Yazılar

German Firms to Submit Separate EU Bids for AI Data Centre, Report Says

Several major German companies — Deutsche Telekom, Ionos, and the Schwarz Group’s IT subsidiary — plan to submit separate bids to the European Union for funding to build an AI data processing centre, according to Germany’s Tagesspiegel newspaper.

The European Commission has announced plans to allocate $20 billion to support the construction of AI data centres aimed at helping Europe catch up with the U.S. and China in artificial intelligence capabilities.

Under the current German government coalition agreement, Chancellor Friedrich Merz’s conservative party and the Social Democrats have prioritized having at least one of these AI centres located in Germany.

In May, Deutsche Telekom revealed it had partnered with SAP, Ionos, and the Schwarz Group to jointly seek EU funding for an “AI gigafactory” — a specialized facility designed to meet the massive computing demands of AI. However, the Tagesspiegel report noted that SAP is no longer involved in the bid.

SAP did not comment on the bidding process itself but said it is not pursuing a role as operator or investor in AI gigafactories. Instead, SAP aims to contribute as a technology and software provider to future AI data centre projects in Germany and Europe.

Ionos told Reuters that the expression of interest being submitted to Brussels this Friday is an initial step, with a formal application planned later this year alongside partners.

The Schwarz Group declined to confirm whether it will submit a separate bid, stating that if a German consortium is formed, all relevant parties will be invited to contribute to creating the fastest, most reliable, and most convincing AI gigafactory.

Deutsche Telekom did not respond to Reuters requests for comment.

Poor Grid Planning Threatens Europe’s Data Centre Hubs, Ember Report Warns

Europe’s top data centre locations, including Frankfurt, London, Amsterdam, Paris, and Dublin, risk losing their dominance unless governments improve long-term grid planning, according to a new report released Thursday by energy think-tank Ember.

The surge in demand for data centres, driven by the rise of artificial intelligence (AI) and its energy-intensive computing needs, is shifting investment priorities. Developers are increasingly choosing locations with faster and easier access to electricity, rather than remaining loyal to traditional hubs plagued by long grid connection delays.

The report warns that by 2035, up to 50% of Europe’s data centre capacity could relocate outside the current main hubs. This could divert billions of euros in economic activity to emerging markets, with significant implications for GDP and job creation. For example, data centres in Germany generated €10.4 billion in GDP in 2024 — a figure expected to more than double by 2029. Losing momentum in such a high-growth sector could harm economic prospects in these countries.

While France is likely to retain investment due to a relatively unconstrained grid, others could suffer delays of up to 13 years in connecting new data centres. The average wait time in the legacy hubs is 7–10 years, compared to only 3 years in Italy and even less in some emerging regions.

Grids are ultimately deciding where investments go,” said Elisabeth Cremona, Senior Energy Analyst at Ember. “If Europe wants to maintain its competitiveness and achieve economic growth, it must prioritise grid development.”

She emphasized that the issue extends beyond data centres to all sectors undergoing electrification. Without updated grid infrastructure, industries could struggle to scale or relocate entirely to regions with faster energy access.

Electricity demand from data centres is projected to triple in Sweden, Norway, and Denmark by 2030, and increase three- to fivefold in Austria, Greece, Finland, Hungary, Italy, Portugal, and Slovakia by 2035.

The findings highlight an urgent need for European policymakers to treat grid planning as a strategic investment tool, not just a utility service, in order to retain tech-sector leadership and support industrial transformation.

Deutsche Telekom and Nvidia Team Up to Build AI Cloud for European Manufacturers in Germany

Deutsche Telekom and Nvidia announced a strategic partnership on Friday to create an artificial intelligence cloud tailored for European manufacturers, based in Germany. Nvidia will provide 10,000 graphics processing units (GPUs) to be integrated into Deutsche Telekom’s existing data centers as part of this initiative.

The project aims for implementation by 2026 and represents a critical step toward building large-scale data centers, which are key to Germany’s push to modernize its industrial sector. This move also aligns with the European Union’s broader goal to close the AI infrastructure gap with global competitors like the U.S. and China.

Nvidia CEO Jensen Huang revealed plans for the AI cloud platform in Germany earlier this week and announced Deutsche Telekom’s participation during a meeting with German Chancellor Friedrich Merz.

Merz emphasized the significance of investments in AI infrastructure for Germany’s innovation and economic future, praising Nvidia and its partners’ commitment. He described the partnership as a vital step for strengthening Germany’s digital sovereignty.

GPUs have become essential for advancing AI capabilities. Under Chancellor Merz’s leadership, Germany plans to develop large data centers supported by state subsidies covering 35% of costs, while the industry is expected to cover the remaining 65%. The government aims to secure up to 100,000 GPUs to meet growing AI demands.

Earlier this year, the European Commission unveiled a $20 billion funding plan to accelerate AI data center construction across Europe.

Deutsche Telekom recently announced collaborations with companies like SAP, web hosting firm Ionos, and retailer Schwarz to seek EU support for building large data centers in Germany.

Nvidia is also working with European AI company Mistral to develop AI computing platforms powered by 18,000 Nvidia GPUs to serve European businesses.