Submer launches Corenix subsidiary to manufacture prefabricated AI data centre modules
Submer has established Corenix, a new venture specializing in modular data centre infrastructure built for artificial intelligence workloads. The company will design and deliver complete data centre modules around Nvidia reference architectures, integrating IT, power, cooling and networking components before factory testing and shipping to customer sites. Corenix joins two other Submer Group entities—Rubix Data Centres and Radian Arc—to address growing capacity demands from AI facilities worldwide.
Corenix represents Submer's strategic response to accelerating demand for AI infrastructure by specializing in pre-assembled data centre units. Rather than building facilities from scratch on-site, the company manufactures complete integrated modules incorporating computing, electrical systems, thermal management and connectivity within controlled factory environments. This approach allows customers to receive fully tested, ready-to-deploy infrastructure that reduces uncertainty around timelines and performance specifications.
The leadership appointment of Martin Renkis, whose background spans hardware innovation and large-scale infrastructure management at Johnson Controls, signals Submer's commitment to operational reliability and customer accountability. Corenix operates within a broader Submer ecosystem that includes campus development and managed GPU services, positioning the group to serve different segments of the AI infrastructure market simultaneously.
The modularization of data centre infrastructure could accelerate AI deployment globally by reducing construction timelines and technical risks for operators. This may benefit organizations seeking rapid capacity expansion while potentially concentrating manufacturing expertise among specialized providers. The approach could also influence how companies balance capital expenditure through factory-integrated solutions versus traditional construction, though widespread adoption would depend on cost competitiveness and demonstrated reliability across diverse deployment environments.