When people talk about the race for artificial intelligence, they usually think of new models from OpenAI, Google, Anthropic or DeepSeek. In reality, however, the next phase of AI development is increasingly being decided elsewhere: in computing infrastructure.
Germany is therefore investing billions of euros in new data centres. As of today, more large-scale AI data centre projects are being planned or built than ever before. The objective is clear: expand domestic computing capacity, strengthen digital sovereignty and reduce dependence on the major US cloud providers.
Germany Is Rapidly Expanding Its AI Infrastructure
For many years, Germany was regarded primarily as a strong location for conventional data centres. That focus is now shifting towards dedicated AI data centres, often referred to as AI factories. Unlike traditional facilities, these are specifically designed to train large language models and power AI applications using thousands of high-performance GPUs.
Rather than concentrating investment in a single location, these projects are emerging across several federal states. New sites are being developed wherever sufficient energy supplies, robust electricity grids and suitable land are available. At the same time, the German government, regional authorities and private companies are working to establish a broader European AI ecosystem.
Brandenburg Is Becoming a Flagship Project
The largest project currently underway is in Lübbenau, Brandenburg, where Schwarz Digits, the digital division of the Schwarz Group, is building one of Europe’s largest AI data centres.
Around €11 billion is being invested in the site. The facility is being developed on the grounds of a former lignite-fired power station and is initially designed for a power capacity of around 200 megawatts. It is expected to house up to 100,000 GPUs for AI and cloud computing workloads.
What makes the project particularly noteworthy is not only its scale but also its overall concept. Waste heat generated by the facility is expected to be fed into the local district heating network, while the entire site is planned to operate using renewable energy. The project has also become a symbol of the Lusatia region’s transition from coal mining to digital technologies.
Major Projects Are Emerging Across Germany
Brandenburg is only one part of a much broader national trend.
In North Rhine-Westphalia, several large-scale developments are planned around former power station sites. Energy companies are increasingly exploring how existing electricity infrastructure can be repurposed for AI data centres, taking advantage of high-capacity grid connections already in place.
Meanwhile, Bavaria is positioning itself with the Blue Swan project as a potential home for a European AI gigafactory. The aim is to provide high-performance computing infrastructure for research, industry and applications in areas such as healthcare, mobility and security.
At the same time, numerous companies are investing in their own cloud and AI infrastructure in an effort to reduce reliance on international hyperscale providers.
Why AI Data Centres Are Different
AI data centres differ significantly from traditional facilities.
Where conventional data centres primarily host websites, cloud services or enterprise data, AI systems require enormous numbers of specialised processors. Thousands of GPUs operate simultaneously, consuming vast amounts of electricity and generating substantial heat.
As a result, issues such as power supply, cooling systems, fibre connectivity and grid infrastructure have become just as important as the AI technology itself.
Digital Sovereignty Is Becoming a Strategic Advantage
Many of these projects are driven by a broader ambition: strengthening Europe’s technological independence.
German companies such as Schwarz Digits are investing not only in data centres but also in cloud platforms, AI services and cybersecurity solutions. The long-term goal is to provide infrastructure for businesses, public authorities and research institutions while reinforcing Europe’s digital sovereignty.
The Biggest Challenges Are No Longer the AI Models
Ironically, the biggest obstacles today are not the AI models themselves but the underlying infrastructure.
Planning approvals, electricity grid connections, energy supply and access to skilled workers are increasingly determining where new AI data centres can actually be built. At the same time, there is growing pressure to meet rising energy demand in a sustainable way while making productive use of waste heat.
As a result, building AI data centres is becoming not only a technological challenge but also an industrial and energy policy priority.
The Real Competition Is Only Just Beginning
The global AI race will not be decided solely by who develops the most capable language model.
It will also depend on who can provide sufficient computing power to train and operate those models at scale. That is why not only the United States and China are investing billions in AI infrastructure—Germany is now accelerating its own efforts as well.
These projects demonstrate a clear shift in focus: away from isolated data centres and towards a national AI infrastructure. Those who build and control that infrastructure will create the foundation for the next generation of AI research, industrial innovation and economic growth.
