Artificial intelligence (AI) has emerged as a foundational driver of the ongoing technological revolution and industrial transformation. Drawing on case studies of leading firms in the automotive sector, this study develops a theoretical framework grounded in a tripartite empowerment logic—behavioral, modular, and systemic—to explicate the mechanisms through which AI enables industrial upgrading and transformation. The findings reveal that AI, by mobilizing core infrastructural capacities in data, computing power, algorithms, models, and connectivity, facilitates intelligent behaviors of perception, analysis, and decision-making. These capabilities restructure value chains through modular decomposition, integration, and optimization; simultaneously, they catalyze the expansion of collaborative networks, foster multi-actor value co-creation, and enhance the adaptability and resilience of innovation ecosystems—culminating in multi-level innovation outcomes. Furthermore, the study identifies three distinct AI empowerment pathways: (1) an all-encompassing transformation via five-dimensional synergy, (2) a rapid, cross-domain transformation focusing on strategic leverage points, and (3) a foundational transformation driven by externally embedded AI capabilities. This research contributes to the theoretical enrichment of AI-enabled industrial transformation from an innovation ecosystem perspective and offers actionable implications for policy and practice.