| Field | Value |
|---|---|
| Type | Bipartisan nonprofit national-security research (think tank) |
| HQ | Washington, D.C. |
| Focus | National security and defense policy; military AI, autonomous systems, US-China technology competition, compute and export controls |
| Relevant programs | AI Security and Stability project; Artificial Intelligence and Global Security Initiative; Technology and National Security program |
The Center for a New American Security (CNAS) is a bipartisan Washington, D.C. think tank focused on national-security and defense policy. Its technology research spans military applications of AI, autonomous weapons, the US-China AI competition, and the compute and export-control levers tied to them.
Programs and activities
CNAS conducts its AI work through several programs. The AI Security and Stability project states its aim as "to inform government decision-making on the most critical AI policy issues that will affect the future of AI" (Source: cnas.org). The Artificial Intelligence and Global Security Initiative examines how the AI revolution could shift global power (Source: cnas.org). The broader Technology and National Security program conducts policy research aimed at securing and extending US AI leadership, including identifying bottlenecks to continued progress (Source: cnas.org).
CNAS analysis appears in the AI-and-national-security debate on export controls, compute governance, and defense-adoption questions, including those formalized in NSPM-11 — Artificial Intelligence in the National Security Enterprise (Trump, June 5, 2026). It is one of several Washington institutions, alongside Center for Strategic and International Studies (CSIS), cited on these topics.
Publications on AGI and geopolitics
CNAS published Superpowers and AGI: How U.S.-China Competition Could Be Shaped by the Emergence of Artificial General Intelligence on August 27, 2026, by senior fellow and Indo-Pacific Security Program deputy director Jacob Stokes. The report stipulates a world in which AGI is imminent or exists and analyses the geopolitical consequences, arguing that neither state would convert AGI into dominance automatically and assessing that "AGI is not the North Star for the AI community in China the way it is in Silicon Valley — at least not yet." Its nine recommendations to US policymakers include tightening export controls, standing up allied and US-China diplomatic channels on AI, tasking the intelligence community with tracking AGI inputs and performance thresholds, running a scenario exercise on intervening to stop a Chinese AGI program, and preparing to reverse engineer or acquire AGI technology if China reaches it first (Superpowers and AGI: How U.S.-China Competition Could Be Shaped by the Emergence of Artificial General Intelligence (CNAS, August 2026)).
Public-vs-private AI investment gap finding
CNAS is the source of a public-vs-private AI investment gap figure used in Dominance by Understanding (Frazier-Rozenshtein policy frame) and Dominating AI Requires Understanding AI (Frazier + Rozenshtein, Lawfare, May 12 2026): "For every dollar allocated to NAIRR, the private sector is investing roughly $23,000 in AI." The figure is used to argue that US public-sector AI capacity lags private investment, a point connected to debates on state capacity, the National Security Commission on AI funding recommendations, and Center for Strategic and International Studies (CSIS) analysis.
Relationships
- related: Center for Strategic and International Studies (CSIS), National Security Commission on Artificial Intelligence (NSCAI), Dominance by Understanding (Frazier-Rozenshtein policy frame), Defense / Military — AI Deployment, DOD — Department of Defense (AI Deployer), Compute Governance
- supports: analysis cited in Dominating AI Requires Understanding AI (Frazier + Rozenshtein, Lawfare, May 12 2026)
- related: Superpowers and AGI: How U.S.-China Competition Could Be Shaped by the Emergence of Artificial General Intelligence (CNAS, August 2026) — its August 2026 report on AGI and U.S.-China competition
- related: US-China AI Competition: Different Races, Different Metrics, Artificial General Intelligence (AGI)