At the MWC Barcelona 2026 keynote, Akira Shimada (President of NTT) revealed significant progress and future prospects for the "Innovative Optical and Wireless Network" (IOWN) initiative. With the generative AI market projected to grow nearly 40-fold over the next decade, the resulting explosion in data center power consumption has become a critical global concern. Shimada pointed out the physical limitations of current AI infrastructure centered on electrical wiring, emphasizing a strategy to drastically improve power efficiency through the comprehensive introduction of optical technology without sacrificing computational performance.

The cornerstone of this vision is Photonic-Electronic Convergence (PEC), which integrates optical and electronic circuits at a high level. Optical communication possesses the physical characteristic of maintaining low power consumption even as data volume scales, a property NTT aims to apply to internal server communications and chip-to-chip interfaces. The company has set a highly ambitious goal of achieving 100 times the power efficiency of conventional systems by 2032. While first-generation devices for data center interconnects are already in practical use, the commercialization of second-generation (PEC-2) devices for rack-to-rack and board-to-board optical connections is planned for fiscal 2026. This implementation will be accelerated through collaborations with global leaders such as Broadcom and Accton Technology to build large-scale AI supercomputers.

Furthermore, the company presented a major milestone in the development of optical quantum computers, which apply light technology directly to computation. Unlike mainstream superconducting methods that require extreme cold or vacuum environments, optical quantum systems can operate at room temperature and atmospheric pressure, offering exceptional versatility and scalability. The roadmap targets a 10,000-qubit system by 2027 and 1 million qubits by 2030, with future plans to scale up to 100 million qubits. This vertical approach, unifying everything from communication networks to individual computing chips through light, is expected to be a decisive factor in creating sustainable infrastructure for the AI era.

At the MWC Barcelona 2026 keynote, Akira Shimada (President of NTT) revealed significant progress and future prospects for the "Innovative Optical and Wireless Network" (IOWN) initiative. With the generative AI market projected to grow nearly 40-fold over the next decade, the resulting explosion in data center power consumption has become a critical global concern. Shimada pointed out the physical limitations of current AI infrastructure centered on electrical wiring, emphasizing a strategy to drastically improve power efficiency through the comprehensive introduction of optical technology without sacrificing computational performance.

The cornerstone of this vision is Photonic-Electronic Convergence (PEC), which integrates optical and electronic circuits at a high level. Optical communication possesses the physical characteristic of maintaining low power consumption even as data volume scales, a property NTT aims to apply to internal server communications and chip-to-chip interfaces. The company has set a highly ambitious goal of achieving 100 times the power efficiency of conventional systems by 2032. While first-generation devices for data center interconnects are already in practical use, the commercialization of second-generation (PEC-2) devices for rack-to-rack and board-to-board optical connections is planned for fiscal 2026. This implementation will be accelerated through collaborations with global leaders such as Broadcom and Accton Technology to build large-scale AI supercomputers.

Furthermore, the company presented a major milestone in the development of optical quantum computers, which apply light technology directly to computation. Unlike mainstream superconducting methods that require extreme cold or vacuum environments, optical quantum systems can operate at room temperature and atmospheric pressure, offering exceptional versatility and scalability. The roadmap targets a 10,000-qubit system by 2027 and 1 million qubits by 2030, with future plans to scale up to 100 million qubits. This vertical approach, unifying everything from communication networks to individual computing chips through light, is expected to be a decisive factor in creating sustainable infrastructure for the AI era.