China Securities Co., Ltd.: "Prospects for Investment Opportunities in Artificial Intelligence+"

AI+ Education In the global wave of technological revolution and educational transformation resonance, artificial intelligence is deeply reconstructing the education ecosystem and has become the core engine driving the digitization upgrade of education. The education sector, due to its clear scenes, rich data, and rigid demands, has become the golden track for the landing of AI technology. In terms of scenes, the quantifiable nature of the entire teaching process enables AI to seamlessly embed in the core links of "teaching, learning, assessment, and management" through intelligent lesson preparation systems and adaptive learning engines to achieve efficiency enhancement; in terms of data, the massive multimodal behavioral data naturally precipitated during the education process provides the foundation for building a holistic portrait of learners and precise decision-making; in terms of demand, the national strategic goals and deployment of nearly a billion terminal devices form a dual drive of policy and market, overlaid by the rigid demand of the billion K12 group for high-quality resources, jointly establishing the demand base for the landing of technology. The deep coupling of these three factors is catalyzing the paradigm upgrade of education from tool empowerment to ecosystem reconstruction, forging a new high ground of new quality productivity with both fairness and a scale of hundreds of billions. The core advantages of AI education are reflected in the following aspects: 1. Paradigm innovation of teaching modes. By mining real-time learning data through intelligent algorithms, AI systems can accurately identify individual learning differences, dynamically generate adaptive learning paths, and shift traditional "one-size-fits-all" teaching to customized cultivation for each individual. This data-driven personalized teaching significantly reduces the marginal cost of tailored teaching. At the same time, the deep integration of AI technology reshapes the boundaries of teaching, relying on its powerful multimodal content generation capabilities to dynamically construct diverse course forms, significantly enhancing the immersive experience and cognitive involvement of the learning scenario. 2. Scale-up of educational efficiency. In the classroom teaching scenario, multimodal data analysis technology can capture students' engagement levels and knowledge mastery states in real time, allowing teachers to dynamically adjust the pace of teaching and improve the efficiency of interaction between teachers and students in a single lesson. In the post-classroom stage, automated homework correction systems provide fast feedback, improve teachers' grading efficiency, enhance teaching efficiency, and address the long-standing contradiction in the education field of balancing "scale expansion" and "personalized service." 3. Universalization revolution of educational resources. By carrying well-known school courses and virtual teacher services on low-cost terminals, the threshold for accessing high-quality educational resources in remote areas is reduced, significantly narrowing the regional education gap. After AI technology takes over mechanical work, teachers can shift to creative teaching and humane care, promoting the transition of educational resources from "basic coverage" to "quality balance." At the same time, technology enhances the efficiency of resource allocation between regions through open platforms and data sharing mechanisms, forming a sustainable path for fair progress. The explosive development of generative artificial intelligence technology is driving AI + education into a new stage. According to Market Research forecasts, the market size of generative AI in the education sector will soar from $215 million in 2022 to $2.74 billion in 2030, with a high compound annual growth rate (CAGR) of 37.5%. Among them, applications targeting learners (students) contribute nearly half of the market share, becoming the core engine of industry growth. The "double reduction" policy combined with policy dividends empowers the development of the AI + education industry. In 2021, the "double reduction" policy aimed to alleviate the heavy academic burden of primary and secondary school students and restrict the excessive development of subject-specific training institutions while emphasizing the public welfare attributes of education. This policy led to a large-scale clearance on the supply side of the industry, with many small and medium-sized training institutions closing or transforming, while industry leaders began adjusting their business directions, shifting towards quality education, vocational education, and compliance in the middle and high school exam retake market. With marginal relaxation of policies since 2024, the education industry has seen new development opportunities. From 2024 onwards, the policy orientation has shifted from "restriction" to "norm," no longer restricting the approval of new institutions, and encouraging the development of non-subject-specific training. In 2024, the "Outline of Building an Education Powerhouse" for the first time set rigid goals such as "AI assistants covering 90% of mandatory education schools" and "building a dedicated education large-scale model." In 2025, the "Opinions on Accelerating the Digitization of Education" from nine departments further detailed "subject-specific model research and development (mathematics/ideology first)," "popularization of intelligent learning companions," and "algorithm security filing system" as the three main focuses, signaling that the policy focus is shifting from tool application to system restructuring. At the same time, the policy supports educational and training institutions to participate in school after-school services, providing a clearer development path for the industry, promoting the deep integration of AI with the teaching process, and enhancing education equity and quality. Educational Informatization Educational informatization has now entered the 2.0 era, with deep integration of technology, comprehensive extension of scenarios, and diversified service objects as core features, driving education from "tool empowerment" to "ecosystem reconstruction." In this stage, teaching scenarios are no longer limited to traditional classrooms, but extend to after-school services, teaching management, educational decision-making, and the entire ecosystem. The service objects have expanded from students and teachers to educational managers and policy makers. In terms of technology applications, artificial intelligence, big data, cloud computing, 5G, and other technologies are gradually replacing single hardware integration, and becoming the central nervous system of educational informatization. For example, AI learning systems use emotion computing and biometric identification technology to provide personalized learning path recommendations, significantly enhancing learning efficiency. In terms of data, AI technologies leverage vast multimodal behavioral data naturally accumulated during the education process to build accurate learner portraits and provide precise decision-making support. With continuous optimization of hardware devices and platforms, the entire educational ecosystem is experiencing rapid changes. More advanced AI tools and technologies are being integrated into educational practices, leading to more personalized and adaptive learning environments. AI-driven educational software, such as intelligent tutoring systems and adaptive assessment platforms, are playing an increasingly important role in providing personalized, engaging, and effective learning experiences. Educational software is the product of long-term accumulation and evolution of industrial innovation knowledge, relying on the integration of software and hardware to achieve data fusion and real-time decision support. Existing educational software is based on a robust system for acquiring, processing, and analyzing data from multiple sensors, enabling informed decision-making, resource optimization, and enhanced teaching and learning experiences. In the educational context, visualization tools help educators and administrators understand student performance and progress, providing valuable insights for personalized instruction and continuous improvement. Innovative educational software solutions are driving the transformation of traditional teaching methods and educational practices, paving the way for a more efficient, interactive, and student-centered learning environment. AI+ Industrial Artificial intelligence is gradually being recognized as a decisive force that is changing today's world and impacting the future global landscape. This is not only directly demonstrated in the direct transformation of the economy, society, and the military field by artificial intelligence technology, but also indirectly influences international strategies, technological competitions, and great power games. Various countries are exploring the use of different types of AI tools to enhance battlefield performance in different military branches and at different levels. The use of AI-assisted Beijing Vastdata Technology libraries for battlefield situational awareness and decision-making has become the focus of current attention. Summary of AI battlefield usage scenarios 1) Autonomous combat systems. AI-driven unmanned systems (such as drones, unmanned ground vehicles, etc.) can perform tasks such as reconnaissance, strikes, and logistics on the battlefield, reducing the direct exposure risk to human soldiers. They can autonomously operate on the battlefield, adapting to complex conditions and responding at a faster pace than humans. The US military uses autonomous drones (such as the MQ-9 Reaper drone and the X-47B unmanned combat aircraft) in multiple projects, which not only conduct reconnaissance missions but also carry out strike missions. The US is also developing autonomous unmanned ground vehicles (UGVs) for adaptive autonomous operations. These systems can patrol, conduct reconnaissance, etc., with no human intervention. 2) Situational awareness and real-time decision-making. The battlefield situation is constantly changing, and AI can integrate information from multiple sensors (such as satellite images, drone data, ground reconnaissance, etc.) for real-time battlefield awareness. AI can provide real-time battlefield awareness by analyzing patterns and data, helping commanders make more precise decisions and optimizing resource allocation. By automatically processing sensor data in real-time, AI can identify and classify targets on the battlefield, enhancing precision strike capabilities. Especially in complex environments, AI can quickly identify high-value targets from massive target information, optimizing attack timing and resource usage. The US military currently uses AI-driven target identification and precision strike technology, such as "MAVEN," which uses image analysis data to identify high-value targets on the battlefield. This system is primarily used for strike missions by drone formations, automating the identification and attack of high-value enemy targets. 3) Operating in non-ideal environments. Artificial intelligence technology can be applied in special environments, such as nuclear radiation, high temperatures, lack of oxygen, etc., where humans cannot operate for extended periods. AI technology can improve operational efficiency and safety in these environments, which are challenging and dangerous for humans, such as mine sweeping robots, etc. 4) Intelligence analysis and prediction. AI can quickly extract valuable information from a large amount of open-source intelligence, surveillance data, etc., and predict trends. Through pattern recognition and data mining, AI can predict enemy behavior, tactical changes, possible battlefield areas, and provide decision-making support. The US Defense Advanced Research Projects Agency (DARPA) has developed several AI-based intelligence analysis systems, which can extract information from social media, communication data, satellite images, etc., predict enemy movements and intentions, and provide decision support. Anduril Pulsar electronic warfare system: A defense system against small and medium-sized drones Anduril Industries, founded in 2017 by Palmer Luckey, the founder of Oculus VR, is dedicated to developing advanced AI-driven military technologies. The company is known for its autonomous systems, sensor fusion, and AI platforms, especially in anti-drone, electronic warfare, command and control, and extended reality (XR) fields. Anduril's AI-driven systems currently consist of several components. Its core product is the Lattice platform, an open AI operating system that serves as the core of command and control and task autonomy. Lattice supports the integration of various sensors and platforms to achieve data fusion and real-time decision support. In the anti-drone field, Anduril's Pulsar system, launched in 2024, can quickly identify and counter current and future threats using AI, especially small and medium-sized drones. The Pulsar system can rapidly identify new threats and formulate corresponding defense measures using AI tools, significantly reducing response times. After a system identifies a new threat signal, it is analyzed within a few hours, and response strategies are pushed to other systems for quick sharing and coordination