China’s Private AI Firms Capture 74% of Military Contracts, Outflanking U.S. Export Controls
Agile tech companies displaced state-owned defense giants in PLA procurement, creating intelligence blind spots that undermine sanctions targeting.
Private Chinese technology firms captured nearly three-quarters of AI-related military procurement contracts between January 2023 and December 2024, according to Georgetown University’s Center for Security and Emerging Technology, marking a structural shift that accelerates weapons development while exposing critical gaps in U.S. export controls. The 74% share went to nontraditional vendors with no reported state ownership ties—not the established defense contractors Western intelligence agencies typically monitor.
The data reveals an institutional reshuffling inside China’s military-industrial complex. Of 338 entities winning AI contracts from the People’s Liberation Army, nontraditional vendors secured 764 deals—the largest slice of any category, outpacing state-owned enterprises and research institutions. Two-thirds of these firms were founded after 2010, reflecting Beijing’s deliberate pivot toward agile commercial players capable of rapid prototyping and deployment cycles that legacy SOEs cannot match.
The Military-Civil Fusion Advantage
China’s military-civil fusion doctrine—centrally coordinated by Xi Jinping since 2017—systematizes innovation across civilian and defense sectors through procurement platforms and joint research mandates. Private firms benefit from structural incentives: their technical capacity and speed-to-deployment align with MCF’s operational tempo, while entrenched SOEs remain bureaucratically slower. Foreign Policy Research Institute analysis from February 2026 contrasts this centralized pipeline with the U.S. approach, where fragmented acquisition processes create delays between procurement and battlefield integration.
The shift is most visible in DeepSeek-specific contracts, where private companies won the majority of PLA tenders, per The Diplomat. These firms—obscure IT providers with no presence on U.S. restricted entity lists—are integrating large language models into battlefield scenario planning, autonomous vehicle navigation, and geospatial intelligence systems. The vast majority remain exempt from U.S. sanctions because they lack documented state ownership or publicly acknowledged military work.
“The firms winning the PLA’s major AI integration contracts are obscure private IT companies—firms that do not appear on any existing restricted entity list.”
— The Diplomat analysis
Export Control Blind Spots
The Pentagon updated its Chinese Military Companies List on February 14, 2026, adding Alibaba, Baidu, and Tencent to bring the total to 72 entities, according to Caixin Global. The designations signal belated recognition of Big Tech’s military integration, but they leave the broader NTV ecosystem untouched. Georgetown CSET found that most nontraditional vendors winning PLA contracts operate outside existing sanctions frameworks—a gap that renders entity-based Export Controls structurally inadequate.
U.S. chip export restrictions, while constraining access to cutting-edge semiconductors, have not prevented Chinese firms from advancing Military AI applications. DeepSeek founder Liang Wenfeng stated in 2024 that “money has never been the problem for us; bans on shipments of advanced chips are the problem,” yet the firm’s technology now powers PLA procurement, as documented in AI Frontiers reporting from July 2025. Chinese AI companies compensate through access to domestic surveillance data and operational experience unavailable to Western competitors—advantages detailed in Center for Strategic and International Studies analysis.
Baidu’s AI Cloud revenue grew 45% year-over-year in Q1 2025, reflecting commercial momentum that feeds dual-use capabilities. The company now appears on the Pentagon’s military companies list, but its cloud infrastructure serves hundreds of smaller firms—many of which remain invisible to U.S. intelligence collection.
Intelligence Collection Gaps
The scale of NTV involvement surprised Georgetown researchers. “Just in this data set, the sheer ambition of what they’re trying to do is surprising,” said Cole McFaul, a CSET analyst quoted in Wall Street Journal coverage. “That there’s such a wide range of these technologies speaks to our limited ability to really hammer or constrain China’s military modernization.”
The intelligence gap extends beyond contract monitoring. Private firms like iFlyTek and SenseTime have gained military experience through domestic surveillance applications before pivoting to PLA work—a trajectory that traditional defense contractor tracking misses. Defense One reported in September 2025 that these firms deploy dual-use technologies including drone navigation systems and battlefield geospatial tools, yet most operate without sanctions exposure.
- Entity-based sanctions cannot track diffuse NTV ecosystem without documented ownership ties
- Commercial AI firms transfer surveillance-derived capabilities directly to military applications
- U.S. acquisition speed lags China’s procurement-to-deployment pipeline by institutional design
- Chip export controls constrain but do not prevent military AI development at scale
Policy Reckoning
The Georgetown data forces a reckoning with export control design. Sam Bresnick, a CSET researcher, warned against reflexive escalation: “There’s a lot of momentum toward the sledgehammer approach because it’s easy.” But blanket restrictions on Chinese AI firms risk collateral damage to U.S. commercial interests while failing to address the structural advantage MCF creates—thousands of agile vendors with plausible civilian cover stories and no visible state ownership.
Council on Foreign Relations analysis from December 2025 noted that private firms like Alibaba and Baidu are developing proprietary chips to circumvent U.S. restrictions, further decoupling the military AI supply chain from Western dependencies. The strategy mirrors MCF’s core logic: distribute innovation across a broad private sector base that resists targeted interdiction.
What to Watch
Georgetown CSET’s dataset covers January 2023 through December 2024; updated Q1 2026 contract data will indicate whether the NTV trend accelerated following the Pentagon’s February 2026 designations. Monitor whether U.S. policymakers shift from entity-based sanctions to technology-specific controls that target capabilities rather than corporate structures. The effectiveness of chip export restrictions will hinge on whether Chinese firms achieve performance parity through domestic semiconductor alternatives—a timeline Huawei’s progress suggests may arrive sooner than Western assessments assume. Finally, track whether U.S. defense acquisition reforms narrow the speed gap that currently favors China’s centralized procurement model, as institutional agility may matter more than individual technology advantages in the AI arms race.