Arrcus, a provider of distributed networking infrastructure, has announced a collaboration with Fujitsu and 1Finity to develop a secure, energy-efficient architecture aimed at supporting the next phase of artificial intelligence deployments. The initiative centers on FUJITSU-MONAKA, Fujitsu’s next-generation Arm-based CPU, and is designed to address infrastructure requirements emerging from the growth of physical, agentic, and training AI workloads.
As AI inference increasingly shifts from centralized cloud data centers to edge environments, enterprises and service providers are facing new challenges related to latency, power consumption, data sovereignty, and security. The companies said their joint effort will focus on building a Secure Sovereign AI Infrastructure capable of supporting highly distributed AI workloads, from real-time edge inference to scalable AI training, while meeting performance and governance requirements.
The collaboration integrates FUJITSU-MONAKA-powered compute platforms with Arrcus’ ArcOS®, a disaggregated, software-defined network operating system, and 1Finity’s high-speed optical interconnect technology. Together, the components are intended to enable intelligent traffic orchestration, secure connectivity, and automated operations across distributed AI environments. The companies said the architecture is designed to help service providers transition from traditional connectivity roles to programmable AI infrastructure platforms capable of delivering next-generation AI services.
Shekar Ayyar, Chairman and CEO of Arrcus, said the shift of AI inference closer to where data is generated — including factories, hospitals, warehouses, and regional networks — is redefining infrastructure needs. He noted that combining distributed networking software with the FUJITSU-MONAKA platform aims to create infrastructure that can scale from edge to core while optimizing power efficiency and operational flexibility.
FUJITSU-MONAKA is engineered to deliver high-performance AI inference and data processing with improved power efficiency and built-in confidential computing capabilities. These features are intended to provide hardware-rooted security and support data sovereignty requirements. When integrated with Arrcus’ network operating system, the solution enables dynamic traffic steering, workload-aware routing, and end-to-end orchestration across distributed environments, functions considered essential for real-time and physical AI applications.
Toshio Yoshida, Vice President in charge of processor development at Fujitsu’s Advanced Technology Development Unit, said the collaboration is focused on delivering infrastructure that is powerful, energy-efficient, and secure by design.
The joint architecture is expected to support use cases including smart factories, robotics, logistics automation, healthcare systems, enterprise AI workloads requiring localized data governance, and service provider-delivered distributed AI hosting platforms.
