--- title: "Milestone moment for domestic AI computing power: Sugon deploys 3 sets of scaleX WanKa superclusters, with the largest domestic AI computing power pool now in operation" type: "News" locale: "en" url: "https://longbridge.com/en/news/275097868.md" description: "On February 5th, Sugon launched 3 sets of scaleX WanKa superclusters in Zhengzhou, becoming the first and largest domestic AI computing power pool in operation nationwide, marking the entry of domestic WanKa clusters into the stage of large-scale deployment. This cluster enhances computing power density and performance through the innovative scaleX640 super node, supporting trillion-parameter model training and high-throughput inference, aiming to lower the R&D threshold for AI clusters" datetime: "2026-02-06T08:43:59.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/275097868.md) - [en](https://longbridge.com/en/news/275097868.md) - [zh-HK](https://longbridge.com/zh-HK/news/275097868.md) --- # Milestone moment for domestic AI computing power: Sugon deploys 3 sets of scaleX WanKa superclusters, with the largest domestic AI computing power pool now in operation On February 5th, the national supercomputing internet core node was launched in Zhengzhou for trial operation, deploying three sets of Sugon scaleX 30,000-card superclusters, becoming the first in the country to achieve a deployment of 30,000 cards and the largest domestic AI computing power pool in actual operation. Currently, large artificial intelligence models are advancing towards trillion-level parameters, multi-modal, and world models. On December 18, 2025, the scaleX 30,000-card supercluster made its first appearance at the HAIC conference, and less than two months later, it was implemented with an AI computing power scale exceeding 30,000 cards, marking the official entry of domestic 30,000-card clusters into the stage of large-scale deployment and practical application. Public information shows that with the world's first single cabinet-level 640-card super node scaleX640, the scaleX 30,000-card supercluster has increased the computing power density of a single cabinet by 20 times, achieving a performance improvement of 30% to 40% for MoE trillion-parameter large model training and inference scenarios compared to traditional solutions. In addition, the scaleX640 super node has passed a 30-day + long-term reliability test, ensuring the expansion deployment of ultra-large-scale clusters at the 100,000-card level. The scaleX640 was officially released on November 6, 2025, driven by a "open architecture + system innovation" dual approach. Its "one-to-two" high-density architecture design not only realizes ultra-high-speed bus interconnection of 640 cards in a single cabinet, constructing a large-scale, high-bandwidth, low-latency super node communication domain, but also allows two scaleX640 super nodes to form a thousand-card-level computing unit. At the same time, in the context of adjustments in the global information technology industry ecosystem, the scaleX640 super node adopts an "AI computing open architecture." This architecture was jointly released by Sugon and more than 20 industry chain enterprises in September 2025, opening up multiple key technological capabilities aimed at lowering the R&D threshold for AI clusters and avoiding redundant investments. It is reported that the scaleX640 super node supports multi-brand accelerator cards at the hardware level and is compatible with mainstream computing ecosystems at the software level, supporting MoE trillion-parameter large model training, high-throughput inference, and scientific intelligence (AI4S) and other cutting-edge scenarios. According to Sugon's official WeChat account, this computing power pool comprehensively covers large-scale AI computing scenarios such as trillion-parameter model training, high-throughput inference, and AI for Science in practical applications: for ultra-large-scale model training, it supports the whole machine training and fault tolerance recovery of trillion-parameter models; for high-throughput inference scenarios, it has served the core intelligent business of several leading internet users and continuously improved inference efficiency through joint deep optimization; in the field of AI for Science, it supports a domestic material research large model to top international authoritative rankings, helping top domestic research teams increase protein research efficiency by 3-6 orders of magnitude; at the same time, it is paired with the OneScience scientific large model one-stop development platform, significantly lowering the innovation threshold for interdisciplinary research Li Bin, Senior Vice President of Sugon, stated that the wave of intelligence is reshaping the world with unprecedented force. Sugon will take the core node launch as a starting point to continue deepening technology research and development and application practices, promoting domestic intelligent computing power to serve various fields of economic and social development more efficiently, stably, and equitably. According to previous reports by this newspaper, Li Jun, the acting chairman of Sugon, stated at the "Artificial Intelligence +" and Digital Economy event organized by the Zhengzhou Municipal Government and co-hosted by Shanghai Securities Journal, Zhengzhou Municipal Bureau of Commerce, and Zhengzhou Office in Shanghai on January 18, that Sugon will next establish an Advanced Computing Technology Research Institute in Zhengzhou, planning to build a 160,000 square meter R&D space, with an initial plan to gather a thousand researchers focusing on R&D innovation in AI and high-performance computing applications. Risk Warning and Disclaimer The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial conditions, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment based on this is at one's own risk ### Related Stocks - [603019.CN](https://longbridge.com/en/quote/603019.CN.md) ## Related News & Research - [Fireworks AI Eyes $15 Billion Valuation In New Funding Talks](https://longbridge.com/en/news/287806419.md) - [CertiK Launches AI Skill Scanner, An Antivirus Software for the AI Age](https://longbridge.com/en/news/287832378.md) - [AI Hyperscalers Are Taking on Debt, But the Broader Market Looks More Leveraged](https://longbridge.com/en/news/288150358.md) - [Billionaire investor Dan Loeb dismisses AI bubble talk: 'We're barely scratching the surface'](https://longbridge.com/en/news/288060733.md) - [Redwood AI Announces Letter of Intent for Potential Acquisition of Quantum.IQ | RDWCF Stock News](https://longbridge.com/en/news/287977147.md)