Verification of Practical Performance for a Remotely Distributed AI Infrastructure Between Tokyo and Fukuoka Utilizing IOWN APN, Based on Workload Characteristics

GMO Internet Group's GMO Internet, Inc. (Headquarters: Shibuya-ku, Tokyo; President and Representative Director: Masaru Ito; hereinafter referred to as GMO Internet), NTT East Corporation (Headquarters: Shinjuku-ku, Tokyo; President and Representative Director: Naoki Shibuya; hereinafter referred to as NTT East), NTT West Corporation (Headquarters: Osaka-shi, Osaka; President and Representative Director: Ryota Kitamura; hereinafter referred to as NTT West), and QTnet, Inc. (Headquarters: Fukuoka-shi, Fukuoka; President and Representative Director: Yoshio Ogura; hereinafter referred to as QTnet) have completed a technical demonstration of a remotely distributed AI infrastructure between Tokyo and Fukuoka utilizing IOWN (Innovative Optical and Wireless Network)'s APN (All-Photonics Network).

In this demonstration, conducted from November 2025 to February 2026, a live IOWN APN circuit was established between Tokyo (storage) and Fukuoka (GPU). The performance of an AI development platform connecting GPUs and large-capacity storage of "GMO GPU Cloud" was measured and evaluated for AI workloads. As a result, for large language model (LLM) training, a performance decrease of only approximately 0.5% was observed compared to a local environment, confirming that the impact is extremely limited. For image classification tasks involving data loading, it was confirmed that practical level processing is possible even in a remote environment through optimization of training data, etc. This demonstration proved that practical AI development is possible in a remotely distributed environment through design tailored to workload characteristics.

Prior to this demonstration, the four companies conducted a performance test in a pseudo-remote environment simulating the Tokyo-Fukuoka distance (approximately 1,000 km) as a preliminary demonstration (Phase 1) in July 2025, and have published the details as a technical report. Press Release: https://www.ntt-west.co.jp/news/2510/251002a.html Technical Report: https://www.ntt-west.co.jp/news/2510/251002a_1.html Based on the results of this demonstration, the four companies will continue to promote initiatives towards the practical implementation of remotely distributed AI infrastructure tailored to customer needs.

【Background and Objectives】 With the recent proliferation of generative AI and large language models (LLMs), the demand for AI development infrastructure has rapidly expanded. Traditionally, GPUs and large-capacity storage were required to be physically adjacent. However, to address space constraints in data centers and the need to manage data within one's own facilities, the realization of distributed AI development infrastructure that transcends geographical limitations has become necessary. The four companies have been exploring the technical feasibility of connecting GPUs and storage in remote locations by utilizing IOWN APN, which features high speed, large capacity, and low latency.