Germany Plans Major Expansion of Domestic AI Compute

• A German start‑up, Polarise, is preparing to build a 30‑megawatt AI data centre to strengthen the country’s locally operated computing capacity.
• The facility could eventually scale to 120 MW, marking a significant step toward greater European control over critical digital infrastructure.
• Rising geopolitical tensions and diverging global tech regulations are driving Europe’s push for more sovereign AI capabilities.

A Strategic Move Toward Digital Sovereignty

Germany may soon see a substantial boost in its domestically run AI computing resources. Start‑up Polarise has announced plans for a 30‑megawatt data centre in the Bavarian town of Amberg, with operations expected to begin in mid‑2027. The company said the site could later expand to 120 MW, making it one of the largest facilities of its kind in the country. This development comes as European governments seek greater control over the infrastructure that powers artificial intelligence.

Current figures from industry group Bitkom show that Germany had around 530 MW of AI data‑centre capacity at the end of last year. Much of that capacity, however, is operated by non‑German providers, raising concerns about reliance on foreign companies for critical computing needs. European policymakers have become increasingly vocal about the importance of digital sovereignty amid global tensions, trade disputes and conflicting rules on online content. The Polarise project aligns with these priorities by expanding locally controlled infrastructure.

The company already operates 13 data centres in Germany and abroad. Its new facility would significantly increase the share of domestic capacity under German ownership. Large technology firms such as Google and Amazon Web Services typically run data centres of 100 MW or more, so Polarise’s planned expansion would place it among major players in the region. The project has not been previously reported, highlighting its potential impact on Germany’s AI landscape.

Polarise declined to disclose the full investment required for the new site. The company said the final cost depends on how many customers install their own servers or rent computing power. This flexibility reflects the varied needs of organizations adopting AI technologies. It also underscores the financial complexity of building large‑scale data‑centre infrastructure, where hardware choices can significantly influence total spending.

Investment Scale and Energy Considerations

A source familiar with the project said the first phase of construction would cost in the “triple‑digit million euro range.” This estimate covers the main infrastructure but excludes the chips themselves, which represent a major portion of AI‑related expenses. The final cost will depend on the number and type of chips required by customers. High‑performance AI processors can dramatically increase the overall investment needed to bring such a facility online.

Polarise recently opened a 12‑MW data centre in Munich, which Deutsche Telekom estimated at around 1 billion euros. That facility alone doubled Germany’s locally operated capacity, illustrating how limited domestic infrastructure has been until now. The new Amberg site would continue this trend by adding significantly more power for AI workloads. These expansions reflect growing demand for compute resources across industries adopting machine learning and data‑intensive applications.

Energy consumption is a central concern for data‑centre operators, especially as oil prices have risen above $100 per barrel. AI facilities are particularly power‑hungry due to the computational demands of training and running large models. Polarise said it plans to address this challenge by partnering with WV Energie to install a dedicated wind and solar power plant for the new centre. The project will also include battery storage to help stabilize electricity supply during peak usage.

Renewable energy integration has become a key factor in the development of modern data centres. Operators face increasing pressure to reduce carbon emissions while maintaining reliable performance. The combination of wind, solar and battery systems suggests that Polarise aims to balance sustainability with operational needs. This approach may also help mitigate long‑term energy costs, which are a major factor in the economics of AI infrastructure.

Europe’s Broader Push for AI Infrastructure Control

The Polarise project fits into a wider European effort to strengthen control over digital infrastructure. Governments across the continent are concerned about geopolitical risks, including supply‑chain disruptions and conflicting international regulations. AI systems rely heavily on data‑centre capacity, making local control a strategic priority. Germany’s push for more domestically operated facilities reflects this shift toward greater technological independence.

European policymakers have also emphasized the importance of aligning AI development with regional values and regulatory frameworks. Locally run data centres can help ensure compliance with European data‑protection rules and content standards. The expansion of domestic infrastructure may also support emerging AI companies that prefer to operate within the EU’s regulatory environment. Polarise’s project could therefore play a role in strengthening Europe’s broader AI ecosystem.

The planned facility in Amberg highlights how private companies are responding to these policy trends. By investing in large‑scale infrastructure, Polarise positions itself as a key contributor to Germany’s AI ambitions. The company’s decision not to seek state subsidies suggests confidence in market demand for domestic compute capacity. It also indicates that private investment may play a significant role in Europe’s pursuit of digital sovereignty.

If the project reaches its full 120‑MW potential, it would represent a major milestone for Germany’s AI infrastructure. The expansion could help reduce reliance on foreign providers and support a more resilient technological foundation. As AI adoption accelerates across industries, the availability of local compute resources will become increasingly important. Polarise’s initiative may therefore signal the beginning of a broader shift in Europe’s approach to AI infrastructure development.

Germany’s interest in expanding domestic AI capacity mirrors similar initiatives across Europe. France and the Netherlands have also announced plans to increase local compute resources, while the European Commission is exploring ways to coordinate infrastructure investments across member states. These efforts reflect a growing recognition that AI competitiveness depends not only on software innovation but also on the physical infrastructure that powers it.