Datacenters: Why Capital Is Moving to the Nordics and Germany

Europe's datacenter map no longer revolves only around FLAP markets. Capital is moving toward Scandinavian hydropower and Finnish district heating, while Germany retains its gravitational pull. These are two different investment theses: energy and climate in the north, demand and interconnection in Germany.

The Nordic shift is tangible. Microsoft committed $3.2 billion to Swedish cloud and AI infrastructure. In 2026, Equinix and CPP Investments acquired atNorth at an enterprise value of about $4 billion, with 1 GW of secured power for expansion. Investors are buying future power access as much as buildings.

Cold weather supports free cooling and new-site PUE targets around 1.1 to 1.2, depending on load and design. District-heating networks can monetise waste heat. But design PUE is neither measured annual PUE nor guaranteed margin.

Two Distinct Investment Cases

The Nordics offer low-carbon energy, land and cooling advantages. Germany offers enterprise demand, sovereign-cloud requirements and DE-CIX interconnection. Google's €5.5 billion programme for 2026-2029, including Dietzenbach and Hanau, demonstrates the depth of that demand.

Germany also has the most visible technical framework. Under the current EnEfG, new facilities commissioned from July 2026 must reach PUE 1.2 and staged energy-reuse requirements of 10%, 15% and 20%. Reform discussions make transaction-date legal checks essential.

Pros and Cons Beyond PUE

Nordic advantages include renewable energy, climate, heat networks and land. Drawbacks include construction costs, distance from some customers, grid queues and changing taxation. Finland moved datacenters into its general electricity-tax class on 1 July 2026, adding 2.19 euro cents per kWh according to the government.

Germany offers industrial customers and exceptional interconnection, but power prices, Frankfurt grid constraints and heat-reuse duties raise the bar. A low-latency inference site and a massive HPC training campus do not call for the same geography.

How We Apply This at FINXIA

Within TITAN DC AI, we compare the full cost per delivered megawatt: grid connection, tax, construction, cooling and commissioning delay. We also separate technically recoverable heat from heat actually sold or injected into a network.

Our brown-to-green approach favours electro-intensive brownfields only where existing infrastructure genuinely shortens delivery. Legacy power may not support high-density NVIDIA racks, and nearby district heating matters only with a credible counterparty.

Key Takeaways

The Nordics are an energy-efficiency thesis; Germany is a demand-and-interconnection thesis under regulatory pressure. Workload, latency, full power cost and connection timing should determine the choice.

Watch zonal power prices, actual grid queues, EnEfG reform and whether operators publish measured PUE and heat-reuse performance rather than design targets.