A high-tech enterprise specializing in the research and development of optical fiber communications and manufacturing.
Language
A high-tech enterprise specializing in the research and development of optical fiber communications and manufacturing.
CloudHQ's Mexico campus plan includes six facilities with a total computing capacity of up to 900 MW, designed to meet the demanding needs of future AI services. Such projects place extreme demands on data center infrastructure, particularly in terms of data transmission bandwidth, connection density, and energy efficiency. Kexint recognizes that the efficient operation of AI computing clusters depends on the underlying infrastructure's ability to achieve breakthroughs in high bandwidth, high density, and high speed. To address the core needs of these hyperscale data centers, Kexint offers comprehensive solutions: High-density connectivity and space optimization: For the massive connectivity requirements between servers and switches within the campus, Kexint's MPO/MTP high-density fiber optic patch cords and high-core-count fiber optic cables (288 cores or more) significantly reduce cabling space and simplify network architecture.
Kexint's standard fiber optic cable series: 24-core, 48-core, 96-core, 144-core, and 288-core, suitable for small to medium-scale network interconnects, balancing cost and practicality;
Kexint's high-density packaging series: MPO/MTP fiber optic patch cords, with multi-core integration design (commonly 12-core or 24-core), significantly reduce cabling space and improve connection efficiency, especially suitable for high-speed interconnects between switches and servers in AI clusters, reducing link latency and maintenance complexity.
Its 1U/144-core and 4U/576-core high-density rack solutions perfectly address the sustainability and space efficiency requirements of the CloudHQ project, helping customers overcome "space bottlenecks" and reduce energy consumption and footprint.
1U rack: Supports 144-core fiber optic connections, suitable for edge nodes or space-constrained areas, achieving high-density termination within a 1U (1.75 inch) height;
2U rack: Supports 288-core fiber optic connections, balancing space and capacity, suitable for aggregation layer deployment in small to medium-sized AI computing clusters;
4U rack: Supports 576-core fiber optic connections, designed for the core data center of large-scale AI data centers, meeting the ultra-high-density connection needs of tens of thousands of GPU clusters, reducing the number of racks and lowering data center energy consumption and footprint.
High-Speed Data Transmission Assurance: AI training and inference generate terabyte-scale data streams. Kexint's 400G/800G optical modules have become the mainstream choice for core interconnects in such data centers, while the forward-looking 1.6T optical module provides ample bandwidth headroom for future computing power upgrades, ensuring seamless data transmission and directly supporting the training efficiency of large-scale AI models with hundreds of billions of parameters.
100G Optical Modules: Suitable for basic interconnect scenarios in small and medium-sized AI data centers, such as connecting servers to access layer switches;
400G Optical Modules: The current mainstream speed for core interconnects in AI data centers, widely used for high-speed interconnects between aggregation and core layer switches, and for AI server clusters;
800G/1.6T Optical Modules: Forward-looking products for future ultra-large-scale AI clusters (such as intelligent computing centers), supporting terabyte-scale data transmission and providing "uninterrupted" high-speed support for training large-scale AI models with hundreds of billions of parameters and real-time processing of massive data. Our Core Advantages: Empowering global leading AI infrastructure projects with one-stop customized solutions.
Our Core Advantage: Empowering Global AI Infrastructure Flagship Projects with One-Stop Customized Solutions
Kexint's value lies not only in its individual products, but also in the synergy and forward-thinking nature of its overall solutions. Its products offer complete end-to-end compatibility, effectively mitigating deployment risks associated with using multiple brands, and enhancing overall system stability. Furthermore, Kexint possesses strong customization capabilities, providing comprehensive services from planning to operation and maintenance for projects of various scales, from edge nodes to super-large AI data centers—perfectly aligning with the needs of large-scale projects requiring highly customized designs, such as CloudHQ's project.
Kexint stated, "CloudHQ's investment in Mexico is a strong signal of the global development of AI infrastructure. We are committed to building a robust computing and data transmission foundation for the global AI industry through 'higher density, faster speed, and smarter' AI data center solutions. The successful implementation of such flagship projects will once again demonstrate Kexint's core value: leveraging cutting-edge technology to enable customers' long-term evolution."
