Abstract:As the core of the fifth research paradigm, AI for Science has been widely used in multiple research fields of natural sciences and high technologies. In contrast to that the application of AI in natural sciences mainly focuses on discovering new theories, principles, and laws, the application of AI in high technologies mainly focuses on creating new plans, tools, and products, in order to resolve concrete problems in related fields. This study first summarizes the typical characteristics and scientific problems of the application of AI in high technologies, i.e., AI for Technology, and then introduces a successful case study of AI for Technology, that is, fully automated CPU design. Finally, this study points out that the main targets of AI for Technology are not only to accelerate the innovation process and thus reduce human investments, but also to endow machines with higher creative abilities than human experts eventually.