{"id":520209,"date":"2026-06-05T14:32:13","date_gmt":"2026-06-05T14:32:13","guid":{"rendered":"https:\/\/www.europesays.com\/ie\/520209\/"},"modified":"2026-06-05T14:32:13","modified_gmt":"2026-06-05T14:32:13","slug":"huawei-chips-refine-deepseek-model-in-major-leap-for-chinas-ai-self-reliance","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ie\/520209\/","title":{"rendered":"Huawei chips refine DeepSeek model in major leap for China\u2019s AI self-reliance"},"content":{"rendered":"

A research team that includes Huawei Technologies says it has successfully used the firm\u2019s Ascend 910C chips<\/a> to complete post-training for the DeepSeek-V4-Pro model, marking a major step forward as China\u2019s semiconductor industry tries to leap from supporting basic AI inference to more complex model training amid tightening US sanctions.<\/p>\n

While Chinese chipmakers have found success in supporting AI inference \u2013 the relatively simple process of running an already-finished model to answer user prompts \u2013 they have struggled with training, the far more complex process of building or refining a model\u2019s brain.<\/p>\n

If initial \u201cpre-training\u201d teaches a model how to speak by absorbing massive amounts of data, post-training teaches it how to work by following human instructions, safety rules and specific tasks.<\/p>\n

\"AA Huawei Ascend 910 processor is displayed during PT Expo China in 2023. Photo: Shutterstock Images<\/p>\n

To achieve this, the researchers ran DeepSeek\u2019s largest model to date \u2013 boasting 1.6 trillion parameters \u2013 on a computing cluster powered by at least 1,000 Huawei chips, according to a social media post from the Shenzhen government on Friday.<\/p>\n

The team successfully conducted \u201cfull-parameter\u201d post-training, meaning the model\u2019s entire architecture was updated and refined without cutting corners, the post said.<\/p>\n

Previously, domestic computing power was primarily used for inference, \u201cmuch like building a one-way road for the model: input a question, output an answer\u201d, the post explained. The project, however, allowed a model to self-reflect and adjust.<\/p>\n

This added \u201ccomplex flyovers and loops to that one-way road, instantly multiplying the computational and communication demands by several times\u201d, it added.<\/p>\n

The exploration \u2013 jointly conducted by Huawei, the Shenzhen Loop Area Institute, the Shenzhen campus of Harbin Institute of Technology and Shenzhen Research Institute of Big Data \u2013 \u201cwill help enhance the self-reliance of China\u2019s AI industry chain\u201d, the post said.<\/p>\n","protected":false},"excerpt":{"rendered":"A research team that includes Huawei Technologies says it has successfully used the firm\u2019s Ascend 910C chips to…\n","protected":false},"author":2,"featured_media":520210,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[261],"tags":[291,289,290,7102,224788,8535,2078,8736,224787,18,5568,20102,5218,19,17,8206,14338,292,224789,224790,14664,82,2480],"class_list":["post-520209","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-ascend","tag-ascend-910c","tag-chip","tag-computing-power","tag-deepseek","tag-deepseek-v4","tag-eire","tag-generative-ai","tag-gpu","tag-huawei","tag-ie","tag-ireland","tag-large-language-model","tag-llm","tag-nvidia","tag-post-training","tag-pre-training","tag-semiconductor","tag-technology","tag-training"],"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@ie\/116697992807288297","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts\/520209","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/comments?post=520209"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts\/520209\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/media\/520210"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/media?parent=520209"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/categories?post=520209"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/tags?post=520209"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}