{"id":41096,"date":"2025-08-10T12:55:20","date_gmt":"2025-08-10T04:55:20","guid":{"rendered":"https:\/\/www.1ai.net\/?p=41096"},"modified":"2025-08-10T12:55:20","modified_gmt":"2025-08-10T04:55:20","slug":"%e6%b6%88%e6%81%af%e7%a7%b0%e5%8d%8e%e4%b8%ba%e5%8d%b3%e5%b0%86%e5%8f%91%e5%b8%83-ai-%e6%8e%a8%e7%90%86%e9%a2%86%e5%9f%9f%e7%aa%81%e7%a0%b4%e6%80%a7%e6%88%90%e6%9e%9c%ef%bc%9a%e9%99%8d%e4%bd%8e","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/41096.html","title":{"rendered":"Huawei to release breakthrough in AI inference soon: reduces dependence on HBM, boosts performance of large domestic models, sources say"},"content":{"rendered":"<p>Aug. 10, 2011 - According to the Daily Kotaku<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%8d%8e%e4%b8%ba\" title=\"_Other Organiser\" target=\"_blank\" >Huawei<\/a>will be held on August 12 at 2025 Financial <a href=\"https:\/\/www.1ai.net\/en\/tag\/ai%e6%8e%a8%e7%90%86\" title=\"[SEE ARTICLES WITH [AI REASONING] LABELS]\" target=\"_blank\" >AI reasoning<\/a>At the Application Landing and Development Forum, the breakthrough technical achievements in the field of AI reasoning were released. It was revealed that this achievement may be able to reduce China's AI reasoning on the <a href=\"https:\/\/www.1ai.net\/en\/tag\/hbm\" title=\"_OTHER ORGANISER\" target=\"_blank\" >HBM<\/a>(high-bandwidth memory) technology dependence, upgrading domestic <a href=\"https:\/\/www.1ai.net\/en\/tag\/ai%e5%a4%a7%e6%a8%a1%e5%9e%8b\" title=\"[View articles tagged with [AI Big Model]]\" target=\"_blank\" >AI Big Model<\/a>inference performance, a key part of improving China's AI inference ecosystem.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-41097\" title=\"e13562d5j00t0rhmn0148d000v900nfp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2025\/08\/e13562d5j00t0rhmn0148d000v900nfp.jpg\" alt=\"e13562d5j00t0rhmn0148d000v900nfp\" width=\"1125\" height=\"843\" \/><\/p>\n<p>1AI notes that Huawei's technological breakthroughs in the field of AI inference already have precedents.2025 In March, Peking University and Huawei released the DeepSeek full-stack open-source inference solution, which is based on Peking University's self-researched SCOW computational platform system and the Hesse scheduling system, and integrates community open-source components, such as DeepSeek, openEuler, MindSpore, and vLLM\/RAY, to realize efficient inference on Huawei's Rise. The solution is based on NU's own SCOW computing platform and Hesse scheduling system, and integrates DeepSeek, openEuler, MindSpore, and vLLM\/RAY.<\/p>\n<p>In terms of performance, Huawei Rise has realized a number of breakthroughs. For example, when CloudMatrix 384 supernodes were deployed with DeepSeek V3\/R1, the single-card Decode throughput exceeded 1920 Tokens\/s under the 50ms latency constraint, and the single-card throughput of Atlas 800I A2 inference server reached 808 Tokens\/s under the 100ms latency constraint.<\/p>\n<p>The cooperation between KU Xunfei and Huawei has also achieved remarkable results, with both parties taking the lead in realizing large-scale cross-node expert parallel cluster reasoning for MoE models on domestic arithmetic power, which improves reasoning throughput by 3.2 times and reduces end-to-end latency by 50%.<\/p>","protected":false},"excerpt":{"rendered":"<p>On August 10, Huawei will release a breakthrough technical achievement in the field of AI reasoning on August 12 at the 2025 Financial AI Reasoning Application Landing and Development Forum, according to Sci-Tech Board Daily. It was revealed that this achievement may be able to reduce China's AI inference dependence on HBM (high-bandwidth memory) technology, improve the performance of domestic AI large model inference, and improve a key part of China's AI inference ecosystem. 1AI notes that Huawei's technological breakthroughs in the field of AI inference have precedents.2025 In March, Peking University and Huawei released the DeepSeek full-stack open-source inference program, which is based on Peking University's self-developed SCOW computational platform system and the Hesse scheduling system, integrating the DeepSeek<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[433,7405,7404,1117],"collection":[],"class_list":{"0":"post-41096","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"hentry","6":"category-news","7":"tag-ai","9":"tag-hbm","10":"tag-1117"},"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/41096","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=41096"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/41096\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=41096"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=41096"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=41096"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=41096"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}