Deep Turmoil in the PLA's Military-Industrial System: The Dual Dilemma of Intelligent Transformation and Internal Purges

March 2026, China's military-industrial system is undergoing an unprecedented internal upheaval. As the People's Liberation Army (PLA) accelerates its push for intelligent military modernization, a top-down anti-corruption storm is sweeping through the entire national defense research and industrial system. From the technological breakthroughs of the J-35 stealth fighter to the rapid iteration of artificial intelligence weapon systems, and the collective downfall of over a dozen academicians and executives, this series of seemingly contradictory events reflects the deep-seated structural contradictions in the PLA's modernization process.

Since the display of the new generation of unmanned combat systems during the Victory Day parade in September 2025, the PLA's intelligent construction roadmap has fully transitioned from the planning stage to the implementation stage. According to publicly available PLA equipment procurement documents, over the past three years, the PLA has been massively procuring artificial intelligence systems capable of driving unmanned combat vehicles, detecting and responding to cyberattacks, tracking maritime targets, and identifying and striking land, sea, air, and space targets. These systems not only cover tactical-level automated weapons but also include strategic decision support tools capable of processing, analyzing, and enhancing vast amounts of data, as well as deepfake technologies used for the dissemination of false information.

At the tactical level, the PLA is developing drone swarm systems capable of identifying, tracking, and coordinating attacks on enemy targets. These systems bear a striking resemblance to the U.S. Department of Defense's Replicator Initiative, both aimed at the large-scale procurement of low-cost, expendable drones for future conflicts. Notably, the PLA's procurement documents have revealed requirements remarkably similar to the U.S. Department of Defense's Combined Joint All-Domain Command and Control initiative, which utilizes artificial intelligence and computer networks to connect command centers, troops, and weapon systems across different military branches.

At the strategic level, the PLA's investment in artificial intelligence-assisted decision-making systems is particularly noteworthy. The Chinese political and military leadership has limited trust in the PLA's command chain and is concerned that the command system may struggle to respond in rapidly evolving conflicts. Artificial intelligence decision support systems are seen as key tools to help PLA leadership understand and anticipate adversary movements, potentially compensating for the PLA's disadvantages in real combat experience. Additionally, the PLA is vigorously developing cognitive warfare technologies, viewing deepfake-generated images, videos, and audio as powerful tools to influence public opinion and manipulate adversary perceptions and decisions.

The J-35 stealth fighter project is a typical case of the PLA's strategy to reduce costs and increase efficiency. This medium-sized stealth fighter, developed by Shenyang Aircraft Corporation, has a development history that serves as a textbook example of technological iteration and cost control. From the poorly performing FC-31 prototype that was first publicly unveiled in 2014 to the J-35A/B models now serving on the Fujian aircraft carrier, this aircraft has undergone multiple key improvements: gradually replacing the Russian RD-93 engine with the domestically produced WS-19 engine, reducing the number of structural components by 50% through the extensive use of 3D printing technology, and introducing civilian supply chains to lower costs.

According to a report by China Daily at the end of 2024, Air Force expert Fu Qianshao stated that the cost of the J-35A will be significantly lower than that of the J-20, allowing the PLA Air Force to procure on a larger scale. A subsequent assessment by Xinhua positioned the J-20 as a heavy stealth fighter primarily used for air superiority, while the J-35A is classified as a medium multi-role stealth fighter capable of performing both air superiority missions and striking ground and maritime targets. This positioning strategy clearly replicates the U.S. Air Force's high-low mix model of the F-22 and F-35, but in terms of cost control, China has achieved goals that the U.S. promised in the 1990s but never fulfilled.

However, just as the PLA's intelligent construction and equipment modernization are accelerating, an unprecedented anti-corruption storm is shaking the very foundations of the military-industrial system. On March 17, 2026, Caixin reported that Yang Wei, the chief designer of the J-20 and former deputy general manager of the Aviation Industry Corporation of China, had been removed from the list of academicians of the Chinese Academy of Sciences. The day before, Sina Finance reported that three academicians, including Wu Manqing, vice president of the Chinese Academy of Engineering, Zhao Xiangeng, and Wei Yiyin, former deputy general manager of China Aerospace Science and Industry Corporation, had also disappeared from the official list.

The breadth of this purge is shocking. According to Ming Pao statistics, since the 20th National Congress of the Communist Party of China in 2022, at least 10 academicians have been removed, including nuclear reactor design expert Luo Qi, missile control theory expert Xiao Longxu, naval propulsion system expert Jin Donghan, and former chairman of the aviation engine group Cao Jianguo. These removed academicians are distributed across China's top ten state-owned defense groups, including China National Nuclear Corporation, Aviation Industry Corporation of China, China Aerospace Science and Technology Corporation, China Shipbuilding Industry Corporation, and China Electronics Technology Group, covering almost all defense industrial fields from nuclear weapons and missiles to aviation engines, naval vessels, and electronic warfare systems.

Even more concerning is the systemic corruption problem exposed by this purge. Online comments point out that corruption exists at multiple levels: at the research level, some academicians are accused of falsifying research data to claim technological breakthroughs; at the procurement level, executives are suspected of accepting bribes during equipment approval and procurement processes; at the operational level, grassroots officers are reportedly attempting to cover up equipment defects while diverting resources. Some netizens bluntly state that the Communist Party system lacks competition and oversight, and wherever the Party exists, deception, corruption, and unfinished projects will follow.

The impact of this purge on the PLA's intelligent construction is incalculable. When the chief expert of radar systems falls, can the so-called Eagle Eye system, which claims to detect stealth fighters, still function normally? When the leading figure in missile guidance technology is embroiled in a corruption investigation, will the so-called Dongfeng Express's precision strike capability be misled by falsified chips and algorithms? When nuclear weapons experts are purged, the issue is no longer just about technical inflation but directly touches the credibility of the country's nuclear deterrent.

From a technological competition perspective, the arms race in artificial intelligence between China and the U.S. is entering a phase of rapid iteration. Many of the artificial intelligence application concepts shown in the PLA's procurement documents are highly similar to major projects of the U.S. Department of Defense. Including parallel experiments in target algorithms and cyber operations, this seems to lock Washington and Beijing into a rapid technological transformation cycle. The outcome of the competition will depend on which military can develop and scale new capabilities and operational improvements faster. Given the speed of evolution in many defense technologies, technological advantages may be hard-won but ultimately fleeting.

However, the balance of how the two militaries approach artificial intelligence-assisted decision-making may become a decisive factor in the competition. The U.S. military requires that artificial intelligence systems involve appropriate levels of human judgment by trained and experienced personnel, while the PLA may lean towards using artificial intelligence decision support systems as a substitute for its weak and inexperienced officer corps. But over-reliance on computer-generated analyses may lead to misinterpretations of military and diplomatic signals and unreasonable military decisions.

Additionally, some of the artificial intelligence decision-making technologies mentioned in the PLA's procurement documents rely on open-source data. This may incentivize various parties to manipulate the information environment, such as flooding social media with false signals or disrupting commercial satellite imagery providers to enhance the effectiveness of information warfare. These attacks aim to confuse, deceive, and undermine the adversary's artificial intelligence tools, potentially leading to unintended escalations.

Despite the chaos brought about by the internal purge, the PLA's path toward intelligence will not stagnate. As demonstrated by the war in Ukraine, developing technologies like autonomous drones is one thing, but effectively deploying them in contested battlefields is entirely another. The ability to rapidly train drone operators, formulate attack plans, and prevent drone misfires is as important as the technological capabilities of the drones themselves. Speed of adoption will become a key dimension of competition. Furthermore, training artificial intelligence systems for military use requires vast amounts of data that are not easily available on the internet, such as classified images of military platforms or the electromagnetic characteristics of various radars and weapons. The PLA's limited real combat experience lacks many of these datasets.

Even if certain artificial intelligence systems fail, the PLA's rapid prototyping and experimentation will accelerate its learning and improvement. China is preparing for the rapid and effective adoption and deployment of operational artificial intelligence, thereby keeping the gap between the Chinese and U.S. militaries narrow. This reality should prompt Washington to develop more flexible military experimentation, procurement, and adoption processes. The U.S. Congress and the Trump administration have begun to take action for reform: the 2026 National Defense Authorization Act includes a series of provisions aimed at reducing unnecessary delays, expanding commercial procurement pathways, and granting defense contractors greater flexibility when new technologies emerge.

In the long run, the core challenge facing the PLA is how to find a balance between anti-corruption and modernization. Excessive purging may destroy the morale and innovative capacity of the entire military-industrial research and development system, but allowing corruption to fester will fundamentally weaken the military's combat effectiveness. The third phase of military modernization that Beijing has promised is fully underway, with the PLA committed to developing artificial intelligence to gain a military advantage over the U.S. However, many of the applications proposed by the Chinese military, particularly its complex artificial intelligence decision support systems and cognitive warfare tools, exacerbate the risks of misjudgment and instability.

As China continues to modernize its military through the adoption of emerging technologies, the U.S. should also commit to rapid artificial intelligence iteration, responsible experimentation, and prudent commercial partnerships to maintain its edge. In this technological race, speed, transparency, and system reliability will be decisive factors. For the PLA, how to maintain the vitality of technological innovation while rooting out corruption will be the key test of whether its intelligent transformation can succeed.