SIASUN(SZ:300024) Xin Lang Cai Jing·2025-06-18 05:15Core Viewpoint - The partnership between Innovusion (02121.HK) and KUKA Robotics marks a significant step in the integration of AI and manufacturing, focusing on the development of intelligent industrial robot systems driven by the AInnoGC industrial model [1][10]. Company Overview - Innovusion was established in 2018 and is backed by Innovation Works, specializing in enterprise-level AI solutions and "AI + manufacturing" solutions across various sectors including steel metallurgy, semiconductor panels, high-tech, automotive equipment, energy, food and beverage, and new materials [3]. - The company's development strategy is based on a "one model, one body, two wings" approach, leveraging its self-developed AInnoGC industrial model as the technical foundation and AI agents as the driving force in industrial software and robotics [4]. Industry Trends - The integration of AI into industrial robotics is increasingly recognized as a trend, with major companies like ABB and others planning to deepen their collaboration with AI technologies by 2025 [5][6]. - The concept of industrial embodied intelligence robots, which utilize large models to enhance robotic capabilities, is gaining traction, allowing robots to adapt to flexible production and complex tasks [6]. Market Position - Innovusion ranks third in China's computer vision market, fourth in machine learning platforms, first in AI industrial quality inspection solutions, and seventh in the large model application market according to IDC [7]. - Despite its strong market position, Innovusion has not yet achieved profitability, reporting adjusted net losses of 142 million, 138 million, 154 million, and 117 million yuan from 2021 to 2024 [8][9]. Financial Performance - The company reported a 21.1% decrease in R&D expenses in the previous year, amounting to 355 million yuan, attributed to improved efficiency and effective management of R&D spending [9]. - Innovusion emphasizes that the intelligent industrial robots powered by large models are still in the early stages of application, with the market landscape yet to be fully established [10].