SoftBank Group will invest up to 75 billion euros (approximately 131 trillion won) to build the largest artificial intelligence (AI) data center in Europe in France. This initiative aims to expand AI computing infrastructure, which is currently concentrated in the United States, while establishing France as a new base due to its high reliance on nuclear power for stable electricity supply.
Nihon Keizai Shimbun (Nikkei) reported on May 31 that SoftBank Group plans to construct a total of 5 gigawatts of data center capacity in France. The total investment is expected to reach 75 billion euros (about 14 trillion yen), making it the largest AI data center project in Europe if all goes according to plan.
Masayoshi Son, chairman of SoftBank Group, will officially announce this plan at the investment promotion event "Choose France" hosted by the French government on June 1. This follows a request for investment made by President Emmanuel Macron during his visit to Japan from late March to early April.
This marks SoftBank's first establishment of an AI base in Europe. In the first phase, the company plans to invest 45 billion euros (approximately 8.4 trillion yen) over the next five years to build data centers exceeding 3 gigawatts in the Hauts-de-France region, including Dunkirk and Bosquel. They are also considering options to lease computing resources not only within France but also to neighboring countries.
However, not all of the 75 billion euros will come from SoftBank's direct investment. According to Nikkei, SoftBank will oversee the project while its own investment will be limited. The servers and AI semiconductors used in the data centers will be supplied directly by clients such as cloud service providers, and SoftBank will secure funding through project financing.
According to the German research firm Statista, the number of data centers in the U.S. exceeds 4,000, far surpassing other countries. The UK and Germany each have over 500, but they still lag significantly behind the U.S., while France has only about 300.
SoftBank is also building a supply chain. In partnership with French electrical equipment company Schneider Electric, it will establish a power supply facility for data centers at the port of Dunkirk. This move aims to expedite the construction of data centers by securing equipment directly.
AI data centers require substantial capital, power, cooling systems, and land. While the competition for AI has previously centered around securing advanced semiconductors like NVIDIA's graphics processing units (GPUs), the focus is now shifting to where and how these semiconductors will be powered.
The choice of France is also influenced by its stable electricity supply due to a high proportion of nuclear power. Additionally, the cost of building data centers in France is lower than in the U.S. The French government announced last February as part of its "AI National Strategy" that it would invest over 100 billion euros in public and private funds in the AI sector over several years. There is also a growing movement to collaborate with India and Global South countries in AI development and utilization, aiming to establish a distinct "third pole" in AI separate from the U.S. and China.
AI computing resources are not merely private facilities; they are directly linked to national competitiveness. Without large-scale data centers in their own countries or regions, businesses and research institutions have no choice but to rely on the infrastructure of American cloud companies, making it imperative for Europe to secure its own AI infrastructure for data sovereignty and industrial competitiveness.
U.S. tech giants also view France as a promising investment destination. Amazon Web Services (AWS) plans to invest over 15 billion euros in AI infrastructure from 2026 to 2028, while Microsoft has announced a 4 billion euro investment in AI-related fields in 2024.
Son's AI investments are also ongoing in the U.S. SoftBank is planning to build a data center in Ohio worth 80 trillion yen. However, as SoftBank continues to announce large-scale investments in AI, its debt is increasing, raising concerns that if AI demand does not grow as expected, recovering investments may become challenging.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.