A data-centre development near Munich has changed hands before construction starts, showing how electricity connections and planning approvals are becoming valuable infrastructure assets as European compute demand collides with the slower pace of grid development.

PGIM has sold the 30MVA project in Unterschleissheim to an unnamed European infrastructure investor. Its European Value Partners II fund acquired the property in 2024 and subsequently progressed the planning, environmental, and grid work required to make the site ready for development.

The scheme now has a building permit, a grid connection, and approvals under Germany’s Federal Immission Control Act, while the proposed facility targets a power usage effectiveness figure of 1.2. Neither the price nor the buyer has been disclosed.

The commercial value lies partly in what has already been removed from the buyer’s risk. In constrained European markets, securing enough electricity and navigating the planning system can take as long as — or longer than — construction itself.

Power moves into the development phase

European data-centre demand is rising through conventional cloud expansion and the much higher electrical densities associated with AI. Land around major cities remains limited, while grid operators face connection queues containing industrial electrification, housing, transport, and new generation alongside digital infrastructure.

A parcel of land may therefore appear suitable while remaining commercially unusable if a developer cannot establish when sufficient electricity will arrive. Environmental and planning approvals introduce another source of uncertainty as local authorities weigh investment against energy demand, noise, cooling, and competing uses for industrial sites.

PGIM explicitly identified the grid connection and planning consent as sources of value in the Munich transaction. The project demonstrates a development model in which an investor can acquire property, secure the difficult permissions and utility position, and exit before assuming the full construction and operating risk.

The buyer still has plenty left to resolve. Construction cost, customer commitments, equipment design, cooling, and whether the allocated power suits the density of future workloads will determine how valuable the completed facility ultimately becomes.

Munich still has strong demand

Greater Munich remains attractive because it combines large enterprise customers, research institutions, fibre connectivity, and an established technology economy. Those same characteristics make suitable infrastructure sites harder to replace once they reach an advanced permitting stage.

AI is strengthening that scarcity because the electrical profile of a modern facility depends increasingly on high-density accelerator clusters. A development designed around conventional server loads can become less commercially attractive if customers arrive asking for significantly more power per rack.

Germany is also considering much larger layers of AI compute infrastructure, adding another potential source of demand to a market already accommodating cloud and colocation growth.

The targeted PUE of 1.2 would represent an efficient design by conventional facility measures, although PUE describes overhead electricity use rather than the total environmental performance of a data centre. Water, grid impact, equipment utilisation, and the source of electricity still affect the wider footprint.

Development risk moves earlier

Infrastructure investment usually concentrates development risk around construction and leasing, but data centres move a growing share of it into power negotiations and permitting. Investors have to decide whether a location can support the computing hardware likely to be deployed several years later, not merely whether the building can be erected.

That gives projects with advanced grid positions a form of scarcity value. It also encourages specialist development strategies in which investors monetise the entitlement process itself rather than operating data centres over the long term.

The trend ties the digital economy increasingly closely to utilities and planning systems that move on different timescales from software and semiconductor product cycles. An accelerator generation can change in two years; a major transmission upgrade may require considerably longer.

The Munich transaction is modest beside Europe’s largest planned AI campuses, but it captures the shift clearly. In the most constrained data-centre markets, much of a project’s value can now be created before a server, rack, or cooling loop reaches the site.