Germany's rapid expansion in artificial intelligence (AI) is encountering a serious obstacle: insufficient electricity supply to support it.
Electricity demand from German data centers is expected to surge to 35 billion to 38 billion kWh by 2030, placing unprecedented pressure on the existing grid. Germany's government and businesses are investing heavily in new data centers to avoid falling behind in the global digital transition. However, electricity supply, the most basic infrastructure requirement, has become the primary constraint on growth. German small and medium-sized enterprises face an increasingly difficult challenge, not only in securing digital competitiveness but also in gaining access to increasingly scarce electricity resources.
Germany's AI future depends entirely on a physical prerequisite: robust electricity connections. Although political and industry leaders are jointly pressing ahead with policies to expand AI, the national electricity infrastructure needed to support them faces physical limits. According to the latest research report from German IT industry association Bitkom, data center electricity demand is rising much faster than initially anticipated. Power shortages at data centers are becoming more than an energy issue. They are emerging as a key factor in location and investment decisions for the small and medium-sized enterprises that underpin German industry.
Investment in Germany's data center sector is reaching record levels. Data center-related investment in Germany is estimated to reach €23.5 billion, approximately 35.25 trillion won, in 2026, up 57% from the previous year. Of this, around €18 billion, approximately 27 trillion won, will be concentrated on high-performance servers and data storage equipment. The remaining €5.5 billion, approximately 8.25 trillion won, will be allocated to new data center buildings and advanced air-conditioning technologies.
The expansion of AI technology is the largest driver of this investment. Around 20% of Germany's total data center capacity is currently allocated to AI and high-performance computing (HPC) workloads. Bitkom expects that share to expand substantially to around 55% by 2035.
More advanced computing capabilities mean substantially higher energy consumption. Data centers and smaller IT facilities are consuming around 23 billion kWh of electricity in 2026. Drawing on expert assessments, however, Bitkom estimates that this figure will reach 35 billion to 38 billion kWh in 2030 and could rise to as much as 70 billion kWh by 2035.
The most serious challenge at present is not simply finding the funding to build advanced new facilities. The real problem is a severe shortage of infrastructure connection capacity to physically connect data centers to the grid in many parts of Germany. In a Bitkom industry survey, around 80% of data center operators and energy experts identified grid connection capacity limits as the most serious obstacle to business expansion.
Technological advances have substantially improved the energy efficiency of modern data centers compared with older facilities. However, rapidly increasing demand for training very large AI models and performing complex computations is driving electricity consumption growth that more than offsets the energy savings achieved through technological innovation.
For this reason, Germany's IT industry considers the government's data center infrastructure expansion target for 2030 effectively unattainable. The German government originally announced that it would secure around 6,000 megawatts (MW) of data center electricity connection capacity by 2030. However, projections based on the current pace of grid expansion suggest that achievable capacity is expected to be limited to 5,500 MW.
Bitkom President Ralf Wintergerst issued a strong warning that countries and businesses falling behind in building basic infrastructure, including grids and data centers, would ultimately lose their competitiveness in AI technology and business ecosystems as well.
Moreover, electricity constraints are not a problem confined to the data center industry. Across German industry, traditionally based on manufacturing, demand for electricity connections is also rising sharply as companies electrify production equipment, expand factories and introduce automation equipment to achieve carbon neutrality.
In recently announced grid measures, Germany's Federal Ministry for Economic Affairs and Climate Action officially acknowledged that there was a severe shortage of line capacity capable of supplying large amounts of electricity. The article interprets this as meaning that power-intensive AI data centers, large battery facilities that store electricity, and manufacturers building or electrifying factories are competing intensely to secure limited electricity connection rights.
In response, the German government plans to move away from its existing first-come, first-served grid connection approval process and introduce a more detailed prioritization system that gives preference to projects fully ready for immediate connection and operation.
However, considerable skepticism remains over whether this reform of government electricity allocation policy can substantially accelerate grid connections for small and medium-sized enterprises. Concerns have also been raised that large data center projects entering the capital region or particular industrial hubs could monopolize local electricity resources and crowd out other manufacturing investment.
Nevertheless, the rapid expansion of data center infrastructure also provides substantial business opportunities for technologically capable smaller German suppliers, even as it creates challenges. Reliable operation of high-performance computing equipment requires advanced liquid cooling, efficient power distribution equipment, emergency generation and uninterruptible power supplies, smart control systems and advanced building management systems. German small and medium-sized enterprises possess world-class technical expertise in these fields.
By region, data center construction is particularly active in Hesse and Brandenburg. According to Bitkom's research, around 2,000 MW of new data center capacity is currently under construction in Hesse and around 1,300 MW in Brandenburg.
However, large investment announcements do not automatically translate into benefits for German domestic businesses. The decisive factors will be how much of the procurement for advanced equipment and facilities needed to build data centers is actually awarded to German manufacturers, and whether electricity infrastructure can be completed on schedule without disruption.
Despite record investment, Germany remains in a highly uncertain position compared with the largest participants in the global market. Bitkom's long-term forecast expects installed data center capacity to reach 160 gigawatts (GW) in the United States and 90 GW in China by 2035. Germany, Europe's largest economy, is expected to reach only around 10 GW.
That does not mean it is time to conclude that German industry has lost the competition for AI competitiveness. Even if the United States and China dominate general-purpose large language models (LLMs) and extensive data center infrastructure, Germany could maintain strong global competitiveness in its traditional areas of strength: factory automation, advanced manufacturing technologies, specialized industrial applications and AI models tailored to niche markets.
However, even excellent software and industrial algorithms are of no use without the computing resources to run them and the electricity to support those resources. Ultimately, becoming a leading AI nation depends on how quickly and coherently digital infrastructure and the national grid can be expanded together. The article cautions that if electricity infrastructure construction continues to face difficulties, Germany's AI expansion could instead become an obstacle to new investment across industry.
Original Korean article: cucucu.co.kr
English brief by Korea Wind Intelligence, koreawindintel.com
