Iceotope launches KUL BOX to boost AI edge computing in factories
Iceotope has launched KUL BOX, a rackmount system designed for artificial intelligence and high-performance edge computing. The KUL BOX aims to address the demands of AI workloads in locations not traditionally equipped for datacentre infrastructure, such as factories, hospitals, and telecom facilities.
Edge computing shift
As AI adoption accelerates, more organisations require intensive compute power closer to the origin of data. This shift presents unique challenges, with many environments lacking the specialised cooling and noise control of standard datacentres. The KUL BOX is intended to bypass these constraints by offering localised compute capabilities without the need for significant infrastructure upgrades.
The system is delivered as a compact rack, housing up to 24 NVIDIA H200 NVL GPUs distributed across six KUL AI liquid-cooled chassis. Each unit also integrates networking, 6TB of memory, and 115TB of distributed storage. The entire solution can be operated using standard power outlets, with no requirement for external water supplies for cooling, broadening suitability across different sites.
Thermal management
Unlike conventional air-cooled systems, the KUL BOX leverages liquid cooling to manage GPU and CPU heat output. This approach is designed to reduce operational noise to below 40 decibels and bring significant gains in energy efficiency. According to Iceotope, the technology can offer cooling cost reductions of up to 83% alongside water usage savings of 96% when compared to traditional air-cooled servers.
"More and more compute horsepower is needed in unusual places close to the action," said Neil Edmunds, VP of Product, Iceotope. "With KUL BOX, we've packaged an entire AI and inferencing solution, removing the traditional complexity and maintenance burden of a custom back end."
The modular design is intended for scalability, allowing racks and cooling infrastructure to be expanded as AI workloads grow. The absence of fans and the reliance on sealed liquid cooling are intended to reduce on-site servicing and operational complexity. KUL BOX aims to simplify deployment for operations teams by providing an integrated solution rather than requiring custom installations.
Deployment partnership
Iceotope is working with UNICOM Engineering, a subsidiary of UNICOM Global, to offer a fully integrated rack assembly. UNICOM Engineering delivers all server, networking, and cooling hardware pre-installed and tested before delivery. This package includes all fluids, pipework, and ancillary equipment required, and is supported by end-to-end services including logistics, on-premise installation, training, and support.
Customers receive a three-year parts and service plan as well as warranty coverage. A single point of contact is provided throughout the product's lifecycle for support and maintenance. The solution is currently being adopted in retail, healthcare, telecoms, and manufacturing settings, where it supports AI-driven applications, secure diagnostics, and data processing tasks at the edge.
Market trends
Industry forecasts predict the global edge AI market will reach USD $66.47 billion by 2030, up from an estimated USD $20.78 billion in 2024. Compound annual growth rates of over 21% are predicted during this period, signalling an increasing demand for high-performance, on-site compute infrastructure. Products such as KUL BOX are targeting this growth by providing capabilities tailored to environments which may lack conventional datacentre resources.
The system aims to offer a balance of performance, energy efficiency, noise management, and operational simplicity suitable for deployment in a wide range of non-traditional computing settings.
