What happened
JERA, Japan’s largest power generation firm, disclosed plans to develop an AI‑focused data center on the site of its Chiba Thermal Power Station. The project, valued at ¥2.3 trillion (about $14.5 billion), will be built with U.S.‑based Dell Technologies and U.K.‑based RHAELM. The center will have a power capacity of roughly 400 MW, making it one of the biggest AI data facilities in the country. Operations are slated to begin in phases around 2028, reaching full capacity in 2029. JERA will supply the power under a long‑term contract of 15‑25 years, while Dell will provide standardized servers and racks, and RHAELM will handle development, construction, operations, and financing. Apollo Global Management will act as a strategic investor for RHAELM. The partners intend to use the Chiba model to expand into overseas markets after establishing integrated standards for energy procurement and equipment sourcing.
On 1 May 2024, JERA announced a ¥2.3 trillion joint venture with Dell Technologies and RHAELM to build an AI data center at its Chiba LNG‑fired thermal power station in Chiba City. The memorandum of understanding was signed the same day, outlining a collaborative model that integrates power supply, equipment procurement, and operational standards.
The facility will be capable of delivering about 400 MW of electricity directly from the plant, eliminating the need for new transmission lines that typically delay data‑center projects for up to a decade. Dell will supply the core IT hardware, while RHAELM, backed by Apollo Global Management, will manage construction, financing, and ongoing operations.
Phase‑1 operations are targeted for 2028, with full capacity expected by 2029. JERA will commit to a 15‑ to 25‑year power supply contract, ensuring long‑term stability for AI workloads. The partners intend to use the Chiba project as a template for future deployments at other JERA plants, potentially aggregating several million kilowatts of AI‑grade power across Japan by the 2030s.
The announcement cites rising demand from AI operators and hyperscalers seeking reliable power in the Tokyo metropolitan area. It also references similar co‑location strategies in the United States, such as Google’s Texas wind‑solar data centers and Meta’s Ohio thermal‑plant integration.
Source details: finance.biggo.com ↗
Why it matters
The announcement tackles a critical bottleneck for AI in Japan: securing reliable, high‑capacity power without lengthy grid‑connection delays. By co‑locating the data center with an existing LNG‑fired plant, JERA sidesteps the typical 5‑10‑year wait for new transmission infrastructure, offering a faster path to scale AI workloads. The 400 MW capacity exceeds other domestic proposals, positioning the facility as a for AI infrastructure and potentially attracting hyperscalers, government agencies, and financial firms that need stable, low‑ compute. The involvement of Dell signals that the hardware stack will be enterprise‑grade, while RHAELM’s focus on “sovereign AI” suggests the center may prioritize domestic data residency and security requirements. The scale of investment also reflects growing confidence in AI‑driven services despite Japan’s broader energy transition challenges, highlighting LNG’s role as a bridge fuel for continuous compute. If successful, the model could be replicated at other JERA plants, creating a network of AI‑ready sites across the country and influencing how power utilities globally consider data‑center integration.
Power availability is the primary constraint on AI data‑center expansion in Japan. Grid congestion has forced many projects into prolonged approval processes, limiting the country’s ability to host large‑scale AI . By leveraging existing generation capacity, JERA’s model offers a faster, more predictable route to scale.
The 400 MW capacity surpasses other domestic proposals, such as the 500 MW Akita project backed by UAE investors, making JERA’s effort a potential new industry standard. This scale can attract major AI service providers who require continuous, high‑throughput , thereby strengthening Japan’s AI ecosystem.
The involvement of Dell signals that the data center will use enterprise‑grade hardware, which may lower entry barriers for Japanese firms seeking to run AI workloads without building their own infrastructure. RHAELM’s focus on “sovereign AI” suggests the center will prioritize data residency and security, aligning with government and financial sector requirements.
If the integrated standards prove successful, they could be exported to overseas markets, influencing global AI‑infrastructure strategies and encouraging other utilities to consider co‑locating data centers with power generation assets.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Which description best fits "narrow AI", the kind of AI in use today?
What to watch next
Key developments to monitor include: (1) the timeline and any regulatory approvals for construction, especially environmental reviews related to LNG use; (2) the identity of early AI customers and whether they secure long‑term power contracts; (3) pricing terms for power and services, which have not been disclosed; (4) progress on the planned overseas expansion and whether the integrated standards are adopted by other operators; and (5) any policy shifts in Japan that could affect the viability of LNG‑based AI data centers versus renewable‑based alternatives.
Regulatory approvals: Environmental impact assessments for the LNG plant’s additional load and any local opposition could affect the construction schedule.
Customer acquisition: Early contracts with AI firms or cloud providers will indicate market appetite and price points, which have not been disclosed.
Pricing and access: The terms of power and pricing remain unknown; transparency will be crucial for broader industry adoption.
International rollout: The partners’ plan to replicate the model abroad will depend on securing similar power‑plant partnerships and navigating foreign regulatory environments.
Energy policy shifts: Japan’s future energy mix, especially moves toward renewable sources, could impact the long‑term viability of LNG‑based AI data centers.