【资料图】

BEIJING, August 3 (TiPost)— Head of Huawei’s smartphone business signals an artificial intelligence (AI)-empowered version of operating system HarmonyOS to be come after market share of the tech giant’s mobile just regained top five in home market.

Credit：Visual China

“You can’t imagine what I posted just now to invite everybody to join in this year’s HDC (Huawei Developer Conference). Together was written by the upcoming updated Xiaoyi ! ” said Yu Chengdong, or Richard Yu, CEO of Huawei Consumer Business Group as well as CEO of Huawei Intelligent Automotive Solution Business Group, through his account at Chinese Twitter-like social media Weibo. Xiaoyi, also known as Celia, is Huawei’s voice assistant launched in 2018.

HarmonyOS will soon bring a brand new experience with more intelligence and consideration, Yu added in his Weibo post, attaching to a video about how he made Xiaoyi to accomplish the task. According to the video, Yu first asked Xiaoyi to create an invitation draft for HDC on August 4, calling on visitors to witness latest development of HarmonyOS ecosystem, the intelligent assistant then responded with a paragraph that called for either developers or users to find out latest developments of the Harmony world and unlock endless possibilities, adding a hashtag of HDC.

Yu’s post can be deemed as a spoiler for this week’s HDC, specifically speaking, he suggested HarmonyOS 4, the latest version of HarmoyOS that is set to be release at the HDC, will have a new major feature that powered by large AI models, which enables an automated content generation, Shanghai Securities News commented. The state-run newspaper learned that both Huawei and many other a mobile vendors like vivo and Honor, the brand formerly owned by Huawei, are ramping up AI developments, and introducing large language models (LLMs) into mobiles is one of their key targets.

Yu’s signal came days after IDC, the market intelligence firm on tech industry, unveiled Huawei became the rare spotlight in China’s lackluster smartphone market in the second quarter of the year. IDC estimated 65.7 million smartphones shipped in China in that quarter, a narrower year-over-year (YoY) decline of 2.1% compared with the previous quarter’s decrease of 12.1%. IDC noted smartphone sales dropped more than 5% even in 618, the second largest online shopping festival in China, highlighting the he sluggish consumer demand recovery. It expected the challenging environment will continue in the short term. However, Huawei mobile notched 13.0% of market share, reaching the Top 5 again by having a tie with Xiaomi, which had a share of 13.1% in the second quarter. IDC believed the return of Huawei was mainly supported by a better product launching pace as well as the favorable sales performance of its P60 series and foldable Mate X3 model. Huawei’s share surged 76.5% YoY, becoming one of the only two vendors with a positive YoY growth in the top 5 ranking. Apple had a share of 15.3% just ahead of Xiaomi, compared with a 14.1% share a year earlier.

Huawei smartphone is poised to get more traction from AI tools. The company last month rolled out Pangu Models 3.0, the latest version of its Pangu pre-trained deep learning AI model. Unlike the mainstream large models such as ChatGPT, Pangu Models 3.0 doesn’t know how to compose poems, but it can do concrete works with three purposes for innovation—reshaping industries, honing technologies and sharing success, to continue to build core competitiveness and serve customers, partners in the industry and developers with better services, said Zhang Pingan, executive director of Huawei and CEO of Huawei Cloud.

As an industry-oriented large model, Pangu Models 3.0 is made up of three layers with so-called “5+ N+ X” architecture. The foundation layer, L0, represents the “5”. It provides various skills to meet the needs of different industry scenarios with its five foundational models in natural language processing, computer vision, multimodal learning, prediction and scientific computing. Pangu Models 3.0 is available in serialized foundational models with four different parameters numbers including 10 billion, 38 billion, 71 billion and 100 billion, to satisfy customers’ diversified industry demands in terms of scenario, latency, and response speed.