China's Military AI Revolution: How DeepSeek Powers Beijing's Next-Generation Warfare Technology

Unitree robot dogs climb a flight of stairs during a demonstration at a park in Hangzhou.

Beijing/Singapore:

China's state-owned defense giant Norinco unveiled a military vehicle in February capable of autonomously conducting combat-support operations at speeds of 50 kilometers per hour. This vehicle is powered by DeepSeek, the artificial intelligence model that has become the pride of China's technology sector.

The release of the Norinco P60 was celebrated by Communist Party officials in press statements as an early demonstration of how Beijing is leveraging DeepSeek and AI to gain ground in its arms race with the United States, at a time when leaders in both nations are urging their militaries to prepare for potential conflict.

A comprehensive review of hundreds of research papers, patents, and procurement records reveals a systematic effort by Beijing to harness artificial intelligence for military advantage.

While the specific workings of China's next-generation weapons systems remain state secrets, procurement records and patents offer insights into Beijing's progress toward capabilities such as autonomous target recognition and real-time battlefield decision support systems that mirror American efforts.

It remains unclear whether all patented technologies have been built, as patents don't necessarily indicate operational technology.

The People's Liberation Army (PLA) and affiliated organizations continue to utilize and seek Nvidia chips, including models subject to US export controls, according to papers, tenders, and patents reviewed.

It could not be determined if these chips were stockpiled before Washington imposed restrictions, as the documents do not specify when the hardware was exported. Patents filed as recently as June demonstrate their use by military-linked research institutes. In September 2022, the US Commerce Department prohibited exports of Nvidia's popular A100 and H100 chips to China.

Nvidia spokesperson John Rizzo stated that while the company cannot track individual resales of previously sold products, "recycling small quantities of old, second-hand products doesn't enable anything new or raise any national security concern. Using restricted products for military applications would be a nonstarter, without support, software, or maintenance."

The US Treasury and Commerce Departments did not respond to inquiries about these findings.

The Chinese military has also increased its use of contractors in 2025 that claim to exclusively use domestically-manufactured hardware such as Huawei AI chips, according to Sunny Cheung, a fellow at the Washington-based Jamestown Foundation defense policy think-tank, who analyzed several hundred tenders issued from the PLA Procurement Network over a six-month period this year.

This shift would align with Beijing's public campaign urging domestic firms to utilize Chinese-made technology.

A review of procurement notices and patents filed with China's patent office revealed demand for and use of Huawei chips by PLA affiliates, though not all tenders viewed by Jamestown could be independently verified. Jamestown is releasing a report this week that was provided early to journalists.

Huawei declined to comment regarding military deployment of its chips. The Chinese defense ministry, DeepSeek, and Norinco did not respond to requests for comment about their use of AI for military applications. The universities and defense firms associated with the patents and research papers also did not respond to similar inquiries.

DeepSeek's AI models were mentioned in a dozen tenders from PLA entities filed this year, while only one referenced Alibaba's Qwen, a major domestic competitor.

Alibaba did not respond to inquiries about military use of Qwen.

Procurement notices related to DeepSeek have accelerated throughout 2025, with new military applications regularly appearing on the PLA network, according to Jamestown.

DeepSeek's popularity within the PLA also reflects China's pursuit of "algorithmic sovereignty" - reducing dependence on Western technology while strengthening control over critical digital infrastructure.

The US Department of Defense declined to comment on the PLA's use of AI.

A State Department spokesperson responded that "DeepSeek has willingly provided, and will likely continue to provide, support to China's military and intelligence operations."

Washington will "pursue a bold, inclusive strategy to American AI technology with trusted foreign countries around the world, while keeping the technology out of the hands of our adversaries," the spokesperson added.

China is exploring AI-powered robot dogs that scout in packs and drone swarms that autonomously track targets, as well as visually-immersive command centers and advanced war game simulations, according to the documents reviewed.

In November 2024, the PLA issued a science fiction-like tender for AI-powered robot dogs designed to scout together for threats and clear explosive hazards.

It remains unclear if this tender was fulfilled. China has previously deployed armed robot dogs from AI robotics manufacturer Unitree in military drills, according to images released in state media.

Unitree did not respond to queries about its PLA work.

The review of patents, tenders, and research papers published over the past two years demonstrates how the PLA and affiliated entities are looking to AI to improve military planning, including developing technology to quickly analyze images captured by satellites and drones.

Researchers at Landship Information Technology, a Chinese company that integrates AI systems into military vehicles including those made by Norinco, stated in a February white paper promoting their services that their technology built on Huawei chips can rapidly identify targets from satellite imagery while coordinating with radars and aircraft to execute operations.

The time required for military planners to progress from finding and identifying targets to executing operations has also been shortened by AI, according to Xi'an Technological University.

Researchers at this institute reported in May that their DeepSeek-powered system was able to assess 10,000 battlefield scenarios - each with different variables, terrain, and force deployments - in just 48 seconds.

Such a task would have required 48 hours for a conventional team of military planners to complete, they claimed.

These researchers' claims could not be independently verified.

Chinese military entities are investing in increasingly autonomous battlefield technology, the documents suggest.

Two dozen tenders and patents reveal military efforts to integrate AI into drones for target recognition and tracking, as well as enabling them to work together in formations with minimal human intervention.

Beihang University, known for its military aviation research, is utilizing DeepSeek to improve drone swarm decision-making when targeting "low, slow, small" threats - military terminology for drones and light aircraft - according to a patent filing this year.

Chinese defense leaders have publicly committed to maintaining human control over weapons systems amid growing concerns that conflict between Beijing and Washington could lead to uncontrolled deployment of AI-powered munitions.

The US military, also investing heavily in AI, aims to deploy thousands of autonomous drones by the end of 2025, which officials describe as an effort to counter China's numerical advantage in unmanned aerial vehicles.

Chinese defense contractors like Shanxi 100 Trust Information Technology have promoted in marketing materials their reliance on domestically produced components such as Huawei's Ascend chips, which enable AI model operation.

The firm did not respond to questions about its relationship with Huawei and the PLA.

Despite the transition to domestic processors, Nvidia hardware continues to be frequently cited in research by military-affiliated academics, according to a review of patent filings from the past two years.

Thirty-five applications were identified referencing use of Nvidia's A100 chips by academics at the PLA's National University of Defense Technology (NUDT) and at the "Seven Sons" - a group of Chinese universities under US sanctions with a history of conducting defense-related research for Beijing.

These entities filed 15 patents related to AI applications citing Huawei Ascend hardware, designed as a substitute for Nvidia chips, during the same period.

As recently as June, the PLA Rocket Force University of Engineering filed a patent for a remote-sensing target detection system, which reportedly used A100 chips for model training.

Senior Col. Zhu Qichao, who leads an NUDT research center, stated last year that US restrictions have impacted their AI research "to some degree," though they remain determined to narrow the technological gap.

Nvidia's Rizzo downplayed PLA demand for Nvidia hardware, saying that China "has more than enough domestic chips for all of its military applications."