Industrial AI is a broader field than embodied intelligence, and its potential market size is even larger.

Industrial scenarios have always been one of the most important areas for the commercialization of AI. In the past two years, many companies have begun to widely apply AI technology on devices, in data centers, and in online human-machine interfaces (HMIs). According to IDC forecasts, the penetration rate and market share of AI are rapidly increasing, whether in general-purpose software, industrial programming software, industrial vision software, or other industry software. Correspondingly, the demand for AI computing power, whether on the device side, on the device side, or in data centers, will also continue to expand.

In addition to embodied intelligence, specific application scenarios for industrial AI also include the following:

Machine vision: As a pioneer in the implementation of artificial intelligence, convolutional neural networks (CNNs) have long been widely used in classification and detection throughout its development. In recent years, with the rise of large models, more AI technologies have been introduced to machine vision, such as using big data models to generate samples or detect anomalies, thus overcoming the limitations of traditional CNN networks, which require retraining when faced with new problems.

Industrial control: In the control field, classic algorithms were previously used. However, in the past two years, reinforcement learning has gradually become a popular deployment trend, both in robotic motion control and other traditional controllers.

Industrial digitalization: Previously, the focus was on functions such as production scheduling optimization and operations and maintenance. These functions, based on machine learning, have gradually incorporated more AI algorithms. In the past two years, the rise of large models has led to significant progress in RAG (Retrieval Augmentation Generation). Global leading manufacturers, as well as many domestic ODMs and even ISVs, have widely adopted RAG in their software products to reduce costs and increase efficiency.