Executive Summary
AI is increasingly becoming an important component of military power, used in a wide range of defence functions. As a result, AI companies are no longer considered solely commercial technology providers but are now regarded as occupying a role comparable to that of defence contractors. Their economic significance is also on the rise. The AI military market is projected to reach 35 billion USD by 2035, driven by traditional defence contractors developing AI solutions as well as cybersecurity firms, cloud and compute providers, data analytics companies, and frontier AI developers. As military capabilities become more dependent on privately developed technologies, cooperation between governments and AI companies is becoming a matter of national security. Accordingly, governments are seeking greater influence over strategic firms and models. Measures justified on national security grounds, including the use of the “supply chain risk” designation, reflect authorities’ efforts to secure domestic technological capabilities amid intensifying geopolitical competition. However, these circumstances raise important questions regarding governance, the use of coercive measures, and the future relationship between states and frontier AI developers.
Key Points
- The military use of AI extends beyond autonomous vehicles. It extends to command and control (C2), intelligence collection and analysis, logistics, maintenance, training, decision support, and combat operations.
- The emerging AI military-industrial complex includes a diverse set of actors, such as traditional defence contractors, cybersecurity firms, data analytics companies, cloud and compute providers, and frontier AI developers.
- Companies such as Shield AI, Anduril, Amazon Web Services, Oracle, OpenAI, Microsoft, NVIDIA, AMD, HENSOLDT, SparkCognition, and Helsing SE are playing an increasingly important role in national security and defence operations.
- The diversity of actors influences the nature of cooperation with authorities. Conventional defence contractors such as Anduril, Palantir, and Helsing have traditionally operated in close cooperation with public authorities, but frontier AI and other dual-use technology companies such as Anthropic and OpenAI maintain stronger ties with the civilian market and often have looser cooperation links in national security matters.
- The Anthropic designation as a “supply chain risk” illustrates governments’ willingness to use coercive tools to ensure control over frontier AI developers.
Analysis
The military use of AI is redefining the relationship between governments and developers. Unlike traditional defence contractors, many leading AI developers operate primarily in the civilian sector and derive most revenue from commercial markets – creating new challenges for cooperation. Governments increasingly consider advanced AI capabilities as strategic assets that contribute to national security, military operations, and geopolitical influence. As a result, cooperation with AI companies has become a critical component of securing access to advanced technologies and reducing dependency on foreign suppliers.
Efforts to secure alignment with national security priorities are generating friction, since they are restricting AI companies’ autonomy, commercial activities, and international partnerships. The designation of Anthropic may have important implications for government-private sector cooperation, as it demonstrates how federal procurement processes can be used as leverage over frontier AI developers.
Policy Recommendations
- Design cooperation mechanisms between governments and AI companies that balance national security requirements with the commercial realities and safety commitments of frontier AI developers.
- Prioritise incentives and cooperative approaches over restrictive or coercive regulatory measures. Excessive reliance on the latter can undermine trust and discourage long-term cooperation.
- Develop institutional mechanisms for dialogue and dispute resolution to address the demands, needs, and concerns of frontier AI developers and incentivise cooperation.
- Dual-use technologies have been governed by different approaches: decentralised, where the private sector leads with government support; structured coordination, involving government authorities, academia, and the private sector; and centralised, where the state defines the approach. Choosing the correct model will enable more effective cooperation.
Bibliography
Aebi, T., Pankov, A., Malkov, A., Merhaba, A., and Stella, C. (2025) “Unlocking the Strategic Power of Dual-Use Technologies”, Arthur D. Little. Available at: https://www.adlittle.com/en/insights/viewpoints/unlocking-strategic-power-dual-use-technologies (Accessed: 25 June 2026).
AI Magazine (2025) “Artificial Intelligence (AI) In Military Market Projected to Grow to $35 USD Billion By 2035 as Global Uncertainty Increases”, AI Magazine. 12 Nov. Available at: https://aimagazine.com/globenewswire/3186361 (Accessed: 22 June 2026).
Nazil, A.R. (2025) “AI at War: The next revolution for military and defense”. World Journal of Advanced Research and Reviews, 27(1), 1998-2004. Available at: https://doi.org/10.30574/wjarr.2025.27.1.2735
Bego, K. (2026) “How a surge in defence and dual-use technology investment could reconfigure the global AI race”. Chatham House. Available at: https://www.chathamhouse.org/2026/04/how-surge-defence-and-dual-use-technology-investment-could-reconfigure-global-ai-race (Accessed: 22 June 2026).
Besliu, R. (2026) “Europe’s AI Act Leaves a Gap for Military AI Entering Civilian Life”, Tech Policy Press. 10 May. Available at: https://techpolicy.press/europes-ai-act-leaves-a-gap-for-military-ai-entering-civilian-life (Accessed: 20 June 2026).
Blanchard, A., Boulanin, V. and Bruun, L. (2026) “Mapping the military AI industry”, SIPRI. 23 April. Available at: https://www.sipri.org/commentary/topical-backgrounder/2026/mapping-military-ai-industry (Accessed: 21 June 2026).
Gvaryahu, A. (2026). “These Aren’t AI Firms, They’re Defense Contractors. We Can’t Let Them Hide behind Their Models”, The Guardian. 15 March. Available at: https://www.theguardian.com/us-news/ng-interactive/2026/mar/15/ai-defense-warfare-companies (Accessed: 23 June 2026).
Harper, J. (2026) “Centcom Commander Touts Use of AI in Fight against Iran during Operation Epic Fury”, DefenseScoop. 11 March. Available at: https://defensescoop.com/2026/03/11/us-military-using-ai-against-iran-operation-epic-fury-adm-cooper/ (Accessed: 24 June 2026).
O’Connor, N. (2026). “Artificial Intelligence at War: How the US and Israel Is Using AI in Its Bombing of Iran”, The Journal. 12 March. Available at: https://www.thejournal.ie/ai-artificial-intelligence-targeting-iran-war-6980425-Mar2026/ (Accessed: 26 June 2026).
Quay-de la Vallee, H. and Anex-Ries, Q. (2026) “Chain Reaction: What the Pentagon-Anthropic Dispute Means for Civilian Agencies Across All Levels of Government”, Center for Democracy & Technology. 12 May. Available at: https://cdt.org/insights/chain-reaction-what-the-pentagon-anthropic-dispute-means-for-civilian-agencies-across-all-levels-of-government/ (Accessed: 25 June 2026).
Rashid, A.B., Kausik, A.K., Hassan Sunny, A.A. and Bappy, M.H. (2023) “Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges”, International Journal of Intelligent Systems. Available at: https://doi.org/10.1155/2023/8676366
Robins-Early, N. (2026) “Anthropic-Pentagon Battle Shows How Big Tech Has Reversed Course on AI and War”, The Guardian. 13 March. Available at: https://www.theguardian.com/technology/2026/mar/13/anthropic-pentagon-artificial-intelligence (Accessed: 27 June 2026).
The Guardian (2026) “Pentagon Inks Deals with Eight AI Companies for Classified Military Work”, The Guardian. 1 May. Available at: https://www.theguardian.com/us-news/2026/may/01/pentagon-us-military-pairs-with-spacex-google-openai (Accessed: 22 June 2026).