A new investigation reveals that domestic artificial intelligence models from the Chinese company Deepseek are being increasingly integrated into China’s military research, signaling a significant step in the nation's push for technological self-reliance in defense.[1] An extensive review of patents, academic papers, and government procurement documents shows that Deepseek's powerful and often open-source AI is being harnessed for a range of applications, from autonomous drone swarms and combat vehicles to advanced battlefield simulations.[2][3][4] This growing reliance on indigenous AI highlights Beijing's strategy of "military-civil fusion," which aims to leverage private sector technological advancements for military modernization, and accelerates the global AI arms race.[5] The trend places a spotlight on the challenges of controlling dual-use technology and the effectiveness of Western export controls designed to slow China's military progress.[6][7]

The applications of Deepseek's AI within the People's Liberation Army's research ecosystem are both diverse and sophisticated, according to public records. Military-affiliated universities and defense contractors are exploring the technology to enhance autonomous operations and strategic planning.[8] For instance, state-owned defense giant Norinco unveiled an autonomous combat-support vehicle, the P60, which is powered by Deepseek's AI and capable of operating at high speeds.[9][4][10] Research from institutions like Beihang University, known for its aviation research, includes patents for using Deepseek to improve the decision-making of drone swarms when targeting threats.[1][2][11] Beyond autonomous hardware, the technology is being used to revolutionize military logistics and command. Researchers at Xi'an Technological University claim a Deepseek-powered system can evaluate 10,000 potential battlefield scenarios in just 48 seconds, a task that would traditionally take human commanders 48 hours.[12][13] The technology is also being applied to analyze satellite imagery for rapid target identification and to power AI-driven robot dogs for reconnaissance missions.[1][3][10] This widespread adoption is reflected in procurement data, which shows a significant increase in military tenders referencing Deepseek's models in 2025, far surpassing mentions of rival domestic models like Alibaba's Qwen.[2][11][10]

This pronounced shift towards Deepseek is a clear manifestation of China's broader strategic goal to achieve "algorithmic sovereignty."[1][11] By prioritizing homegrown AI, Beijing aims to reduce its dependence on foreign technology, thereby insulating its military-industrial complex from Western sanctions and export controls.[2][11][3] The nation's "military-civil fusion" policy actively encourages the flow of technology from the commercial sector to the defense sector, creating a dynamic where a private startup like Deepseek becomes a key enabler of military advancement.[5][14] Founded in 2023 by Liang Wenfeng, the co-founder of the hedge fund High-Flyer, Deepseek quickly gained prominence for producing powerful AI models that rivaled Western counterparts at a fraction of the cost.[15][16][17] While the company focuses on open-source models, its technology is also being adopted for non-combat military roles, such as improving treatment plans in PLA hospitals and assisting with personnel management.[5] This dual-use nature of AI complicates global efforts to regulate its military application, as models developed in the open can be freely adapted for defense purposes.[18][19]

The growing integration of advanced AI into China's military apparatus is unfolding against a backdrop of an intense technological rivalry with the United States. Washington has implemented stringent export controls since 2022, specifically targeting high-performance AI chips, such as Nvidia's A100 and H100 models, to curb their use in advancing China's military capabilities.[20][21][22] Despite these restrictions, Chinese military research papers and patents continue to reference the use of these Nvidia chips, though it is unclear if they were procured before the bans took effect.[4][10] In response to the controls, there is a clear and growing pivot toward domestic hardware, with procurement notices increasingly specifying Chinese-made alternatives like Huawei's Ascend AI chips.[11][4][10] This situation is further complicated by accusations from the U.S. government. A senior State Department official stated that Deepseek has "willingly provided" support to China's military and intelligence operations and has actively sought to evade U.S. export controls, including attempts to acquire restricted chips through shell companies in Southeast Asia.[6][7][23]

The rise of Deepseek as a key partner in China’s military research carries profound implications for the global technology landscape and international security. It demonstrates how quickly cutting-edge, commercially-developed AI can be repurposed for military ends, blurring the lines between civilian innovation and defense applications.[2][4] The proliferation of powerful, open-source AI models presents a formidable challenge for policymakers attempting to prevent an unchecked AI arms race, as the very accessibility that drives innovation also makes the technology difficult to contain.[24][19] As both the U.S. and China push forward with plans to deploy thousands of AI-enabled autonomous systems, the focus on domestic champions like Deepseek underscores a world rapidly moving toward distinct and competitive technological ecosystems.[11][10] This trend not only intensifies the strategic competition between the two powers but also signals a new, more sophisticated, and potentially destabilizing era of AI-driven warfare.[2]