Japanese financiers target AI infrastructure as large data-center plans advance in Akita and Tomakomai
Редакция ЦОД:Главное
Nippon Life Insurance plans to raise infrastructure financing, including lending for AI data-center construction, to JPY2 trillion by fiscal 2035, while Mitsubishi UFJ Bank has decided to invest in an AI-infrastructure fund exceeding JPY4 trillion in size. In parallel, a 500 MW grid-connected data-center concept is advancing in Akita Prefecture and SoftBank is developing a 300 MW grid-connected facility in Tomakomai, Hokkaido.
- Nippon Life Insurance is targeting JPY2 trillion of infrastructure financing, including AI data-center construction lending, by fiscal 2035.
- Mitsubishi UFJ Bank’s decision concerns an AI-infrastructure fund exceeding JPY4 trillion in size; the bank’s own investment amount was not disclosed.
- The Akita proposal is a concept with 500 MW of grid-connected capacity, not confirmed IT load, and an anticipated operating date in the first half of the 2030s.
- SoftBank is developing a Tomakomai data center with 300 MW of grid-connected capacity.
- Power availability, environmental permits and local regulation are becoming material factors in data-center underwriting and site selection.
Japanese financial institutions are increasing their focus on AI-related infrastructure financing as large-scale data-center proposals progress domestically.
Nippon Life Insurance has stated that it aims to bring infrastructure financing, including lending for AI data-center construction, to JPY2 trillion by fiscal 2035. Its existing loan portfolio, concentrated primarily in the United States, stood at about JPY1 trillion at the end of March. Separately, Mitsubishi UFJ Bank has decided to invest in an AI-infrastructure fund with a size exceeding JPY4 trillion. The reported fund size should not be read as the bank’s own committed investment amount.
On the supply side, a data-center concept in Akita Prefecture is being advanced with 500 MW of grid-connected capacity and a stated maximum project cost of JPY2 trillion. The concept is expected to use substantial offshore wind potential, with operations anticipated in the first half of the 2030s. The 500 MW figure refers to receiving or grid-connected electrical capacity, not confirmed IT load, and the project remains a concept rather than an operating facility.
SoftBank is also progressing a large data-center development in Tomakomai, Hokkaido, with 300 MW of grid-connected capacity. As with the Akita proposal, this is a power-receipt capacity metric rather than a stated IT-load figure.
The financing and capacity plans sit against a more restrictive permitting backdrop in parts of the United States. New York State has reportedly suspended new environmental permits for up to one year for large data centers with power consumption of 50 MW or more. Virginia, meanwhile, has tax incentives intended to attract data centers that run through 2050. This uneven regulatory environment could influence where developers pursue capacity and reinforce the importance of evaluating grid, water and environmental conditions before financing is committed.
For lenders and operators, the central issue is execution risk. Delays, redesigns or permitting constraints can affect project returns and the credit exposure of financing institutions, even where demand for AI infrastructure and access to capital remain strong.
The developments underline that AI data-center expansion is increasingly determined by more than tenant demand and capital availability. Grid capacity, environmental permitting, water constraints and local regulation are becoming central underwriting and site-selection factors; these variables could redirect demand between markets and raise capital costs for projects without mature power and environmental plans.
- — / 20 сентября 2026 г. · ЯпонскийПервоисточник →
Материал подготовлен редакцией ЦОД: Главное на основе указанных первоисточников.