AI Patent War Map -- What Are Korea, US, and China Investing In? The era of speaking through AI patents: as the speed of technology development accelerates, companies are moving the competitive weight center from simple model performance to intellectual property acquisition and technology preemption. The AI patent landscape: US companies Google, Microsoft, and NVIDIA dominate in foundational AI patents covering machine learning architectures, training methods, and inference optimization; the breadth and quality of their portfolios create licensing leverage that extends their competitive advantage beyond just deployment. China government-directed AI patent ecosystem: unlike the US market-driven approach, China government actively coordinates AI patent strategy across state enterprises and research institutions; the result is a highly organized ecosystem where key patents are cross-licensed between Chinese companies creating a domestic moat; SMIC, Huawei, Baidu, and dozens of state-linked AI labs coordinate to ensure no single chokepoint exists for China domestic AI industry. Korea AI patent position: Samsung and SK Hynix dominate semiconductor-related AI patents (memory, processing architecture); Korea lacks the software AI patent depth of US companies; government investment in AI patent development through ETRI and other research institutes is increasing but patent quality and breadth lag the US and China. The patent war implication: AI leadership will increasingly be determined not by who has the best current model but by who has the broadest patent portfolio covering the techniques that all future models will use.