en.Wedoany.com Reported - Sugon showcased its full-chain AI infrastructure capabilities at the International Supercomputing Conference ISC High Performance 2026, held in Hamburg, Germany, covering core technologies such as chips, computing, storage, networking, cooling, applications, and services.
On June 23, ISC High Performance 2026 opened in Hamburg, Germany. This conference, gathering major technology companies from Europe, the Americas, and Asia, focused on the evolution of artificial intelligence and high-performance computing infrastructure. At the event, Sugon demonstrated AI hardware capabilities from China to European and global professionals, while providing trend references and engineering implementation pathways for next-generation AI infrastructure construction.
As large model parameters continue to grow, super AI clusters impose higher demands on the overall coordination of computing, storage, and networking. Sugon showcased its systematic "Designed in China" capabilities and engineering-oriented "Made in China" capabilities. Its scaleX AI supercluster adopts a tightly coupled design concept for computing, storage, and networking, improving training and inference efficiency. Products displayed alongside, including AI servers, supernodes, all-flash storage, and native high-speed interconnect networks, reflect its innovation capabilities in individual technologies.
The expansion of AI infrastructure scale brings challenges in power and cooling. Sugon presented green computing solutions covering individual machines, clusters, and computing networks. The scaleX AI supercluster leverages phase-change immersion liquid cooling technology to reduce cooling energy consumption and improve the efficiency of converting electricity into computing power. Additionally, it lowers power costs through renewable energy sources such as wind and hydro power, along with system-level computing-electricity cycling methods. Relying on the supercomputing internet, Sugon interconnects and dispatches computing resources from advantageous regions to alleviate issues such as uneven national computing resource distribution and high power costs.
With the growth of large model parameters and the development of intelligent agents, AI has permeated fields such as basic scientific research, industrial manufacturing, and core business operations. The CPU+GPU heterogeneous architecture is becoming the mainstream computing form, blurring the boundaries between supercomputing and intelligent computing. Future large-scale AI infrastructure will accelerate its transformation toward AI superclusters. Sugon's scaleX series of AI supernodes and supercluster products incorporate full-precision computing capabilities from FP64 high precision to INT8 low precision from the outset of system design, building an ecosystem with high adaptability, compatibility, and openness. This defines a new generation of AI infrastructure oriented toward "super-intelligent convergence" as an application direction. Multiple 10,000-card AI superclusters have been connected to the supercomputing internet operated by Sugon, providing various computing access services such as tokens, APIs, MaaS, and card-hours through industry application software and development tools on the platform.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
