SoftBank Corp., ZutaCore and Foxconn Collaborate on Design and Development of Rack-integrated Solution Optimized for ZutaCore's 2-phase DLC Technology Used in AI Servers that Utilize NVIDIA Accelerated Computing
Achieved pPUE1.03 (actual measured value) as cooling efficiency per rack through operation verification and performance evaluation at SoftBank's data center for practical use.
SAN JOSE, Calif., Feb. 28, 2025 /PRNewswire/ -- SoftBank Corp. ("SoftBank"), ZutaCore® and Hon Hai Technology Group ("Foxconn®") today announced they implemented ZutaCore's two-phase DLC (Direct Liquid Cooling) technology*1 in an AI server using NVIDIA accelerated computing, making it the world's first implementation*2 of ZutaCore's two-phase DLC (Direct Liquid Cooling) technology*1 using NVIDIA H200 GPUs. In addition, SoftBank designed and developed a rack-integrated solution that integrates each component of the server, including cooling equipment with two-phase DLC technology, on a rack scale, and conducted an operational demonstration and performance evaluation at its data center in February 2025. The demonstration results indicated the solution passed NVIDIA's temperature test (NVQual), thereby confirming the compatibility, stability and reliability of this rack-integrated solution. The solution also achieved pPUE (partial Power Usage Effectiveness) of 1.03 (actual measured value) per rack for cooling efficiency*3.
With the spread and increased adoption of AI, demand for AI servers and other computing resources is expected to expand significantly, further increasing power consumption at data centers. At the same time, reducing power consumption from the perspective of reducing carbon dioxide (CO2) emissions has become a pressing issue worldwide, requiring data centers to become more efficient, consume less power, and introduce innovative heat removal solutions. Since May 2024, SoftBank has been collaborating with ZutaCore, a global leader in the development and business deployment of two-phase DLC technology, to develop solutions optimized for low power consumption at AI data centers.
SoftBank, ZutaCore and Foxconn will continue their efforts to achieve higher efficiency and lower power consumption in AI data centers using NVIDIA accelerated computing, and collaborate on the development of rack-integrated solutions optimized with two-phase DLC technology for practical use and deployment in the global market.
*1 A technology that cools the cold plate on the semiconductor chip (processor) inside the server by circulating an insulating refrigerant without water in a two-phase system (liquid and gas).
*2 As of January 2025, according to ZutaCore research.
*3 According to ZutaCore research. PUE is a measure of cooling efficiency in data centers. pPUE is a measure of the efficiency of a particular area or facility, such as a server room or module. The closer the number is to 1.0, the more energy efficient it is.
Features of rack-integrated solution optimized with two-phase DLC technology
1. Plug & Play solution
By integrating server components such as cooling equipment equipped with 2-phase DLC technology, which conventionally had to be optimized individually, on a rack scale, it is possible to provide a "Plug & Play"*4 solution that can be used immediately by simply connecting the equipment. This makes it easy and simple to install in AI data centers.
2. Innovative 2-phase DLC cooling system
ZutaCore's two-phase DLC technology, which cools the chipsets (CPUs, GPUs, memory, etc.) that are the heat sources in the server by bringing them into contact with a cold plate, circulates a special insulating heat transfer (dielectric) fluid with a low boiling point in the cold plate and uses heat absorption during the phase change from liquid to gas caused by the heat source. Efficient cooling is achieved through repeated evaporation and condensation. In addition, the water temperature can be maintained at a higher temperature than in conventional water-cooled facilities, which improves the heat transport capacity of the water used for cooling and reduces the flow rate of water to achieve the required cooling effect. This makes it possible to reduce the pressure of the pump that delivers the water for cooling, contributing to lower power consumption. Furthermore, the use of insulating heat transfer fluid prevents serious damage to the server in the event of a spill of special fluid.
3. AI server with guaranteed compatibility, stability and reliability
Foxconn developed an AI server based on NVIDIA H200 GPUs, optimized with ZutaCore's two-phase DLC technology. This AI server was able to exceed target performance figures in NVIDIA's certification test (NVQual), which determines the compatibility, stability, and reliability of server systems.
4. Versatile and scalable rack for high-density placement of AI servers
SoftBank developed a rack designed to maximize cooling efficiency and improve operational efficiency with its two-phase DLC technology. In addition to being designed to be versatile and scalable for high-density placement of the latest AI servers, the rack has airflow for improved cooling efficiency, and power supply. In addition, power supply and wiring are concentrated in the rear of the rack for operational safety and robot-friendly operation. The rack is also compatible with 21-inch servers and conventional 19-inch servers compliant with the ORV3 standard, which is a specification set by the Open Compute Project (OCP), a non-profit organization that promotes open-source hardware specifications and designs. The specifications are designed to be deployed in accordance with global standards.
Comments from each company
Hironobu Tamba, Vice President and Head of the Data Platform Strategy Division at SoftBank Corp. said,
"As we work to develop one of the largest AI infrastructures in Japan and homegrown largescale language models (LLMs), improving the energy efficiency of data centers is essential to accelerate AI development and realize a sustainable society. The operational demonstration of this rack-integrated solution with optimized two-phase DLC technology has confirmed stable operation and improved energy efficiency of high-density GPU servers with liquid cooling. In the future, we will work on the practical application of this solution for use in AI data centers and AI factories to build a sustainable AI infrastructure and maximize computational efficiency."
Erez Freibach, Co-founder and CEO at ZutaCore said,
"ZutaCore is honored to partner with global industry leaders SoftBank and Foxconn to introduce the world's first 2-phase DLC rack integration solution using NVIDIA accelerated computing for AI servers. This groundbreaking collaboration highlights the transformative potential of our waterless DLC cooling technology to deliver unmatched efficiency while addressing the growing need for sustainable energy solutions. Together, we are not just meeting today's challenges but setting a new standard for energy-efficient innovation in next-generation AI data centers and factories, driving progress while safeguarding the planet."
Foxconn indicated,
"Foxconn, as the world's largest AI server supplier, is delighted to join this collaboration and share our extensive in AI server development. We look forward to driving greater value for the global AI industry together."
Future Developments
SoftBank, ZutaCore, and Foxconn will strengthen cooperation with AI server and rack design companies to improve the efficiency and low power consumption of AI data centers using NVIDIA accelerated computing, with a view to the commercialization of this solution and its deployment in the global market.
*4 This system that automatically recognizes peripherals and makes the necessary settings when peripherals are connected to the main unit of a device.
・SoftBank, the SoftBank name and logo are registered trademarks or trademarks of SoftBank Group Corp. in Japan and other countries.
・Other company, product and service names in this press release are registered trademarks or trademarks of the respective companies.
Share this article