VEP40.3CEN-256NN-MAD-128-NN-FW By reviewing the development process of global industrial robots, we see that industrial robots have experienced the evolution process from special to general, from mechanical to intelligent, and from single to complex. As industrial robots began to be applied to various downstream industries, gradually replacing human repetitive labor, the industrial value began to gradually emerge. Taking ABB as an example, from the point of view of dismantling the industrial chain, the core components of industrial robots include sensing, execution and decision-making, among which the IT architecture is the core of determining the degree of robot intelligence, and the importance of software is expected to gradually increase. At the same time, with the rise of big predictive models such as ChatGPT, robots are beginning to evolve toward embodied intelligence, and industrial change is expected to accelerate.

Core idea

Industrial robots are gradually becoming intelligent/universal, and the industrial application prospect is promising

Global industrial robots have roughly experienced three VEP40.3CEN-256NN-MAD-128-NN-FW stages of development: 1) embryonic stage of technology: the initial industrial robots are mainly used in military, nuclear industry and other high-risk areas, showing that teaching and reproduction robots are mainly; 2) Industrial rise stage: ABB, Kuka, Fanuc and other leading enterprises began to rise, and industrial robots began to have the ability to perceive and process external information; 3) Industrial upgrading stage: Extending to the general industrial scene, and continuously improving the level of intelligence. VEP40.3CEN-256NN-MAD-128-NN-FW According to IFR, 517,000 new industrial robots will be installed globally in 2021, an increase of 31.2%. According to Inkwood Research, the annual CAGR of global industrial robot output from 2021 to 2028 will reach 14.7%. We believe that with the continuous upgrading of industrial robot intelligence, the contraction of population supply and the release of downstream demand drive the growth of the industry, and the prospect of industrial application is promising.

VEP40.3CEN-256NN-MAD-128-NN-FW

IT architecture is the core of determining the degree of robot intelligence, and the importance of software is expected to gradually increase

VEP40.3CEN-256NN-MAD-128-NN-FW Taking ABB as an example, from the point of view of dismantling the core link, the industrial robot industry chain includes upstream core components, midstream robot ontology and software, downstream system integration and application, and the main cost is concentrated in the upstream core components such as the reducer and servo system. From the perspective of functional classification, industrial robot components can be divided into three layers: perception, execution and decision-making. ABB has a layout at all levels, such as sensors, servo systems/controllers, and algorithm software, forming a full-stack solution for industrial robots. We believe that with the development of industrial robots to intelligence, IT has gradually entered the software-defined stage, and robots are gradually evolving into IT products. The differences in software algorithms will have an important impact on the level of intelligence that robots can eventually achieve, and the status of software is expected to gradually improve.

AI empowerment is expected to accelerate industrial change

We can see that for simple robot tasks, ChatGPT has been able to solve in a zero-shot way, and for complex tasks such as course learning and AirSim obstacle avoidance, ChatGPT can be implemented under the on-the-loop interaction of human users. From the perspective of the landing in the industrial field, embodied intelligence is still in the early stage of industrial development, and AI applications are concentrated in some simple scenarios. VEP40.3CEN-256NN-MAD-128-NN-FW Taking ABB as an example, in February 2020, ABB cooperated with Covariant to launch an intelligent sorting robot to improve recognition efficiency and accuracy through machine vision algorithms. In addition, in July 2023, ABB announced a partnership with Microsoft to apply GPT-4 to scenarios such as industrial analysis. Based on this, we believe that with the emergence of large models such as ChatGPT, robots are beginning to evolve towards embodied intelligence, and industrial change is expected to accelerate.