Japan – Launch of a proof-of-concept experiment for a next-generation dispersed AI processing facilities intending to understand the Watt-Bit vision

Introduce of a proof-of-concept experiment for a next-generation dispersed AI processing facilities intending to recognize the Watt-Bit vision

Developing an optical network facilities for the “KYOJO CUP” formula racing series

Tokyo, Aichi and Kawasaki, Japan, July 27, 2026 – (JCN Newswire) – Nautilus Technologies, Inc., Internet Initiative Japan Inc., TEPCO Power Grid, Inc., CHUBU Electric Power Grid, Fujitsu Limited, and 1Finity Inc., a Fujitsu business and leading company of worldwide network options, today revealed the launch of a joint proof-of-concept experiment for a next-generation dispersed AI processing facilities focused on recognizing the Watt-Bit Convergence Vision, under the New Energy and Industrial Technology Development Organization’s (NEDO) “Study on the Validation of Low-Latency Distributed Processing Technologies for Next-Generation Infrastructure”(1 ).

This proof-of-concept experiment intends to recognize the dispersed release of calculating resources pictured by the Watt-Bit Vision, a tactical structure that connects electrical power grids (“watts”and telecoms (“bits”to enhance energy usage and computing power. To this end, an All-Photonics Network (APN) (2) will be built by integrating fiber optics cable televisions owned by TEPCO Power Grid and CHUBU Electric Power Grid with interactions devices supplied by Fujitsu and 1Finity. By managing the operating timing of GPU servers (3) released at 2 geographically apart places, the experiment will examine the expediency of versatile changes in reaction to electrical power need conditions, while likewise verifying the efficiency of low-latency dispersed processing innovations for running these resources as a combined AI processing environment.

Particularly, this proof-of-concept experiment will utilize the last race of the “KYOJO CUP” formula racing series held at Fuji Speedway as its field environment. Real-time 4K video footage and telemetry information( 4) gathered from on board cams and telemetry systems( 5 )set up in race cars and trucks will be collected and transferred by means of the APN through high-bandwidth, low-latency interactions to GPU servers for AI-processing dispersed throughout the TEPCO Power Grid and Chubu Electric Power Grid service locations. This allows the gathered information to be processed by any of the GPU servers situated throughout both service locations.

With need for AI calculating anticipated to continue growing, this proof-of-concept experiment will take a look at the expediency of using low-latency dispersed processing innovations to make use of the power usage of GPU servers as a resource for electrical energy supply-demand balancing. The presentation will assess the useful applicability of these innovations towards attaining that goal.

This effort represents Japan’s very first effort( 6) to collectively run geographically dispersed GPU servers as a combined AI processing environment throughout locations with various energy frequencies (50 Hz and 60 Hz). Looking ahead, the task is anticipated to add to the awareness of Watt-Bit Convergence by making it possible for AI work to be moved to areas with surplus power capability, thus promoting more effective usage of electrical power facilities.

The proof-of-concept duration is set up to range from August 2026 to March 2027, with the proof-of-concept beginning from Round 3 of the KYOJO CUP series. Information of this proof-of-concept and its system confirmation outcomes are anticipated to be provided at market occasions, consisting of Inter BEE( 7 ), from fall 2026 onward.

Referral 1: Roles of getting involved business

The functions of each getting involved business in this proof-of-concept experiment are as follows.

Recommendation 2: Infrastructure setup of the proof-of-concept experiment

The GPU servers set up in the 2 areas are adjoined through fiber optics cable televisions offered by TEPCO PowerGrid and CHUBU Electric Power Grid, together with optical network systems supplied by Fujitsu and 1Finity, allowing them to run as a combined AI processing facilities.

Figure: Infrastructure setup of the proof-of-concept experiment

Referral 3: Key innovations used in the proof-of-concept experiment

* All-Photonics Network (APN)

The All-Photonics Network (APN) is a next-generation interactions facilities that processes info as optical signals anywhere possible. In standard interaction networks, optical and electrical signals are consistently transformed at each network gadget. APN reduces these conversions, allowing low-latency, high-capacity, and energy-efficient interactions.

For this proof-of-concept experiment, an optical Network Interface Card (optical NIC) presently under advancement by 1Finity and set up for industrial release in October 2026 will be released. By using the optical NIC, servers situated more than 100 kilometers apart can be straight adjoined utilizing optical signals without counting on traditional Layer 2 and Layer 3 switches or routers.

This technique even more decreases interaction latency while leveraging exclusive innovations to decrease efficiency deterioration over long-distance connections.

* Next-Generation High-Performance Relational Database “Tsurugi”

This proof-of-concept experiment uses Tsurugi, a next-generation high-performance relational database (RDB) established in Japan, as the core database of the system. Assistance and architecture style for the system are supplied by Nautilus Technologies, Inc.

Tsurugi is an open-source next-generation relational database established in Japan with assistance from the New Energy and Industrial Technology Development Organization (NEDO). It is created to provide high deal processing efficiency and scalability while offering an end-to-end domestic assistance and production structure.

Tsurugi is likewise being improved to support future AI-driven applications. In addition to supporting AI combination innovations such as the Model Context Protocol (MCP) (9 ), it supplies a system for flawlessly using remote GPUs straight from the database through user-defined functions (UDFs) (10 ).

These abilities allow effective combination in between the information management platform and the AI processing facilities, adding to the awareness of low-latency AI processing in this proof-of-concept experiment.

(1) Survey on the efficiency of low-latency dispersed processing innovations for next-generation facilities

This study is being performed as part of efforts to take full advantage of the results of NEDO’s Technology Development Project for AI Chips and Next-Generation Computing for High-Efficiency and High-Speed Processing.

The study intends to gather info on the most recent advancements in low-latency dispersed processing innovations, which are significantly acknowledged as an essential innovation for next-generation details facilities. It likewise looks for to determine usage cases in which such innovations are anticipated to be more efficient than standard interactions innovations and to examine their efficiency through field confirmation. This proof-of-concept experiment is being carried out as part of those field confirmation activities.

(2) A next-generation interactions facilities that processes details utilizing optical signals to the best level possible throughout information transmission.

(3) A high-performance server geared up with Graphics Processing Units (GPUs) for massive computational work such as AI and artificial intelligence.

(4) Driving information such as car speed, engine, speed accelerator position, brake operation, and equipment position.

(5) An onboard measurement gadget that gathers and transfers information connected to automobile operation and habits.

(6) Based on research study performed by the taking part business since July 2026.

(7) Inter BEE: One of Japan’s biggest exhibits for transmitting devices, video and audio innovations, and content circulation services.

(8) Optical Network Interface Card (Optical NIC): A hardware part that allows network connection and information transmission/reception. By straight processing optical signals, it assists in long-distance, ultra-high-speed interaction while enhancing network efficiency and effectiveness.

(9) A basic procedure that makes it possible for AI systems to firmly and effectively gain access to and make use of info by linking AI with databases, service systems, and other information sources.

(10) User-Defined Functions (UDFs) are functions that users can individually specify and contribute to extend database abilities, making it possible for databases to conjure up and use external systems and computational resources such as GPUs.

About IIJ

Established in 1992, IIJ is among Japan’s leading Internet-access and extensive network services service providers. IIJ and its group business supply overall network services that primarily deal with high-end business clients. IIJ’s services consist of top quality Internet connection services, systems combination, cloud computing services, security services and mobile services. IIJ has actually developed one of the biggest Internet foundation networks in Japan that is linked to the United States, the United Kingdom and Asia. IIJ was noted on the Prime Market of the Tokyo StockExchange in 2022. To find out more about IIJ, check out the IIJ Web website at https://www.iij.ad.jp/en/.

About Fujitsu

Fujitsu’s function is to make the world more sustainable by constructing rely on society through development. As the digital improvement partner of option for consumers around the world, our 100,000 staff members work to deal with a few of the best obstacles dealing with humankind. Our series of services and services make use of 5 essential innovations: AI, Computing, Networks, Data & & Security, and Converging Technologies, which we combine to provide sustainability change. Fujitsu Limited (TSE:6702) reported combined profits of 3.5 trillion yen (US$ 23 billion) for the ended March 31, 2026 and stays the leading digital services business in Japan by market share. Discover more: https://global.fujitsu/en-global

About 1Finity

1Finity, a Fujitsu business, is a worldwide service provider of interactions networks for our linked world. We distinctively integrate technological management and know-how in open optical and cordless networking, network automation, and used AI/ML to create, develop, run, and preserve important digital interactions network facilities. Working together carefully with community partners, we provide transformative results for provider and network operators, and allow them to lower TCO, enhance network efficiency, and boost energy effectiveness. To learn more, see www.1Finity.com

Press Contacts