Japan’s next major AI infrastructure push is starting with a $15 billion data center and could eventually grow into a much larger national buildout.
Dell Technologies, Japanese power producer JERA, and UK-based AI infrastructure developer RHAELM have signed a memorandum of understanding to develop a standardized model for AI data centers across Japan. The first project, planned for JERA’s Chiba Thermal Power Station, is expected to cost more than $15 billion and support up to 400 megawatts of capacity.
The Financial Times reports that the broader multi-site effort could involve as much as $140 billion in investment. JERA has confirmed plans to pursue multi-gigawatt AI infrastructure across Japan in the 2030s, but its official announcement does not attach a $140 billion figure to that expansion.
For channel partners, the significance extends beyond one facility. The companies want to turn the Chiba design into a repeatable infrastructure model, potentially creating demand for the technology and services required to deploy and operate additional AI campuses.
Chiba project starts with more than $15B
The first deployment will be built at JERA’s Chiba Thermal Power Station and is expected to require more than $15 billion, or roughly 2.3 trillion yen, across land, power infrastructure, construction, and AI computing equipment.
The project is planned for up to 400MW of capacity, with operations targeted to begin around 2028.
According to JERA, the Chiba development is intended to serve as a model for additional sites. JERA and RHAELM plan to explore similar deployments as they pursue multi-gigawatt AI infrastructure capacity across Japan during the 2030s.
Apollo Global Management also intends to participate as a strategic financing and investment partner for RHAELM on the project.
Power generation is being built into the data-center model
One of the project’s most notable features is its location at an existing JERA power-generation site.
The companies plan to supply electricity behind the meter from JERA’s infrastructure rather than relying solely on a conventional grid connection. JERA says that approach could make computing capacity available sooner than projects that must wait for new grid connections.
That addresses one of the practical constraints facing AI infrastructure development. Building more compute capacity also requires enough electricity to operate it, which can make power availability as important as access to servers and accelerators.
Japan’s government has been trying to coordinate those requirements through its Watt-Bit Collaboration initiative.
The Ministry of Economy, Trade and Industry and Ministry of Internal Affairs and Communications created the program to better align electricity, telecommunications, and data-center development as demand from AI and other digital workloads grows.
The JERA project follows a similar approach by planning the computing environment and its power supply together from the beginning.
Dell AI Factory becomes the compute standard
Dell will provide the computing layer through its Dell AI Factory portfolio.
JERA will supply the site and power-generation infrastructure, while RHAELM will lead project delivery and integrate the energy and computing environments.
The companies say a standardized architecture could reduce the amount of custom engineering required when the model is deployed at future locations.
Dell Infrastructure Solutions Group President Arthur Lewis said the collaboration is intended to create infrastructure capable of supporting AI deployment at national scale.
“Deploying AI infrastructure at national scale requires a foundation built for control, flexibility and growth from day one,” Lewis said in JERA’s announcement.
If the Chiba architecture proves repeatable, JERA and RHAELM intend to apply the model at additional sites rather than designing each deployment from scratch.
Repeatable AI campuses could widen the channel opportunity
For the channel, that emphasis on repeatability may be the most important part of the project.
Large AI campuses need supporting infrastructure and expertise around the core compute environment. That can include networking, storage, cooling, security, integration, and ongoing infrastructure management.
Dell, JERA, and RHAELM have not announced reseller incentives, margins, or a partner program specifically tied to the Chiba project, and there is no guarantee those requirements will be sourced through channel companies.
However, replicating the same architecture across multiple locations could create recurring demand for partners capable of deploying, integrating, securing, and maintaining those environments.
Standardization could be particularly important for integrators. Instead of treating every large AI campus as an entirely bespoke project, partners could potentially develop deployment methods and operational expertise that transfer from one site to the next.
That opportunity depends on whether the Chiba model actually scales.
The confirmed project is a more than $15 billion, 400MW AI campus. The larger channel opportunity will take shape if Dell, JERA, and RHAELM succeed in turning that first deployment into the multi-site, multi-gigawatt infrastructure model they envision for Japan.
Other news: Samsung is investing $1 billion in Nvidia-backed Helix Digital Infrastructure, adding its data center, energy, cooling, and construction capabilities to the firm’s push to scale AI infrastructure globally.