NTT Global Data Centers reports that the artificial intelligence boom is driving unprecedented demand for high density computing, with the United States absorbing most of the first wave of deployments and Europe ramping up for the next phase. Chief executive Doug Adams describes United States artificial intelligence demand as “unquenchable” and notes that Europe currently has a very small installed artificial intelligence base, particularly in direct-liquid-to-chip infrastructure, which underscores significant room for expansion. Even leading artificial intelligence developers such as OpenAI and Anthropic are tenants inside third party data centers, relying on providers like NTT to supply the powered land, supply chains, and specialized designs needed to support advanced chips and mixed artificial intelligence and cloud workloads.
Power availability emerges as the primary constraint on where new artificial intelligence campuses can be built, especially in Europe. Adams says “It’s difficult to get 100-megawatt chunks of power here in Europe” and identifies electricity access as the biggest limiting factor for new projects. NTT reserves power and secures powered land early, using bridging solutions such as gas turbines when utilities cannot immediately deliver full capacity, and Bruno Berti expects these temporary measures to be critical “for the next three to five years”. The company is planning several mega campuses, including one site in Germany that could support roughly 500 megawatts of IT load, but such developments remain rare due to uneven grid capacity and competition with industrial users, renewables, and urban growth.
Europe’s artificial intelligence buildout is fragmenting along energy, regulatory, and sovereignty lines rather than converging on a single hub. Berti says most current artificial intelligence demand is flowing to the Nordics, especially Sweden and Norway, because power is cheaper and renewable energy is abundant, although major markets such as Germany, the United Kingdom, the Netherlands, and Ireland still attract strong demand despite tightening power constraints. Data sovereignty rules are pushing more infrastructure into national markets as governments require sensitive data to remain within borders, which Adams and Berti say is creating smaller pockets of artificial intelligence infrastructure worldwide. At the same time, operators are innovating in cooling with direct immersion and direct-liquid-to-chip systems, experimenting with waste heat reuse, and reworking supply chains in response to geopolitics, with NTT planning equipment in three year increments, avoiding Chinese gear in the United States, and leveraging its scale with “hundreds and hundreds of megawatts” of equipment pre ordered to support expected annual growth above 20%.
