"China's Military AI Wish List" is a February 2026 report by the Center for Security and Emerging Technology (CSET) at Georgetown, authored by Emelia Probasco, Sam Bresnick, and Cole McFaul. It analyzes thousands of Chinese-language open-source requests for proposals (RFPs) published by the People's Liberation Army to characterize the PLA's stated demand for AI-enabled military technology, its procurement posture, and the role of military-civil fusion. The report builds on a companion CSET study, "Pulling Back the Curtain on China's Military-Civil Fusion" (McFaul, Bresnick, Chou, September 2025).
The full PDF returned a 403 error at fetch time; the content summarized here was compiled from the CSET publication page, press materials, and the companion September 2025 report.
Methodology
The primary dataset is thousands of Chinese-language open-source RFPs published by the People's Liberation Army between January 1, 2023 and December 31, 2024. The companion September 2025 report drew on a dataset of 2,857 AI-related defense contract award notices covering the same period, awarded to 1,560 unique entities, with deeper analysis of the 338 entities that won two or more contracts.
Scope
The report focuses on C5ISRT: command, control, communications, computers, cyber, intelligence, surveillance, reconnaissance, and targeting. The authors identify this capability family as having the most near-term AI leverage on warfighting.
Findings
Stated PLA demand
The RFPs indicate PLA demand for AI decision-support systems that ingest open-source data for strategic and tactical use; maritime- and space-domain technologies explicitly oriented to countering perceived US advantages; biometric surveillance including facial recognition, gait recognition, and digital forensics; and deepfake generation and detection, covering both offensive information-operations tooling and defensive integrity tooling.
Procurement posture
Most RFPs carry modest budgets with three- to six-month timelines, which the report reads as evidence of a rapid-prototyping and commercial-vendor experimentation model rather than multi-year programs of record. The authors contrast this with the US Department of Defense's traditional acquisition timeline.
Military-civil fusion patterns
Drawing on the companion September 2025 report's analysis of the 338 entities that won multiple contracts, the authors describe heavy overlap between defense contractors and commercial AI companies. They argue that military-civil fusion (MCF) mechanisms channel commercial AI capabilities directly into military applications without a separate "defense-grade" development track. According to the report, this structure makes export controls less effective: controlling civilian chips constrains military AI indirectly, but the civilian-to-military transfer pathway is short.
Relevance to export-control debates
The authors situate the MCF procurement structure within the debate between unconditional export controls (Dario Amodei — On DeepSeek and Export Controls) and conditional or targeted controls (Geopolitics in the Age of Artificial Intelligence). If PLA acquisition occurs primarily through civilian commercial channels, chip controls affect the whole stack, both civilian and military, which the report frames as the intended effect of such controls but also a source of diplomatic costs.
Connections
The report provides an empirical basis for PLA AI procurement priorities, complementing US-side documentation such as DoD Directive 3000.09 (AI and National Security). The PLA's C5ISRT acquisitions represent upstream demand for autonomous systems (Autonomous Weapons), and the MCF structure illustrates why civilian chip controls carry military implications (Compute Governance). It is a primary publication of CSET (Center for Security and Emerging Technology (CSET)).
Confidence
Confidence is rated medium. The primary PDF was not retrieved, and the content here was reconstructed from CSET public materials and the companion report. The methodology of open-source RFP analysis is transparent, and the claims about MCF overlap are supported by the companion dataset's methodology.
Relationships
- supports: AI and National Security — empirical basis for PLA AI demand priorities
- supports: Compute Governance — MCF structure connects civilian chip policy to military AI capability
- related: CSET + DOJ + CSIS — US-China AI National-Security Axis (composite source summary) — CSET/CSIS national security analysis (prior ingest)
- related: Autonomous Weapons — PLA C5ISRT procurement includes autonomous targeting
- related: China and the US Are Running Different AI Races — adds military procurement dimension to the US-China divergence picture