Runze Technology announced its results for the third quarter of 2024. The 1-3Q24 company achieved revenue of 6.4 billion yuan, a year-on-year increase of 139.1%, and realized a net profit of 1.5 billion yuan to mother, an increase of 35.1% over the previous year. Looking at a single quarter, 3Q24 achieved revenue of 2.8 billion yuan, a year-on-year increase of 184.0%, a year-on-month increase of 21.5%, and realized a net profit of 0.55 billion yuan to mother, an increase of 30.9% year-on-year and 11.2% month-on-month.

The big domestic model got off to a quick start, and the first signs of AI inference were on display. On October 25, Zhipu Huazhang released the latest model, AutoGLM (based on OpenAI AI Agent), which can perform various operations on mobile phones and websites. There is no need for users to perform manual operation demonstrations, and is not bound by simple task scenarios or API calls, and can replace humans in performing operations on electronic devices. Currently, AutoGLM is equipped with a variety of traffic applications, and large-scale daily activity is just around the corner. Currently, AutoGLM is compatible with 8 well-known applications including WeChat, Taobao, Meituan, and Xiaohongshu.

An OS-layer AI agent has been built, and vertical AI has been gradually implemented. Currently, AI operations implemented by Smart Spectrum are mainly focused on the OS layer, and it is expected to lead apps to open up vertical AI in the future. From a data perspective, the app grasps more user data, “understands” user psychology more, and has higher authority. From this perspective, App Nikkatsu is expected to become a weather vane for observing AI agents. We believe that as inference computing power gradually increases, inference tasks usually require frequent data exchange with other servers or data sources, so data centers require high-bandwidth network connections, which also place higher demands on IDC.

Seizing AIDC's new opportunities, the second growth curve achieved remarkable results. The size of computing power clusters has grown exponentially, and demand for intelligent computing has surged. The company keenly grasps the new opportunities brought to the IDC field by AI technology innovation, and attracted many leading AI customers to deploy core training modules in the company's intelligent computing center to achieve the goal of starting, growing and growing with large terminal models, and injecting new impetus into continuous rapid development. 1H24's AIDC business order size grew exponentially, with revenue of 2.05 billion yuan, an increase of 1148% over the previous year.

Innovative and optimized liquid cooling design leads the green development of the computing power center industry. As a leading “liquid cooling” enterprise, the company's self-developed technology is fully applied to new intelligent computing centers in multiple parks across the country, creatively completed the intelligent computing cluster infrastructure design with high-power, high-computing networks, high storage, and multiple modules, and provided all-round support for trillion-scale model training. Continuously optimize the design of liquid cooling systems to lead the industry towards a more efficient and environmentally friendly green intelligent computing center. Continued investment in green power transactions. The total volume of 1H24 green power transactions was nearly 0.4 billion kilowatt-hours, and two additional computing power centers won the title of “2023 National Green Data Center”.

Investment advice: The company is expected to achieve operating income of 6.14 billion yuan, 8.08 billion yuan, and 10.93 billion yuan in 2024 to 2026, and realized net profit of 2.47 billion yuan, 3.33 billion yuan, and 4.61 billion yuan, corresponding EPS of 1.44 yuan, 1.94 yuan, and 2.68 yuan, respectively, corresponding to the closing price of October 29 is 23 times, 17 times, and 12 times, respectively. Under industrial trends such as artificial intelligence, cloud computing, and 5G, data centers are expected to continue to grow at a high rate. Considering the company's outstanding comprehensive competitive advantage and sufficient resource reserves, it is expected to accelerate development. Optimistic about the company's business prospects and maintain a “recommended” rating.

Risk warning: Policy implementation falls short of expectations; industry competition intensifies; risk that performance promises may not meet standards.