Large wind power flange faucet, explore the new layout of computing power leasing. The company is mainly engaged in R&D, production and sales of rolled ring forgings, forged flanges and other free forgings. The products are mainly used in wind power industry, petrochemical industry, metal pressure vessel industry, machinery industry, ships, nuclear power and other industries. The company currently has three products: annular forgings, forged flanges, and free forgings. Its core business is annular forgings, specifically including wind power flanges, supporting parts for the construction machinery industry, and large-scale ring forgings for metal pressure vessels. In order to adapt to the wave of the digital economy and enter a new circuit of digital energy computing power integration, the company and Shanghai Liuzhu Technology Group Co., Ltd. jointly funded the establishment of Shanghai Run Liuzhu Technology Co., Ltd., and plan to cooperate to establish computing power centers in Shanghai, Fuzhou Economic Development Zone, Anhui, and Jining, Shandong, etc., and build a Yangtze River Delta GPU computing center cluster.
Localization of wind power bearings, computing power drives the growth of the industrial chain. In recent years, wind power industries at home and abroad have begun to develop in the direction of specialization, scale and intensification, and the demand for refined, high-intensity and standardized products has continued to increase. China has classified the wind power industry as one of the country's strategic emerging industries. Under the dual role of industrial policy guidance and market demand, the national wind power industry has achieved rapid development. We believe that ChatGPT reflects the current improvement in AI training speed and training accuracy. The fundamental logic behind it is supported by a huge set of training data, and at the hardware level, computation and transmission are the core of ensuring the smooth operation of ChatGPT.
According to Shell Finance Consulting, on February 9, 2023, Beijing time, due to overloaded training, ChatGPT's official website could not be logged in. The page displayed “Overloaded”, confirming that the current network resources/computing power cannot guarantee the massive influx of computing power requirements in the future. The total computing power consumption of ChatGPT is about 3640PF-days. From the perspective of Moore's Law, the rate of development of AI represented by ChatGPT has increased exponentially, doubling about every 3 to 4 months (Moore's Law doubling cycle is about 18 months).
The wind power business is picking up, and the company's revenue is expected to increase. Although the wind power industry in the first half of 2022 was affected by factors such as affordable Internet access and construction progress, wind farm construction started cautiously, and order release slowed. However, judging from the strong policy support given to the wind power industry by national and local governments over a long period of time, the development of the domestic wind power industry has entered a stage of steady development. Offshore power generation and stock market renewal due to the imminent retirement of early wind turbines are opportunities for future development of the wind power industry. In 2014-2022, the company's revenue was 12%, and domestic onshore wind power was installed in 2020. The company achieved revenue of 2,385 billion yuan, the highest year-on-year growth rate in history; net profit increased from 400 million yuan in 2014 to 0.9 billion yuan in 2022, CAGR 9%, and net profit of 440 million yuan in 2020, an increase of 459% over the same period, significantly higher than the revenue growth rate during the same period, reflecting the company's good profitability.
Investment suggestions: We are optimistic about the company's traditional business fundamentals and the future prosperity of the computing power industry. We expect the company's second growth curve in computing power leasing in 23-25 to be 28.4/38.2/4.79 billion yuan, respectively, and net profit of 3.0/5.2/740 million yuan, respectively, corresponding to the closing price of November 10, 2023, which is 64x/37x/26x, respectively.
Risk warning: Technology development falls short of expectations; increased risk of competition in the industry; risk of core personnel reserves and loss.