NTT Global Data Centers, part of Japan’s NTT, is expanding its capacity in response to sustained growth in demand for digital infrastructure linked to cloud and AI.
The company plans to increase its capacity to about four gigawatts via around 34 projects, including several where development has begun. Doug Adams, the subsidiary’s CEO, said the expansion could be complete inside two years. Further growth is expected to take total capacity five gigawatts in the next five years.
The company attributes the reason for its expansion to a demand for more computing power and specialised infrastructure. Adams said, “AI demand is expected to get stronger by the minute, and we’re essentially an enabler,” although the company doesn’t separate its revenue derived from AI workloads from cloud computing.
Sales at NTT Global Data Centers reached approximately $2.4 billion in the fiscal year ending March 2025, up around 30%. Adams said he expects revenue growth above 20% annually, but did not state a timeframe. The future commitment suggests current capacity is largely accounted for. More than 70% of the planned developments are ‘under contract’, although no details of these were made available.
Expansion will be concentrated in established European and APAC markets. In Frankfurt, the company is developing new facilities to provide 500 MW in addition to its existing four data centres. New sites are also under construction in Milan and Osaka, Japan. The company says it’s assessing further opportunities in the Nordic region and parts of South America.
Demand for infrastructure suited to artificial intelligence workloads has driven large capital commitments from technology companies, including a recent multi-billion-dollar agreement between Meta and Nebius. NTT Global Data Centers is one of the largest DC providers globally, ranked behind Equinix and Digital Realty.
Finance for the proposed expansion will comprise of internal revenues and external sources, with Adams stating, “we have more than $3 billion from our own balance sheet to fund expansion,” while also noting that the company may work with partners or use investment to support further growth.
The need for additional funds is obvious if the new capacity were to be entirely dedicated to AI – something that remains unclear. According to Jerome Darling of TD Cowen, the US investment bank, AI-specific data centres cost around $13 million per megawatt of capacity. Thus, providing the shells and surrounding infrastructure for 4GW of facilities would cost $52 billion over the course of the proposed expansion.
Equipping the facilities with today’s GPU racks and associated infrastructure (cooling, networking, etc.), priced by the same source at $30m per megawatt, comes with a bill of $120 billion. However, the cost of succeeding generations of GPU racks looks likely to increase that figure: The current DGX B200 rack retails at c. $500,000, and the reported pricing for next generation of AI compute, the GB200 NVL72 rack, stands at $3m per unit.
NTT made $10.45 billion in pre-tax profits [PDF] in the financial year 2024-2025. Four gigawatts of data centre-based IT load represents more than four times London’s current DC capacity.
NTT Global Data Centers currently operates more than 150 DC facilities in 20 countries. Its expansion plans suggest a demand for capacity that, if realised, will significantly expand its business and potentially, its liabilities. Any growth in demand will, of course, depend on broader economic and technological factors.
(Image source: “NTT DoCoMo Yoyogi Building 2009 cropped” by NTT_DoCoMo_Yoyogi_Building_2009.jpg: Wiiii derivative work: 0607crp (talk) is licensed under CC BY-SA 3.0.)
Want to learn more about Cloud Computing from industry leaders? Check out Cyber Security & Cloud Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information.
CloudTech News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
