NTT Group Launches "AIOWN" AI-Native Infrastructure for Resource Optimization and Operations Aligned with AI Utilization Advancement
NTT Group, including NTT, NTT DATA Group, and NTT DOCOMO Business, is launching "AIOWN," an AI-native infrastructure designed to provide optimal AI usage environments in response to the rapid expansion of AI applications. This initiative focuses on expanding and establishing liquid-cooled data centers to optimize GPU resources and ensure secure operations, reinforcing their leadership in the domestic data center market.
📋 Article Processing Timeline
- 📰 Published: April 27, 2026 at 22:00
- 🔍 Collected: April 27, 2026 at 13:31
- 🤖 AI Analyzed: April 28, 2026 at 04:02 (14h 30m after Collected)
NTT Corporation (Headquarters: Chiyoda-ku, Tokyo; President and CEO: Akira Shimada; hereinafter NTT), NTT DATA Group Corporation (Headquarters: Koto-ku, Tokyo; President and CEO: Yutaka Sasaki; hereinafter NTT DATA Group), and NTT DOCOMO Business Inc. (formerly NTT Communications Corporation; Headquarters: Chiyoda-ku, Tokyo; President and CEO: Katsushige Kojima; hereinafter NTT DOCOMO Business) are launching "AIOWN," an AI-native infrastructure designed to deliver optimal usage environments tailored to customer needs, against the backdrop of the rapidly expanding use of AI.
1. Advancement of AI Utilization and Sophistication of Infrastructure Requirements
With the spread of generative AI, AI utilization is rapidly expanding from general business efficiency improvements to core business and specialized tasks, and furthermore to physical AI that links with various objects like cars and robots. Along with this, AI workloads are shifting from being learning-centric to inference-centric, and a significant expansion in inference demand is expected as AI becomes integrated into corporate activities and social infrastructure.
Due to these changes, infrastructure centered on AI (AI-native infrastructure) now requires even more advanced capabilities than before. These include handling heat generation from high-density GPU racks, low-latency networks connecting distributed data centers and servers, considerations for security and data sovereignty to safely handle sensitive data, and flexible infrastructure design that extends to the edge.
NTT Group will continue to expand data centers and advance liquid cooling capabilities in response to the growing demand for AI. Furthermore, we will deploy "AIOWN," an AI-native infrastructure that optimizes resources such as GPUs, networks, and power according to customers' AI usage, realizing a secure usage environment and integrated operations, including various edge devices.
2. Status of Data Center Expansion and Advanced Liquid Cooling Capabilities
(1) Domestic Data Center Deployment Status - No. 1 Share, Operating in All 47 Prefectures -
NTT Group operates over 160 data centers across all 47 prefectures in Japan. In addition to geographical coverage, we provide an environment that can be used with low latency and security under stable power supply. Through continuous investment in data centers that support these efforts, we have established the No. 1 share in the domestic data center market.
(2) Global Top Runner in Latest AI-Ready Data Center Facilities with "Liquid Cooling Method"
With the increasing performance of GPUs for AI, the power required per rack has surged, rapidly necessitating a shift from conventional air cooling to a liquid cooling method that circulates coolant within servers. NTT's latest data centers can accommodate multiple high-performance GPU-equipped servers per rack, and liquid cooling can be utilized at the rack unit level. Moreover, the liquid cooling method can reduce cooling power consumption by up to 60% compared to air cooling. NTT is a top runner globally, providing 250MW of liquid cooling compatible facilities.
Equipment within data centers where liquid cooling can be introduced
3. Future Plans in Japan
(1) Further Expansion of Domestic Data Centers, Over 3 Times by FY2033
To meet the robust demand for data centers, NTT plans to expand its current IT power capacity from approximately 300MW to over 1GW by FY2033, more than triple the current capacity. We will also progressively expand resource management functions that allow flexible use of GPUs across multiple locations according to customer needs.
This will accelerate the provision of AI-native infrastructure supporting a wide range of applications, from AI learning to inference, and utilization in corporate systems and social infrastructure.
(2) AI-Ready Data Center with Standard Liquid Cooling in Central Tokyo
NTT DOCOMO Business has commenced construction of an AI data center in central Tokyo, designed to accommodate standard liquid cooling and high-heat server equipment, thereby strengthening its supply system for AI-ready data centers. This data center is scheduled to begin service provision in the latter half of FY2029.
This data center is ideally located in a prime urban area, approximately a 5-minute walk from a JR Yamanote Line station, offering excellent convenience for equipment setup, installation, and maintenance work. Furthermore, its proximity to connection points for various cloud providers and IX (Internet Exchange) points simplifies communication paths, enabling connection to customer sites via low-latency and highly reliable networks.
In this data center, liquid-cooled floors optimized for AI applications will be standardly equipped, housing high-performance GPUs necessary for large-scale AI model computations.
1. Advancement of AI Utilization and Sophistication of Infrastructure Requirements
With the spread of generative AI, AI utilization is rapidly expanding from general business efficiency improvements to core business and specialized tasks, and furthermore to physical AI that links with various objects like cars and robots. Along with this, AI workloads are shifting from being learning-centric to inference-centric, and a significant expansion in inference demand is expected as AI becomes integrated into corporate activities and social infrastructure.
Due to these changes, infrastructure centered on AI (AI-native infrastructure) now requires even more advanced capabilities than before. These include handling heat generation from high-density GPU racks, low-latency networks connecting distributed data centers and servers, considerations for security and data sovereignty to safely handle sensitive data, and flexible infrastructure design that extends to the edge.
NTT Group will continue to expand data centers and advance liquid cooling capabilities in response to the growing demand for AI. Furthermore, we will deploy "AIOWN," an AI-native infrastructure that optimizes resources such as GPUs, networks, and power according to customers' AI usage, realizing a secure usage environment and integrated operations, including various edge devices.
2. Status of Data Center Expansion and Advanced Liquid Cooling Capabilities
(1) Domestic Data Center Deployment Status - No. 1 Share, Operating in All 47 Prefectures -
NTT Group operates over 160 data centers across all 47 prefectures in Japan. In addition to geographical coverage, we provide an environment that can be used with low latency and security under stable power supply. Through continuous investment in data centers that support these efforts, we have established the No. 1 share in the domestic data center market.
(2) Global Top Runner in Latest AI-Ready Data Center Facilities with "Liquid Cooling Method"
With the increasing performance of GPUs for AI, the power required per rack has surged, rapidly necessitating a shift from conventional air cooling to a liquid cooling method that circulates coolant within servers. NTT's latest data centers can accommodate multiple high-performance GPU-equipped servers per rack, and liquid cooling can be utilized at the rack unit level. Moreover, the liquid cooling method can reduce cooling power consumption by up to 60% compared to air cooling. NTT is a top runner globally, providing 250MW of liquid cooling compatible facilities.
Equipment within data centers where liquid cooling can be introduced
3. Future Plans in Japan
(1) Further Expansion of Domestic Data Centers, Over 3 Times by FY2033
To meet the robust demand for data centers, NTT plans to expand its current IT power capacity from approximately 300MW to over 1GW by FY2033, more than triple the current capacity. We will also progressively expand resource management functions that allow flexible use of GPUs across multiple locations according to customer needs.
This will accelerate the provision of AI-native infrastructure supporting a wide range of applications, from AI learning to inference, and utilization in corporate systems and social infrastructure.
(2) AI-Ready Data Center with Standard Liquid Cooling in Central Tokyo
NTT DOCOMO Business has commenced construction of an AI data center in central Tokyo, designed to accommodate standard liquid cooling and high-heat server equipment, thereby strengthening its supply system for AI-ready data centers. This data center is scheduled to begin service provision in the latter half of FY2029.
This data center is ideally located in a prime urban area, approximately a 5-minute walk from a JR Yamanote Line station, offering excellent convenience for equipment setup, installation, and maintenance work. Furthermore, its proximity to connection points for various cloud providers and IX (Internet Exchange) points simplifies communication paths, enabling connection to customer sites via low-latency and highly reliable networks.
In this data center, liquid-cooled floors optimized for AI applications will be standardly equipped, housing high-performance GPUs necessary for large-scale AI model computations.