Summary

• Wirth Gruppe is developing the Engrida datacentre campus at Philippsburg, using electricity and heat infrastructure as part of the proposition.
• The project is tied to a rare high voltage connection and a former industrial and nuclear energy site.
• It shows how AI infrastructure is becoming a local planning, energy, sovereignty, and industrial redevelopment issue.

Engrida, a datacentre campus project backed by Germany’s Wirth Gruppe, is being developed at Philippsburg in Baden-Württemberg with power, grid, and heat infrastructure built into the proposition.

The project is planned for an industrial site linked to Philippsburg’s former energy economy, including the area around the former nuclear power plant and the former Goodyear works. Wirth Gruppe and the city of Philippsburg have presented plans for a hyperscale datacentre campus with a direct 380kV connection, while sector materials put the agreed grid connection capacity at 707MW.

The pitch reaches beyond server halls. Engrida is being described as a platform combining energy, grid, and data, with waste heat intended for local heating. Public materials refer to Germany’s need for more datacentre capacity and to Philippsburg’s transition from a power location to a data location.

That framing is politically useful and commercially important. Datacentre developers across Europe are competing less on land alone and more on time to power. A site with credible high voltage access, potential heat reuse, and local political engagement can be far more valuable than a cheaper plot in a congested grid area.

Compute follows the power

AI is changing datacentre geography. Earlier cloud regions tended to cluster around connectivity hubs, financial centres, and established metropolitan markets. Those factors still matter, but large scale AI training and high density compute have made power availability a first order constraint. Former industrial sites, energy corridors, and substations are becoming part of the digital economy.

Philippsburg fits that pattern. A former nuclear energy area carries grid infrastructure that is difficult to replicate quickly, while nearby industrial land can support large physical development. The project’s heat reuse argument also reflects pressure on datacentre operators to show local benefit beyond tax income and construction jobs.

Germany has additional reasons to care. Its manufacturing base, research institutions, automotive sector, and enterprise software ecosystem all need access to AI infrastructure, but the country has to balance digital sovereignty ambitions with energy costs, grid constraints, local opposition, and environmental scrutiny. Datacentres that appear to consume scarce power without returning local value will face harder politics.

Engrida’s execution risk remains substantial. Large datacentre proposals depend on financing, customers, power agreements, permits, construction supply chains, and community consent. The project is seeking investors, with preference expressed for German or European backing. That will test Europe’s appetite for domestic AI infrastructure ownership, rather than simply hosting capacity for global hyperscalers.

The project also raises a recurring sovereignty question. A European-owned facility using German grid assets and returning heat to local communities is one model of digital infrastructure. A site ultimately leased or sold to global cloud or AI companies is another. Both may increase compute capacity, but they create different outcomes for control, tax, procurement, and bargaining power.

Philippsburg shows that Europe’s AI infrastructure race is moving through municipal councils, grid operators, industrial land banks, and heat networks. AI policy may be written in capitals, but the capacity to run it is being negotiated one substation and planning file at a time.