NTT leads $500 million AI fund: All-optical network IOWN moves from R&D to commercial ecosystem

From Photonics Revolution to Capital Alliance: How NTT's IOWN AI Fund Is Reshaping AI Infrastructure

As global AI data centers struggle with power consumption and bandwidth bottlenecks, Japanese telecom giant NTT is betting on a fundamental technological paradigm shift: completely eliminating electronic signal processing in the network, allowing data to flow at the speed of light throughout the entire path. This concept, called IOWN (Innovative Optical and Wireless Network), is now accelerating toward commercialization through a dedicated fund of approximately $500 million.

According to the Yomiuri Shimbun, NTT will establish the "IOWN AI Fund" by the end of the month with an initial capital of about $500 million. Participants include regional telecom giants such as South Korea's SK Group and Taiwan's Chunghwa Telecom, policy-oriented financial institutions like the Development Bank of Japan, and core Japanese companies spanning technology and finance, including Toshiba, Sony, Fujitsu, Mitsubishi UFJ, Sumitomo Mitsui, and Mizuho. The breadth of this cross-border alliance reveals that IOWN is no longer just an internal R&D project at NTT, but has evolved into a next-generation network strategy jointly promoted by major East Asian economies.

Why All-Optical Networks Are the Inevitable Choice for the AI Era

The core concept of IOWN is to build an end-to-end photonic network. Although current backbone networks use optical fibers to transmit optical signals, every switching or processing node requires converting optical signals into electrical signals. This photoelectric conversion introduces significant latency, power consumption, and capacity limitations. IOWN aims to eliminate all electrical signal stages, using optoelectronic integration technology to perform routing, switching, and computation directly in the optical domain.

The explosion of generative AI makes this direction particularly urgent. Training a large model often requires electricity equivalent to the annual consumption of thousands of households, while the real-time response demands during inference impose stringent requirements on network latency. In 2024, NTT and Chunghwa Telecom successfully tested a 3,000-kilometer IOWN submarine link from Japan to Taiwan, achieving a one-way latency of only 17 milliseconds. This result demonstrates the overwhelming advantage of all-optical networks in long-distance scenarios.

The Fund Targets Three Technical Pillars

The investment direction of the "IOWN AI Fund" is clearly focused on three levels: optoelectronic integrated devices, AI semiconductors, and IOWN-native AI models. This is less about financial investment and more about ecosystem building—NTT needs global startups to develop dedicated chips, optical modules, and application algorithms around the IOWN architecture to form a complete solution that can replace traditional electronic networks.

Notably, the fund will simultaneously cover North America, Asia, and Europe. This reflects that IOWN's ambition is not a closed Japanese standard, but an aspiration to become a global infrastructure. The participation of SK Group and Chunghwa Telecom ensures access points in the semiconductor and manufacturing ecosystems of South Korea and Taiwan.

An Alternative Breakthrough for Japan's Technological Competitiveness### Another Breakthrough in Japan's Technological Competitiveness

In the global AI race, Japan has not taken the lead in large language models and cloud computing platforms. However, the IOWN pathway showcases a unique competitive strategy for Japan: instead of directly chasing the trends in the model layer, it starts from the underlying network architecture and reshapes the efficiency boundaries of AI infrastructure through disruptive physical layer innovation.

If IOWN succeeds, it will resolve the core contradiction of the AI era—the gap between the exponential growth in computing demand and the linear growth in energy supply. Through this fund, NTT is essentially betting on a more aggressive network evolution route than existing 5G/6G, and has gained collective endorsement from Japanese industrial capital and regional partners.

Challenges Ahead, but Momentum Has Built

The journey of all-optical networks from lab to commercial use still faces huge obstacles such as device costs, standardization, and replacement of existing infrastructure. The industry is generally cautious about the commercial timeline of IOWN. However, by establishing a dedicated fund, NTT brings large enterprises, financial institutions, and startups into the same ecosystem, significantly reducing the market risk of technology incubation. The true value of this fund lies in the creation of an "optical"-centered innovation community, allowing previously scattered optoelectronics, AI chips, and network software teams to co-evolve.

As the energy crisis of AI data centers becomes increasingly urgent, all-optical solutions like IOWN may shift from being "optional" to "essential." This step by NTT is not just about a company's strategic transformation, but may also redefine the underlying logic of global information infrastructure over the next decade.