no approved development plans, and no definitive construction agreements. The assessment — open-ended in duration — will evaluate existing grid links, available land, and industrial infrastructure at the former sites against the strict technical requirements of AI factories.
"AI factories require access to power, land and high-capacity infrastructure, and former industrial sites could provide a strong foundation for this next generation of computing," said Markus Hacker, senior regional director at Nvidia. "Together with PreussenElektra, we will assess how its sites could meet the technical and operational requirements of AI factories and contribute to Germany's high-performance AI infrastructure."
If feasibility checks pass, the eight PEL sites under decommissioning are the likely candidates. That would give Germany a shortcut to large-scale AI compute capacity without the multi-year permitting grind that greenfield projects face.
No reactors in the equation
One point both companies make explicit: the plan does not involve using nuclear power for any future data centers. The reactors are staying offline and coming down. The value sits in the surrounding assets — grid connections, cooling water access, land, and industrial zoning — not the fission hardware.
For Nvidia, the move is one more hedge against the bottleneck constraining the entire AI industry: sites with power and connectivity that can be brought online fast. Whether German nuclear graveyards become AI factories depends entirely on what the assessment finds.
via Tom's Hardware (Source)
More from Grace Kim
Show full bio
Market editor covering regulation, policy and their impact on hardware layer of artificial intelligence at Die Signal.
501 articles
Same lot · LOT-C204
- DSG-807920nmNvidia and PreussenElektra to Assess Former Nuclear Sites for AI Data Centers
- DSG-657745nmNvidia Acts as Matchmaker to Drive Nordic AI Data Center Boom
- DSG-199314nmSWI Compute launches with 3.6GW power bank, $1bn Nvidia order
- DSG-754728nmOpenAI Backs 100,000-Nvidia-Chip Data Center in Europe
- DSG-95397nmNVIDIA and SB Energy Plan 10-GW AI Data Center Complex in Ohio