Currently, the global technological revolution is accelerating, and China is in a critical period of transitioning from a manufacturing powerhouse to a manufacturing powerhouse. The deep integration of artificial intelligence and manufacturing has sparked a fierce competition for the "right to define" the physical world. Faced with the challenge of developed countries competing for dominance in high-end manufacturing, how to make good use of China's super large market advantage and the huge number of "engineer dividends", and build a new type of production relationship that adapts to the digital age, is a major issue that must be answered during the "15th Five Year Plan" period. Research has found that in the leading regions represented by the Pearl River Delta, a group of innovative entities have emerged that transform non-standard innovation into standard services, exploring a new model of "new generation innovation infrastructure". This model achieves exponential reduction in innovation costs and exponential improvement in conversion efficiency through software defined manufacturing and industrial chain element reconstruction, providing important insights for seizing the power to define future industries. Relying on China's complete industrial system, accelerating the construction of a new generation of innovation infrastructure that integrates "domestic software soul casting, flexible intelligent manufacturing shaping, and innovative ecological empowerment" is the key path to achieving the leap from the "world factory" to the global innovation source. Deeply grasping the fundamental transformation of the global industrial competition logic, the injection of new production factors such as data, computing power, and algorithms is driving a fundamental shift in the global industrial competition logic. The competitive form has shifted from a "smile curve" to a "dumbbell model". The "smile curve" theory of the industrial era suggests that the manufacturing process has the lowest added value. But in the era of artificial intelligence, the industrial value chain presents a "dumbbell" shape: one end is cloud computing power and large models, and the other end is perception and execution of the physical world. As artificial intelligence infiltrates the edge, whoever can help global developers create intelligent hardware with the lowest threshold holds the power to define data sources and application scenarios. The phenomenon that more than half of the technology projects on international crowdfunding platforms rely on the Chinese supply chain indicates that China has the ability to leap from "Made in China" to "Defined in China". Factor driven transformation from "demographic dividend" to "engineer dividend". Since the reform and opening up, China has accumulated a complete supply chain system and a large team of science and engineering talents. Data shows that China trains over 5 million graduates in science and engineering every year. However, in traditional manufacturing models, a large number of talents are confined to low-end links. The data of over 8 million registered users and over 1 million active paying users on the hardware innovation service platform created by Jialichuang proves that once there are low threshold innovation tools, this huge "army of engineers" will burst with astonishing vitality. The organizational model is shifting from a "linear chain" to a "networked ecosystem". The traditional linear model of "research and development manufacturing sales" is no longer suitable for the demand of "small order quick response" in the digital economy. The Pearl River Delta region has successfully incubated a group of unicorn enterprises in the field of intelligent hardware through the model of "new engineering education+supply chain integration". This indicates that building an open, shared, and collaborative "tropical rainforest style" innovation ecosystem has become a core means of competing for the discourse power of high-end manufacturing. The theoretical connotation and practical path of the new generation of innovation infrastructure. The so-called new generation of innovation infrastructure refers to the use of new generation information technology to digitize and encapsulate the entire life cycle capabilities of innovation, such as research and development design, pilot testing, and small batch manufacturing, making it a public service that is "accessible, on-demand, and autonomously controllable". Its theoretical connotation and practical effectiveness are reflected in three dimensions. The dominant position in reconstructing the technological foundation with "software defined". The soul of infrastructure lies in standards, and platforms without proprietary software are just "shells". From a theoretical logic perspective, promoting domestic industrial software and underlying technical standards to "code" process rules is the key to solving the problem of "chip shortage and soul shortage". From a practical perspective, independently developing electronic design automation software and opening it up for free can enable millions of engineers to follow domestic process standards at the design stage. Meanwhile, open-source operating systems and open instruction set architectures are becoming the technological foundation for the Internet of Things. This "domestic software+platform service" model has forced the domestic technology stack to iterate and mature in a massive number of applications, achieving self-reliance and self-improvement of the technological foundation. Realize the dialectical unity of economies of scale and scope through "flexible intelligent manufacturing". Resolving the contradiction between personalized customization and large-scale production is a world-class challenge for the transformation of the manufacturing industry. From a theoretical logic perspective, utilizing big data and artificial intelligence algorithms to consolidate massive and fragmented personalized demands into large-scale production can achieve the optimal solution of cost and efficiency. From a practical perspective, through intelligent bundling technology, thousands of circuit board sample orders can be pieced together for production, reducing trial and error costs by 90% and shortening delivery cycles to 12 hours; By installing IoT devices on the weaving machine, cross factory capacity cloud scheduling has been achieved; By utilizing digital tools to collaborate with thousands of factories, we have achieved the ultimate efficiency of "quick response to small orders". These practices have proven that digital restructuring can rejuvenate traditional production capacity. Promote the deep coupling of talent chain and industry chain through the integration of industry and education. Talents are the first resource for innovation, and infrastructure should not only be easy to use, but also have people to use it. From a theoretical logic perspective, by combining methodological output with infrastructure support, we can lower the threshold for hard tech entrepreneurship and cultivate innovative entities. From a practical perspective, "New Engineering" education not only imparts knowledge, but also relies on the supply chain of the Greater Bay Area to enable students to make products with real skills. This dual wheel drive model of "innovative methodology+innovative infrastructure" effectively connects university research and industrial transformation, solving the chronic problem of "many laboratory achievements and few industrialized products". The policy recommendation for building a new industrial digital base is to construct a new generation of innovative infrastructure, which is not only a technical issue, but also a profound transformation of production relations. We must adhere to a systematic approach, coordinate the "hard infrastructure" and "soft environment", and embark on a new path of industrialization with Chinese characteristics. Strengthen top-level design and promote the infrastructure of the pilot platform. It is suggested to include platforms with the characteristics of "software self-developed, open sharing, and industry empowerment" in the scope of national new infrastructure construction. Change the policy orientation of simply supporting production line technological transformation in the past, and focus on supporting the construction of a digital pilot platform open to the whole society. In the field of electronic information, promote the pilot platform of mechatronics integration, integrate precision injection molding and CNC machining production capacity, and provide "building block" sampling services for embodied intelligent entrepreneurs. In the traditional industry field, guide characteristic industry clusters to establish shared intelligent manufacturing empowerment centers, use artificial intelligence visual inspection to unify standards, use shared equipment to reduce costs, and solve the dilemma of homogeneous competition. Adhere to self-reliance and controllability, and implement the "domestic technology application traction" project. The scenario is an accelerator for the maturity of new technologies. We should fully leverage the advantages of a super large market, establish special guidance funds, and promote the "hardware innovation voucher" system. In terms of mechanism innovation, universities, research institutions, and small and medium-sized enterprises are encouraged to prioritize the use of domestic electronic design automation software, domestic operating system modules, and domestic chips in the innovation process. Provide corresponding subsidies for the cost of using domestic technology stacks for sample verification. By providing "government procurement services", we aim to cultivate the first batch of "seed users" for the domestic software and hardware ecosystem, and polish the "national key equipment" in practical situations. Deepen the integration of industry and education, and cultivate a fertile ground for the "engineer dividend". Promote the experience of "new engineering" education reform, and support universities and industrial Internet platforms to jointly build future technical colleges. Incorporate domestic industrial software applications and supply chain management into the engineering education system, hold high-level global hard technology developer conferences, and enable students to become accustomed to using Chinese infrastructure for innovation during their school years, cultivating a large number of compound talents who understand both algorithms and manufacturing. Optimize the evaluation system and establish a "compass" for high-quality development. Explore the establishment of an evaluation index for manufacturing innovation service capabilities, and regularly release the "pilot response speed", "innovation cost reduction rate", "domestic technology penetration rate", and "achievement transformation efficiency" indices for key regions and industries. Breaking away from the "GDP only" theory, guiding localities to shift from simply "competing for investment and land" to "competing for ecology and services", and creating a favorable environment conducive to the emergence of innovative factors. If high-speed rail and bridges are the infrastructure of the physical world, then the new generation of innovative infrastructure that integrates domestic software, flexible manufacturing, and open source ecology is the new infrastructure of the digital economy era. Building this infrastructure is a strategic choice to adapt to the requirements of new quality productivity development and reshape industrial competitive advantages. With a sense of urgency, we should make good use of the complete industrial system and valuable human resources, firmly grasp the initiative of development in our own hands, and provide solid material and technical support for the realization of Chinese path to modernization. (New Society)