In the age of digital transformation, cloud services are helping global organizations modernize their applications, migrate their workload to the cloud and innovate with data, analytics and artificial intelligence (AI). Consequently, data center development has become vital for processing high-quality data to enhance operational efficiencies.
However, in this data-center-operations growth story, a stark reality prevails. According to a new International Data Corporation (IDC) report, AI data center energy consumption is forecast to grow at a CAGR of 44.7%, reaching 146.2 Terawatt hours (TWh) by 2027, with AI workloads consuming a growing portion of total data center electricity usage. The IDC expects global data center electricity consumption to more than double between 2023 and 2028, demonstrating a five-year CAGR of 19.5% and reaching 857 Terawatt hours (TWh) in 2028. Moreover, climate conscious investors and regulators are keen to ensure that such spikes in power do not trigger a huge rise in greenhouse gas (GHG) emissions.
Turning to Liquid Cooling
There is an urgent need to find solutions that can increase data center efficiency in both a sustainable and cost-effective way. Given that data centers operate 24/7, consuming vast amounts of electricity to power the servers and process data, a lot of heat is generated. If this heat is not removed, the electrical components can overheat and fail.
This evolving trend poses newer challenges for the telecommunications industry as fragmented construction of intelligent computing services can drive up costs. For instance, a single 8-card GPU server consumes nearly 10 kW, resulting in extremely low rack space utilization, making air cooling increasingly unfeasible. To combat these effects, experts suggest introducing technological solutions such as improving chip efficiency and introducing liquid cooling solutions, as well as rethinking data center design to increase data center efficiency.
Liquid cooling has become the necessary (and preferred) choice for managing data center temperatures. Building large-scale, low-cost, liquid-cooled intelligent computing centers can optimize computing power and ensure energy-efficient computing networks, significantly reducing costs and improving efficiency.
xFusion's Solution for Telecommunications
xFusion’s flagship product, the FusionPoD for AI—a rack-scaled, liquid-cooling server—has become the ideal choice for experts and practitioners in the telecommunications industry engaged in AI research and application development. China Telecom's Hangzhou Intelligent Computing Center is a prime example of FusionPoD's successful deployment.
xFusion provides an end-to-end complete liquid cooling solution for data centers, featuring a comprehensive secondary side integration delivery scheme. The FusionPoD for AI server can be used for large-scale deployment, significantly enhancing delivery efficiency.
Leveraging a native liquid cooling design, it incorporates measures such as splash protection, liquid leakage diversion, and shut-off to ensure the reliable operation of high-value equipment, adding an extra layer of protection. This further helps clients save electricity, achieving exceptional energy efficiency and ultra-low PUE.
Through collaborative innovation, the FusionPoD has been thoroughly validated and operates stably in intelligent computing centers.
International Recognitions
In addition, FusionPoD for AI server has received international acclaim for its innovative features, including prestigious certifications and awards. Critically, its liquid cooling system has earned TÜV Rheinland certification for its leak-proof design, addressing a key concern in liquid cooling technology.
Furthermore, FusionPoD boasts the TÜV SÜD Certification for its exceptional power usage effectiveness (pPUE) of 1.06 – the lowest in its class – and its intelligent data center (DC) server capabilities.
Its industry-leading design and efficiency were further recognized with the 2023 Interop Tokyo Best of Show Award Grand Prize and the IDA Design Award Gold Winner 2023. The FusionPoD servers have been deployed in over 70,000 nodes globally.
Future Innovations
On September 24, Intel unveiled its groundbreaking Intel Xeon 6 processors with Performance-cores (P-cores), revolutionizing the high-performance computing (HPC) landscape. As a key strategic partner, xFusion announced its full support of these new processors, which form part of its FusionServer lineup.
This collaboration between xFusion and Intel has set the stage for advanced HPC and energy-efficient solutions. Powered by the Intel Xeon 6 processors with P-cores, the next-generation FusionServer offerings will introduce transformative upgrades across the following niches: AI, HPC, liquid cooling, and general-purpose computing products.
The FusionPoD servers are known its efficiency during peak performance, exceptional energy efficiency and ease of maintenance ideal for data-center performance and the development of green data centers.
In Conclusion
The growing global digital economy is expected to drive a substantial increase in demand for data centers, making them a key area for investment and expansion. However, to ensure cost-effective and sustainable growth, there will be a strong focus on energy-efficient solutions. Innovations like xFusion’s FusionPoD for AI servers will play a crucial role by enabling the integration of high-performance computing power with energy-efficient networks, helping to reduce both operational costs and environmental impact.
FusionPoD for AI: https://www.xfusion.com/en/product/rack-scale-servers/fusionserver-fusionpod
xFusion LinkedIn: https://www.linkedin.com/company/xfusion-global/
xFusion Case Center: https://www.xfusion.com/static/en/cases
