Converting Used Ships into Data Centers Mitsui O.S.K. Lines and Hitachi
Mitsui O.S.K. Lines (MOL), Hitachi, Ltd., and Hitachi Systems have agreed to develop, operate, and commercialize a floating data center (FDC) converted from a used ship. The three companies will focus on Japan, where the Hitachi Group already has experience operating onshore data centers, and Malaysia and the United States, where they have experience providing onshore data center-related services. They aim to begin operations after 2027 and will conduct demand verification for the FDC, examine basic specifications and operational procedures, and verify its feasibility.
MOL, based on its experience in ship conversion plans, coordination with port authorities, and evaluation of maritime operations including mooring and maintenance, will be responsible for “planning and promoting ship conversion, leading consultations with port authorities, clarifying maritime operational requirements such as mooring and maintenance, and considering funding schemes.”
Hitachi, Ltd. and Hitachi Systems will be responsible for “technical considerations for the design, construction, and operation of data centers, definition of IT infrastructure requirements such as network and security, utilization of on-site expertise, and cooperation in clarifying customer requirements and developing new customers.”
With the rapid spread of generative AI, demand for data centers is expanding. Especially in recent years, there has been a demand for diverse supply models that take into account location, securing water resources for power and cooling, surrounding infrastructure, and disaster risks. FDCs eliminate the need for large-scale land acquisition and acquisition costs in urban areas, and construction periods can be shortened. Data centers consume large amounts of electricity and generate significant heat, requiring cooling systems, but FDCs make it easy to introduce water-cooled cooling systems utilizing seawater and river water.
※Translating Japanese articles into English with AI
