西安财经大学学报

在全球人工智能技术竞争加剧与我国面临技术封锁的双重背景下，探索中国AI大模型自主创新形成机制具有重要理论与现实意义。以DeepSeek为典型案例，运用纵向单案例研究方法，基于战略管理和组织理论视角，将其发展过程划分为初创期、发展期和转型期三个阶段，构建“动因—过程—结果”分析框架，揭示其从量化投资跨界到通用人工智能的创新逻辑。结果发现，DeepSeek自主创新的主要特征体现为技术路径重构和商业模式变革，在外部情境、创始人特质和组织能力共同作用下，各阶段的技术路线和商业模式选择不断演变，最终形成了由探索式创新、组合创新向重构式创新的跨越。在讨论了长期主义的技术深耕、核心人物的引领作用等五大自主创新的关键要素后，提出三点政策启示，包括构建“开放、协同和包容”的创新生态、培育“市场驱动、组织敏捷和长期主义”的科技民营企业和打造“青年主导、潜力优先和跨界贯通”的人才培养体系。

Against the backdrop of intensifying global artificial intelligence(AI) competition and technological blockade faced by China, exploring the formation mechanism of China's independent innovation in large AI models holds significant theoretical and practical implications.This study takes DeepSeek as a typical case and employs a longitudinal single-case study method.From the perspectives of strategic management and organizational theory, it divides DeepSeek's development into three stages: startup, growth, and transformation.By constructing a “motivation-process-outcome” analytical framework, the study reveals the innovation logic behind its transition from quantitative investment to artificial general intelligence.The findings indicate that DeepSeek's independent innovation is characterized by technological path reconstruction and business model transformation.Under the combined influence of external conditions, founder traits, and organizational capabilities, its choices of technological approaches and business models evolved across different stages, ultimately leading to a transition from exploratory and combinatorial innovation to architectural innovation.After discussing the five key elements of independent innovation, such as long-term technological exploration and the leading role of key figures, the study proposes three policy implications: building an “open, collaborative, and inclusive” innovation ecosystem; fostering “market-driven, organizationally agile, and long-term-oriented” private technology enterprises; and developing a “youth-led, potential-first, and cross-disciplinary” talent cultivation system.