XinKeLCS Launch of a fully prefabricated liquid-cooled container: delivery of a computing unit in 72 hours
As the demand for AI computing power surges exponentially, the delivery speed of computing infrastructure is becoming the sharpest competitive variable in the intelligent era. The traditional data center construction model, characterized by “long cycles, high energy consumption, and heavy civil engineering,” makes each computing power expansion a long wait. Today, Silicon Labs offers a completely new solution.
From civil engineering to prefabrication: A revolution in computing power delivery methods
The 45-foot fully prefabricated integrated liquid-cooled container data center is a “plug-and-play” computing unit developed by Silicon Labs for high-density AI computing scenarios. It integrates five major systems—power distribution, computing power, liquid cooling, automation, and security—into a standard container space. All modules are prefabricated and debugged in the factory, requiring zero on-site civil construction. Only water, electricity, and network connections are needed, and it can be put into operation within 72 hours, reducing the deployment cycle by more than 90% compared to traditional data centers.
Ultimate integration: 300kW computing power packed into a single box
The core configuration showcases integrated capabilities: the cooling system uses the same air and liquid cooling source, with direct cooling and dual-path cooling working in tandem, and liquid cooling directly connected to the chip cold plate, ensuring stable cooling under all operating conditions and supporting simultaneous operation of NVL72 GB300 and B300. Functions and resources are perfectly integrated within a standard container space—a data center that can truly “pack and go.”
Three core values: rapid delivery, green energy saving, and unmanned operation.
In terms of delivery speed, the production and deployment cycle is shortened by more than 90% within 72 hours, allowing computing power to be available on demand. In terms of energy efficiency, the product is based on a closed-loop liquid cooling system, which uses coolant as a heat transfer medium to absorb the heat generated by the server. The high-temperature coolant after absorbing heat enters the cold station to complete heat exchange and cooling, and then flows back to the computing equipment for circulating heat dissipation. This breaks through the power density limit of traditional air cooling, significantly reduces cooling energy consumption, and achieves low PUE operation. In terms of operation and maintenance, it supports 24/7 cloud remote monitoring, automatic fault alarm and energy consumption optimization, and realizes unattended operation and maintenance.
One box supports five high-density computing scenarios
From edge computing and emergency computing power to AI intelligent computing, government supercomputing, and university research, the 45-foot liquid-cooled container, with its closed-loop liquid cooling heat exchange principle and all-domain interlocking automatic control system, provides a complete solution for high-density computing scenarios, offering efficient heat dissipation, reliable operation, and intensive deployment. It serves not only as the foundation for next-generation computing infrastructure but also as a “computing ark” ready to set sail at any time, enabling computing power to reach every corner where it is needed with lower PUE and faster speed.
Establish yourself with standards, and strive for excellence through intelligent manufacturing.
Starting in 2016 with sales and after-sales service of supercomputing servers, Xinke Intelligent officially established itself in 2019 and devoted itself to the research and development of heat dissipation systems. After three years, it built a fully independent intelligent liquid cooling production line, deeply cultivating cold plate liquid cooling technology. It has successively launched indoor and outdoor full-scenario solutions covering data centers from single units to tens of thousands of units, as well as mobile liquid-cooled data center modules. Its maximum installed capacity has exceeded 2.8 MW, with over 50 product types, and it has participated in the construction of overseas supercomputing sites exceeding 200 MW.
As a key driver of industry development, the company participated as a major drafting unit in compiling China’s first national standard for liquid cooling, “GB/T 48023-2026 Technical Specification for Cold Plate Liquid Cooling Systems for Data Centers.” It also led the participation in the group standard “Energy Efficiency and Low Carbon Evaluation Specification for Liquid Cooling Data Centers,” filling a standard gap in the field of cold plate liquid cooling. Furthermore, it pioneered the release of the industry’s first integrated data center liquid cooling skid-mounted solution, promoting the large-scale standardized application of cold plate liquid cooling.
Relying on its 30,000-square-meter modern production base in Tangshan, with an annual capacity of 500MW and a closed-loop manufacturing capability covering the entire process from material preparation, welding, electrical integration, final assembly, comprehensive testing, and intelligent warehousing, Silicon Labs continuously empowers the green transformation of global data centers with one-stop delivery—making computing power more efficient and energy greener.
See you in Langfang from September 11-13.
From September 11th to 13th, Silicon Labs will showcase its new products at the Beijing-Tianjin-Hebei Big Data Innovation Application Center in Langfang City, Hebei Province (Booth No.: C08). We cordially invite you to visit our booth and witness firsthand the incredible speed of “300kW packed into a single box,” and together embark on a green journey for new computing infrastructure.