Kunlun(SZ:300418)2024-05-09 15:08Financial Data and Key Metrics Changes - In FY23, the company achieved revenue of 4.92 billion yuan, a year-on-year increase of 3.8% [1] - The overseas business revenue accounted for 86% of total revenue, growing by 8.4% year-on-year [1] - The gross profit margin reached 78.8%, with a net profit attributable to shareholders of 1.26 billion yuan, up 9.15% year-on-year [1] - In Q1 2024, revenue was 1.21 billion yuan, with overseas revenue accounting for 88.4%, reflecting a year-on-year growth of 6.2% [1] - The gross profit margin for Q1 2024 was 78.4% [1] Business Line Data and Key Metrics Changes - R&D expenses in 2023 were 970 million yuan, a 40.2% increase year-on-year, representing a record high of 19.7% of total revenue [1] - The number of R&D personnel increased by 16.1% to 1,508 [1] - Opera's revenue in 2023 was 397 million USD, a 20% increase year-on-year, with adjusted EBITDA of 93.72 million USD, up 38% [1] - In Q1 2024, Opera generated 102 million USD in revenue, a 17% year-on-year increase, with adjusted EBITDA of 24.91 million USD, a 15% increase [1] Market Data and Key Metrics Changes - The global monthly active users of Opera reached 304 million as of Q1 2023 [1] Company Strategy and Development Direction - The company is focused on AGI and AIGC, aiming to build a diverse AI business matrix including AI search, music, video, social, and gaming [3] - The company plans to enhance its AI capabilities through continuous R&D investment and talent acquisition [1][3] - The company aims to become a leading infrastructure and application platform in the AI era, enhancing user efficiency and enabling high-quality content creation [5] Management Comments on Operating Environment and Future Outlook - Management noted that the AI industry is evolving rapidly, with a focus on both general AI and AIGC [5] - The company believes that the AIUGC platform will be a long-term competitive strategy, enhancing user value and generating profits [5] - Management expressed confidence in the company's ability to keep pace with global AI advancements, particularly in the AIGC sector [11] Other Important Information - The company has released multiple iterations of its TianGong model, with the latest version (3.0) being one of the largest and most powerful open-source models globally [2][4] - The company is collaborating with various industry players, including Alibaba Cloud and Huawei Cloud, to enhance its AI capabilities [4] Q&A Session Summary Question: How does the company assess the development pace of the AI industry and its own product rhythm? - The company categorizes Chinese manufacturers into two directions: general AI and AIGC, noting that domestic models are nearing GPT-4 levels [5] Question: What is the company's positioning in large models and C-end products? - The mission is to achieve general AI, focusing on text large models to meet demands in AI search and AIGC [5] Question: What are the reasons for the lack of blockbuster AI applications overseas? - The core reason is the reliance on paid music models, which limits growth speed [6] Question: What is the future product form of TianGong AI? - The company aims to create an AI Super APP that enhances user creativity and lowers content creation barriers [7] Question: What are the plans for promoting AI music products? - The company plans to launch an independent AI music product similar to a UGC community platform [8] Question: How is the testing data for the AI social application Linky? - Linky has exceeded 1 million downloads and ranks in the top ten in several countries [8] Question: What are the company's plans for video applications? - The company has long-term plans for AI video applications, including the release of a new multimodal model [9] Question: What characteristics will companies need to dominate the AI search market? - AI search is a new product form, and the company believes in a free information service model supplemented by advertising [10] Question: What is the expected capability of global models compared to 2023? - The company anticipates that its models will reach or exceed GPT-4 levels this year [11]