What I am sharing today is the research report on the artificial intelligence AI industry:"The second half of AI, application landing, empowering all industries".
Featured Report**: The world of fantasy filmsSummary of the content of the research reportAs follows, in 2023, ChatGPT will explode across the entire social network, and the artificial intelligence industry will enter a stage of rapid development represented by large models. Foreign large models started relatively early, leading the development trend of the artificial intelligence industry, domestic large models closely follow the international frontier, and at present, the domestic representative large models have been comparable to ChatGPT in Chinese Q&A performance, and have the ability to initially commercialize and monetize. From the perspective of technological frontiers, AI as a whole will develop along a more efficient, intelligent, and segmented path, and important technology trends include hybrid expert model technology, AI agent, and robotics. The landing and application of AI in vertical fields will be the main line in 2024, and we are optimistic about the implementation of AI in education, autonomous driving, device-side equipment, and industrial scenarios. As artificial intelligence enters the era of large models with huge parameters, the growing demand for computing power has ushered in huge opportunities for the AI chip and server market. At the same time, we are optimistic about the localization of large-scale model technology bases, such as EDA and other subdivisions.
In 2023, ChatGPT will explode across the entire social network, and the artificial intelligence industry will enter a stage of rapid development represented by large models。Internet and artificial intelligence companies at home and abroad are actively engaged in the field of large model research, and have accelerated their layout in the field of generative large models. Foreign large models started relatively early, forming a situation of one super and many strong led by OpenAI + Microsoft, Google, Meta and other giants, leading the development trend of the artificial intelligence industry. At present, the representative large models in China are comparable to ChatGPT in terms of Chinese Q&A performance, and have the ability to initially commercialize and monetize. 
AI technology as a whole will look along the way:More efficient, smarter, and more segmented path development. We believe that the core technology trends around these three points are as follows:1) More efficient:Hybrid Expert Model Technology (MOE). Task implementation through gated network The distribution of training data among different expert models allows each model to focus on the task it is best at, which greatly improves the model's ability and training efficiency. 2) Smarter:AI agent. The AI agent receives information about the surrounding environment, makes plans, makes decisions, and takes action. The GPTS and Assistants APIs launched at the OpenAI conference on November 7, 2023 are an important step towards AI Agent, which will broaden the application boundaries of large models and is an important direction to AGI. 3) More vertical:Robotics. In recent years, Google, Meta, and Tesla have led the application of AI in the field of robotics. Robot control algorithms from traditional linear controllers, rule definitions, etc., are shifting to AI technology routes such as strong chemistry Xi, strong chemistry Xi + large language models, end-to-end large models, etc., which has greatly accelerated the development of the robot industry. 
The landing and application of AI in vertical fields will be the main line in 2024, and we are optimistic about the implementation of AI in education, autonomous driving, device-side equipment, and industrial scenarios. The education industry has become a high-quality field of artificial intelligence because of its strong demand for personalized learning Xi, high data richness, and strong willingness to pay。With the explosion of generative AI technology, AI+ education has entered a new stage of development. According to Market Research, the market size of generative AI in the education sector will increase from 2$1.5 billion, up to $27 in 2030$400 million, with a CAGR of 375%。We believe that AI+ education will take the lead in landing in three scenarios: software, education informatization, and intelligent hardware. Focusing on the two needs of learning Xi machine and overseas education informatization. The algorithm architecture is optimized, and the penetration rate of high-end assisted driving is expected to increase。Tesla's latest end-to-end model for autonomous driving, FSD V12, is completely trained on a neural network, without a single line of human-written rules**. The end-to-end architecture is a new algorithm architecture for autonomous driving, and by optimizing the autonomous driving algorithm architecture, the corresponding autonomous driving performance of the corresponding vehicles is expected to achieve a significant increase, thereby driving the penetration rate of high-level autonomous driving to a certain extent. The penetration rate is increasing due to the support of technology, and the era of fully autonomous driving is gradually approaching, and domestic enterprises will gradually start globalization. The core of cloud AI lies in mobile phones and PCs, and AI phones and AI PCs will open a new era. From the Mobile World Conference held in February this year, Qualcomm showed its mobile phone offline running model, to the Microsoft Developer Conference Qualcomm in May to display its PC running AI model, and then to the recent release of AI PC acceleration plan by Intel and Lenovo, etc., It can be seen that domestic and foreign manufacturers continue to make efforts to AI Phone and AI PC, and the end-side AI will enter a new era. Intelligent manufacturing is the trend of the times and the only way for an industrial country to transform into an industrial power。It is recommended to pay attention to the following three sub-tracks: 1) Industrial machine vision: SAM models and GPT-4V have been applied in industrial quality inspection, industrial automation and other fields, with significant improvements in efficiency and accuracy, and AI is expected to further improve the penetration rate of machine vision. At the same time, the recovery of downstream industries such as consumer electronics and the increase in the proportion of visual investment such as lithium batteries will bring more industry demand. 2) Industrial robots: mobile robots such as A** are rapidly entering industrial scenarios to replace traditional manual handling and inspection. Intelligent welding robots have been applied in downstream industries such as steel structures through automatic identification of workpieces, gap finding, path planning, and motion control, filling the gap of a large number of welders caused by the rapid growth of steel structures and other industries. 3) Industrial software: Because of its high technical barriers, it has been monopolized by overseas giants in the past. We are optimistic that under the technological trend of AI + cloudification, with the gradual increase in the localization rate and the legalization rate, more industrial opportunities will emerge. 
As artificial intelligence enters the era of large models with huge parameters, the increasing demand for computing power has ushered in huge opportunities for the AI chip and server market. The U.S. Department of Commerce and the Security Agency issued a new round of chip export restrictions on October 17, 2023, and advanced chips such as NVIDIA are subject to export controls. Although domestic manufacturers have a gap with the former in terms of hardware product performance and industrial chain ecological architecture, they are gradually improving their product layout and ecological construction, and constantly narrowing the gap with industry leading manufacturers. As the demand for computing power continues to increase, the AI server boom will also continue. (Report producer: China Securities Construction Investment).
The following is an excerpt from the original report:
This article is for informational purposes only and does not represent any investment advice from us. The world of fantasy filmsOrganizing and sharing information is only recommended for reading, usersInformation obtainedFor personal Xi only, please refer to the original report for use. The report totals:82 pages。Due to space limitations, only some of the contents are listed.