维持“增持” 评级，下调目标价至49.57 元。根据最新业绩情况，我们下调并新增盈利预测，预计2024-2026 年公司营收分别为5.99(-0.86)、7.33(-2.29)、8.81 亿元，基于下调后的盈利预测，下调目标价至49.57 元，对应2024 年10 倍PS，维持“增持”评级。Introduction to this report:
The company's performance remains relatively stable, the revenue structure continues to improve, gross margin has increased markedly, the AI upgrading and transformation of the company's products continues to deepen, and it is expected to exceed expectations in the future.
Key points of investment:
Maintaining the “Overweight” rating, the target price was lowered to 49.57 yuan. According to the latest performance situation, we have lowered and added profit forecasts. The company's revenue for 2024-2026 is 5.99 (-0.86), 7.33 (-2.29), and 0.881 billion yuan, respectively. Based on the lowered profit forecast, the target price was lowered to 49.57 yuan, corresponding to 10 times PS in 2024, maintaining the “gain” rating.
The company's revenue and net profit remained relatively stable, and losses were reduced slightly after deducting net profit not attributable to mother. In the first half of 2024, the company achieved revenue of 0.14 billion yuan, a year-on-year increase of 1.58%, achieved net profit of 0.191 billion yuan, a year-on-year decrease of 0.92%, achieved net profit deducted from non-mother of 0.206 billion yuan, a slight loss reduction of 2.02%, and remained relatively stable overall.
The revenue structure continued to improve, and gross margin increased markedly. In the first half of 2024, the revenue of the basic software business with a high degree of commercialization and high gross margin increased 22.00% year on year, accounting for 59.78% of revenue, up 10.1 percentage points year on year. Judging from the downstream customer composition, the revenue contribution of financial industry customers with stronger purchasing power reached 39.69%, an increase of 7.31 percentage points over the previous year. The improvement of the product sales structure and customer structure jointly promoted the company's comprehensive gross margin to 58.80%, an increase of 7.54 percentage points over the previous year.
The AI upgrade and transformation of the company's products continues to deepen, and it is expected to exceed expectations in the future. Faced with fluctuations in the market environment, the company continues to increase the AI upgrading of products. In the big data and cloud-based platform product line, the newly launched TDH4.9 version was further upgraded in the Hecangji integrated architecture, adding support for the Python ecosystem, improving the functionality and ease of use of large models and AI application scenarios. At the same time, the distributed vector database Hippo 2.0 built into TDH9.4 supports the storage, indexing, and management of 10 billion vector data converted from text, images, audio and video, greatly improving the efficiency of large-scale model training. In the big data development and intelligent analysis product line, the Sophon LLMops1.3 version of the company's big model operation platform supports the unified listing of multi-modal models, large models, traditional machine learning and deep learning models, which can better meet the management needs of large model customers. The company also launched the Starlink knowledge platform TKH to provide enterprise customers with the ability to manage AI computing power facilities, corpus, models and applications. The company gradually grew from a digital infrastructure supplier to an AI infrastructure vendor. As AI demand continues to be implemented, the company's business is expected to grow beyond expectations.
Risk warning: Risk of AI product iteration failure, risk of increased market competition.
