China’s integration of large language models into military systems marks a transformative shift in warfare, challenging US dominance through rapid innovation, seamless systems fusion, and precision mass deployment across the battlefield.
Image Source: Getty Images
Referring to China’s 2017 Artificial Intelligence (AI) Plan, former Google CEO Eric Schmidt once remarked, “By 2020, they will have caught up. By 2025, they will be better than us. And by 2030, they will dominate the industries of AI.” China’s recent launch of DeepSeek brought Eric’s warning to life, with some claiming the launch as China’s ‘Sputnik moment’.
A Chinese lead in Large Language Models (LLMs) might seem just an economic edge. Nevertheless, a new generation of AI-enabled military systems represents a cognitive leap in future warfare, enhancing China’s systems integration, streamlining kill chains, and potentially outpacing US military decision cycles. Further, China’s advanced Unmanned Aerial Vehicle (UAVs) or drone production capacity aids this integration, providing capabilities like autonomous swarms and enabling affordable and precision mass across the battlefield.
As China begins to weave advanced LLMs into its military architecture, America must accelerate systems integration for synced decision loops across services to maintain its edge in a high-tech, AI-driven battlespace.
While the US races to field its autonomous systems through the Replicator Initiative, experts warn that disjointed software integration is slowing its progress. As China begins to weave advanced LLMs into its military architecture, America must accelerate systems integration for synced decision loops across services to maintain its edge in a high-tech, AI-driven battlespace.
Developed with a fraction of the cost and only one-tenth of the computing power compared to LLMs developed in the West, including ChatGPT4 and Llama 3, DeepSeek holds significant potential for military applications, ranging from UAVs to Command and Control systems (C2). Recent reports suggest that China has deployed DeepSeek’s models in hospitals and soldier training programmes, suggesting a controlled environment for experimentation before expanding into high-stakes combat environments. A recent defence expo in Abu Dhabi also witnessed DeepSeek’s integration into the Xingji-P60, an autonomous military vehicle incorporating self-driving software for dual-use applications. These developments suggest China’s emphasis on rapid experimentation and deployment of LLMs for military use.
A more pressing concern is China’s ability to rapidly deploy LLMs across services for synchronised decision-making in a ‘network-centric’ battlefield. By fusing AI across services, these LLMs could close the gap between data, decision, and action in a fraction of a second for faster sensor-to-shooter loops. Utilising real-time intelligence to guide strike operations, these LLMs could integrate reconnaissance assets with precision-strike capabilities to rapidly close kill chains—a military organisation experts call reconnaissance strike complex (RSC). The PRC’s enhanced RSC capabilities risk redefining battlefield dominance before the US can even match the pace.
China is pursuing the development of what some US analysts call a ‘multi-domain kill-web’ designed to coordinate across aircraft, sensors and missiles.
Chinese reports suggest that DeepSeek can integrate drones, satellites, and radars, swiftly pinpointing key targets to aid military decisions. Recent Pentagon reports also support these claims, indicating that China is pursuing the development of what some US analysts call a ‘multi-domain kill-web’ designed to coordinate across aircraft, sensors and missiles. Emphasising the level of AI integration that China is envisioning, the report states that the PLA expects a range of ‘algorithmic’ and ‘network-centric’ warfare capabilities operating at different levels of autonomy by 2030. Moreover, a centralised command structure to enforce uniform data sharing, along with a civil-military fusion connecting commercial technologies to military pipelines, further accelerates rapid deployment for the PLA.
Currently, the US military still lacks unified AI integration across branches, compromising deployment efforts. Though initiatives like Joint All-Domain Command and Control (JADC2) show some promise, significant challenges persist, including data sharing across services, aligning organisational structures for joint operations, and cultural barriers within the military. Without interoperable software, autonomous platforms cannot communicate or share data in real-time with other platforms and systems, affecting decision-making loops.
Additionally, China’s large drone market intensifies these challenges by enabling precision mass at scale, flooding the battlespace with affordable, AI-enabled platforms that coordinate with each other for mission tasks. An integrated systems approach coupled with affordable precision mass threatens US dominance in the forthcoming age of high-tech warfare. Although China may not have fully caught up with or surpassed US technological edge on many fronts, advanced LLM integration through a ‘system-of-systems’ approach could tilt the balance in its favour if Washington fails to respond with urgency.
The US military still lacks unified AI integration across branches, compromising deployment efforts. Though initiatives like Joint All-Domain Command and Control (JADC2) show some promise, significant challenges persist.
Currently, small drones equipped with AI have become ubiquitous tools in the war in Ukraine. In particular, American companies are supplying Ukraine with technologies like Skynode-S boards—computers roughly the size of a credit card—optimised for LLM applications. These systems enable UAVs to lock on to a target and manoeuvre autonomously, while performing various other tasks. A range of challenges, however, plague these technologies, including compressing sufficient computing power onto a small UAV and training LLM systems to operate reliably and efficiently. This is precisely the sort of technology that China could unlock with advanced LLMs like DeepSeek. These advancements enhance data processing, accelerate decision-making, and enable seamless, natural-language coordination among UAVs and other autonomous platforms.
While recent Department of Defense (DoD) efforts like JADC2, the new Software Acquisition policy, incentivising Other Transactions Authority (OTAs), and Modular Open Systems are meaningful steps, these efforts remain slow due to bureaucratic hurdles and organisational silos. The DoD must prioritise ‘Standardized Reference Architecture’ and integrate software systems across units, services, and departments, accelerate DevSecOps pipelines to integrate workflows throughout the department, and field interoperable LLMs that compress decision cycles in real-time.
Although China may not have fully caught up with or surpassed US technological edge on many fronts, advanced LLM integration through a ‘system-of-systems’ approach could tilt the balance in its favour if Washington fails to respond with urgency.
The rapid development of the PLA’s LLM-powered battlefield capabilities presents a significant challenge to US military dominance. Though the US recognises this threat, addressing persistent issues of disjointed software integration, bureaucratic hurdles, and organisational silos with agility is key to maintaining an edge in the era of ‘software-defined but hardware-enabled’ warfare. China’s AI integration and production scale could tilt future battlefields in its favour, and American deterrence now hinges on swiftly evolving and adapting interoperable software for a network-centric battlefield.
Rohith Narayan Stambamkadi is an Associate at Starburst Accelerator, USA—an innovation consulting firm that fast-tracks the development and deployment of cutting-edge ventures in the defence and aerospace sectors.
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.
Rohith Narayan Stambamkadi is an Associate at Starburst Accelerator, USA—an innovation consulting firm that fast-tracks the development and deployment of cutting-edge ventures in the defence ...
Read More +
PREV
