Collaboration Begins for Domestic Deployment of "IOWN DCI Rack Solution" Based on IOWN Concept
ITOCHU Techno-Solutions (CTC), APRESIA Systems, and Edgecore Networks have partnered to promote the domestic deployment of the "IOWN DCI Rack Solution," based on NTT's IOWN concept. This collaboration aims to fundamentally improve power, heat, and latency issues in large-scale AI processing by utilizing optical technology, thereby building highly efficient next-generation data centers.
📋 Article Processing Timeline
- 📰 Published: May 11, 2026 at 19:00
- 🔍 Collected: May 11, 2026 at 10:31
- 🤖 AI Analyzed: May 12, 2026 at 06:48 (20h 16m after Collected)
ITOCHU Techno-Solutions Corporation (President & CEO: Tatsushi Shingu, Headquarters: Minato-ku, Tokyo, hereinafter: CTC), APRESIA Systems Corporation (President & CEO: Masahiko Suenaga, Headquarters: Chuo-ku, Tokyo, hereinafter: APRESIA), and Edgecore Networks Corporation (President: Mingshou Liu, Headquarters: Hsinchu City, Taiwan, hereinafter: Edgecore) have initiated a collaboration to promote the domestic deployment of the "IOWN DCI※2 Rack Solution," which addresses infrastructure challenges in the AI era based on NTT's proposed IOWN※1 concept. This initiative aims to fundamentally improve the power, heat, and latency issues encountered in large-scale AI processing with conventional electrical wiring, thereby advancing the construction of highly efficient next-generation data centers.
With the spread of generative AI and other technologies, GPU server power consumption and heat generation are increasing, posing challenges for data centers in terms of rising power demand and thermal density. Furthermore, large-scale AI processing involves massive data transfer between servers, leading to communication power consumption, latency, and heat generation. Addressing these challenges through an extension of conventional methods is difficult, necessitating a fundamental overhaul of infrastructure.
The IOWN DCI Rack Solution is a next-generation data center platform that revolutionizes traditional electrical signal-centric connections and transmissions. It leverages optical technology to achieve power-saving, high-speed, and low-latency data transmission. The solution also incorporates the concept of "Watt-Bit Collaboration," which integrally optimizes power (watts) and information (bits), contributing to improved overall power efficiency in data centers.
At the core of this solution is Edgecore's "Nexvec™" platform, which allows flexible integration of GPU, memory, and other resources required for AI processing in distributed server and data center environments. This enables an efficient optical connection structure for servers, switches, and communication elements essential for AI processing.
Furthermore, by adopting a data-centric design philosophy, the solution minimizes data transfer between computation and storage, achieving both energy efficiency and high processing performance required in the AI era.
The roles of each company in this collaboration are as follows:
CTC will be responsible for the entire process, from upstream design to construction and operation, for AI data centers, promoting the development of optimal infrastructure based on optical technology. APRESIA will specialize in supporting GPU network switches and software that enable low-latency and stable communication, as well as the optical layer for inter-DC connections. Edgecore will provide server and switch products, along with related components, that constitute the AI infrastructure.
The three companies will continue to promote the social implementation of highly efficient AI infrastructure by integrating advanced technologies and related solutions through the IOWN concept. They also aim to build next-generation infrastructure that supports Japan's AI industry and contribute to the realization of a carbon-free society.
IOWN DCI Rack Solution Image Diagram
IOWN DCI Rack Solution Image Diagram
Roles of Each Company
CTC
• Prime integrator overseeing customer proposals and the entire project
• Responsible for the design and construction of the entire system, including IOWN DCI Rack and IOWN APN
• Provision of value-added solutions, including AI infrastructure construction and facilities like liquid cooling
APRESIA
• Support for configuration and tuning of switches and Network OS composing the GPU network
• Vendor support (Level 2/3) for the above switches and NOS
• Optical layer level support by engineers proficient in optical wavelength technology and GPU communication technology
Edgecore
• Development and provision of equipment that forms the data center infrastructure utilizing IOWN
• Development and commercialization of highly efficient network/server equipment corresponding to the evolution of opto-electronic convergence devices
• Development and support of OS and management software, and technical collaboration support for domestic partners
In conjunction with this announcement, we have received the following comments:
NTT, in collaboration with partners including Taiwan's Accton/Edgecore, is advancing the development and provision of AI computing infrastructure that can flexibly utilize distributed computing resources via the IOWN optical network, aligning with the shift in AI infrastructure from large-scale training to inference processing.
We welcome this announcement as a step that will swiftly lead to the social implementation of IOWN, a next-generation communication infrastructure characterized by large capacity, low latency, and low power consumption.
Yoshiyuki Shimazu, Head of Technology Strategy and Innovation Office, Technology Planning Department, NTT Corporation
※1 IOWN (Innovat
With the spread of generative AI and other technologies, GPU server power consumption and heat generation are increasing, posing challenges for data centers in terms of rising power demand and thermal density. Furthermore, large-scale AI processing involves massive data transfer between servers, leading to communication power consumption, latency, and heat generation. Addressing these challenges through an extension of conventional methods is difficult, necessitating a fundamental overhaul of infrastructure.
The IOWN DCI Rack Solution is a next-generation data center platform that revolutionizes traditional electrical signal-centric connections and transmissions. It leverages optical technology to achieve power-saving, high-speed, and low-latency data transmission. The solution also incorporates the concept of "Watt-Bit Collaboration," which integrally optimizes power (watts) and information (bits), contributing to improved overall power efficiency in data centers.
At the core of this solution is Edgecore's "Nexvec™" platform, which allows flexible integration of GPU, memory, and other resources required for AI processing in distributed server and data center environments. This enables an efficient optical connection structure for servers, switches, and communication elements essential for AI processing.
Furthermore, by adopting a data-centric design philosophy, the solution minimizes data transfer between computation and storage, achieving both energy efficiency and high processing performance required in the AI era.
The roles of each company in this collaboration are as follows:
CTC will be responsible for the entire process, from upstream design to construction and operation, for AI data centers, promoting the development of optimal infrastructure based on optical technology. APRESIA will specialize in supporting GPU network switches and software that enable low-latency and stable communication, as well as the optical layer for inter-DC connections. Edgecore will provide server and switch products, along with related components, that constitute the AI infrastructure.
The three companies will continue to promote the social implementation of highly efficient AI infrastructure by integrating advanced technologies and related solutions through the IOWN concept. They also aim to build next-generation infrastructure that supports Japan's AI industry and contribute to the realization of a carbon-free society.
IOWN DCI Rack Solution Image Diagram
IOWN DCI Rack Solution Image Diagram
Roles of Each Company
CTC
• Prime integrator overseeing customer proposals and the entire project
• Responsible for the design and construction of the entire system, including IOWN DCI Rack and IOWN APN
• Provision of value-added solutions, including AI infrastructure construction and facilities like liquid cooling
APRESIA
• Support for configuration and tuning of switches and Network OS composing the GPU network
• Vendor support (Level 2/3) for the above switches and NOS
• Optical layer level support by engineers proficient in optical wavelength technology and GPU communication technology
Edgecore
• Development and provision of equipment that forms the data center infrastructure utilizing IOWN
• Development and commercialization of highly efficient network/server equipment corresponding to the evolution of opto-electronic convergence devices
• Development and support of OS and management software, and technical collaboration support for domestic partners
In conjunction with this announcement, we have received the following comments:
NTT, in collaboration with partners including Taiwan's Accton/Edgecore, is advancing the development and provision of AI computing infrastructure that can flexibly utilize distributed computing resources via the IOWN optical network, aligning with the shift in AI infrastructure from large-scale training to inference processing.
We welcome this announcement as a step that will swiftly lead to the social implementation of IOWN, a next-generation communication infrastructure characterized by large capacity, low latency, and low power consumption.
Yoshiyuki Shimazu, Head of Technology Strategy and Innovation Office, Technology Planning Department, NTT Corporation
※1 IOWN (Innovat