《AI共生，有色“需图”重构》系列报告（一）：AI浪潮来袭，撬动有色需求的下一个支点？

Report Industry Investment Rating - Not provided in the content Core Viewpoints of the Report - The rise of AI computing power has opened up new demand space for non - ferrous metals, and the pricing logic and demand landscape of commodities may be reconstructed. The new development in the AI field is becoming an indispensable new driving force for the growth of non - ferrous metal demand. At present, one should maintain sufficient imagination for the non - ferrous metal demand prospects brought by AI while closely tracking the micro - demand and high - frequency data of each variety [2][3][40] Summary by Relevant Catalogs 1. AI Industry from "Algorithm Economy" to "Industrial Integration" 1.1 AI Industry Development History Review - The AI industry has generally gone through six stages: start, reflection, application, depression, steady and booming development. Before 2011, it was in the exploration and accumulation stage focused on algorithm theory and early model verification. From the 1950s to the 1990s, it experienced two booms and troughs. After the 1990s, new technologies injected new life but had limited impact on hardware and materials [7] - From 2012 - 2023, AI entered the breakthrough and expansion periods driven by deep learning and computing power. It was mainly "algorithm - economy" centered, with capital and technology concentrated on software and chips, indirectly affecting the traditional industrial chain [8] - In 2024, it became a double - inflection point of technology and demand. The AI industry entered the acceleration period of industrial integration and infrastructure construction, with the growth focus shifting from "algorithm dividends" to "computing power bottlenecks", and from "virtual innovation" to "physical expansion" [10][11] 1.2 Current State of AI Commercialization - After the "computing power arms race" and capital frenzy, the AI industry has entered a new stage. Commercialization and verifiable profit models have become key factors, and the global commercialization path shows differences [12][13] - In terms of Sino - US development comparison, although the US still leads in some aspects, China is catching up rapidly. The decline in reasoning costs provides conditions for AI commercialization and large - scale application, and the financial reports of tech giants verify the entry into the second half of AI commercialization [14][20][24] - Tech giants are still increasing capital expenditure on computing power and data centers. It is expected that the total capital expenditure of tech giants will exceed $350 billion in 2025, $450 billion in 2026, and $500 billion in 2027 [26] 2. AI Industry Development Spurs New Demand for Non - Ferrous Metals 2.1 Panoramic Insight into the AI Industry Chain - The AI industry chain is a multi - level system composed of the basic layer, technology layer, and application layer, with non - ferrous metals running through it [28] - The basic layer provides computing power, with chips as the core carrier, and AI servers, optical modules, and data centers as important infrastructure, promoting the coordinated development of related industries [29] - The technology layer transforms basic resources into practical capabilities, with algorithm paradigms evolving, and multi - modal large models emerging. The application layer integrates AI with the real economy, promoting efficiency improvement and industrial transformation, and non - ferrous metals become the cornerstone of AI industrial implementation [30][31][32] 2.2 "AI + Non - Ferrous Metals" Leads the Industry to Higher - end and Intelligent Development - China's non - ferrous metal industry has developed rapidly but faces problems such as insufficient resource security and high - end supply. The "Work Plan" aims to promote high - quality development, emphasizing digital transformation and the "AI + non - ferrous metals" action [34][35][36] 3. Conclusion: Symbiosis of Non - Ferrous Metals and AI Industry, Key Reconstruction Period for the "Demand Map" - AI has opened up new demand space for non - ferrous metals, and the pricing logic and demand landscape may be reconstructed. Copper, aluminum, and tin have broad application prospects in the AI field due to their unique physical and chemical properties [40][41][42] - The next report will focus on evaluating the incremental demand for non - ferrous metals from the development of data centers (computing power centers) [43]