Frontier AI in the U.S. Military: Strengthening Military Operations with Artificial Intelligence
The Pentagon’s $200 million contract with OpenAI is the latest in the series of initiatives by the Department of Defence focussed on engaging cutting-edge technology firms to enhance military capabilities and introduce automation and efficiency in modern warfare.
Announced in June 2024, and extending through July 2026, the U.S. Department of Defence (DoD) has signed a $200 million contract with OpenAI, an artificial intelligence research and development organisation, to accelerate adoption of modern technologies in warfighting. This collaboration would be the latest and the biggest step forward in the heavy investments that the U.S. military has been undertaking towards ‘Frontier AI’ to revolutionise combat strategy.
What is Frontier AI?
The earliest recognition of the importance of incorporating new-age technologies in shaping future military operations has been mentioned in previous years in DoD strategy documents. The Defence Innovation Board, which is a civilian advisory committee providing recommendations on emerging technologies and innovation to the Pentagon, in its 2021 report titled, ‘Artificial Intelligence and the Department of Defence: Recommendations for Accelerating AI Adoption’, recognised the importance of technologies such as machine learning, deep learning and reinforcement learning in order for the U.S. to maintain technological superiority over its adversaries. The 2021 National Defence Authorization Act of the DoD supplemented this by reinforcing the significance of integrating AI systems in multi-domain operations and called for a shift towards increasing funding for AI. The National Security Strategy (NSS), a grand strategy released periodically by the executive branch of the government laying out national security concerns and defence objectives, has been using overlapping terms like ‘cutting-edge AI’, ‘next-generation AI’ or ‘emerging AI technologies’ to refer to the shift in how the DoD is seeking to enhance its operational efficiency, decision-making speed, and combat effectiveness through incorporating advanced frontier AI technologies.
In May 2025, the Department of Defence along with the Chief Digital and AI Office launched a Frontier AI strategic focus within its Tradewinds Solutions Marketplace, a digital platform that connects DoD to vendors offering technology solutions to ease swift acquisition of technology solutions. Tradewinds budgeted for $50 million in the financial year 2025 towards companies and research organisations that work on areas of autonomous weapons systems, intelligence, surveillance and reconnaissance, logistics, cybersecurity, training and combat decision support.
This is where Frontier AI gains relevance. Frontier AI refers to advanced artificial intelligence systems that work beyond the narrow, rule-based tasks performed by traditional AI, and are more dynamic in their operation as they make use of machine learning, neural networks, and deep reinforcement learning to complete tasks. The core technologies strengthening Frontier AI are Large Language Models (enabling natural language processing and decision support), autonomous systems (enabling real-time, unsupervised action), reinforcement learning (allowing adaptation through experience) and Edge AI (enabling data processing directly using the device not the network). These features make Frontier AI ideal for military applications as it can adapt and evolve with speed and autonomy based on the circumstances on the warfield.
The table below illustrates Frontier AI technologies and their military applications in detail:
| Technology | Description | Relevant Companies | Applications in Military | Strategic Benefits |
| Large Language Models (LLMs) | AI models capable of understanding and generating human-like text. Used for data processing and natural language tasks. | OpenAI | Cybersecurity: Detecting security anomalies Intelligence: Analysing unstructured data, SIGINT | Catching hackers early by detecting security anomalies and accelerating response to cyber threats to protect critical defence infrastructure. |
| Autonomous Systems | AI-powered robots, drones, and vehicles capable of performing tasks without direct human control. | Shield AI, Anduril Industries | Autonomous Drones: Surveillance, reconnaissance Counter-UAV: Autonomous drones to intercept enemy drones | Can be eyes in the sky in high-threat environments without risking personnel. |
| AI Planning Agents | Algorithms designed to optimise strategic decisions and assist military leaders in operational planning. | Scale AI, Anduril Industries, Microsoft | Operational Decision Support: Helping commanders make data-driven decisions in complex military theaters | Synthesising data, analysing battlefield conditions, logistics needs, and enemy movements to provide data-driven decisions. |
| Edge AI | AI systems that process data locally (at the edge), minimising latency and maintaining functionality in disconnected environments. | Second Front, EdgeRunner AI, Palantir | Edge Operations: Autonomous operations in communication-denied environments, secure local processing of mission data | Empowers autonomous systems to operate with low latency and without network connectivity- right on the battlefield. |
| Decision Intelligence | AI platforms that aggregate data from various sources to help make complex operational decisions. | BigBear.ai | Decision Support: Analysing global force deployment, optimising logistical operations, managing resource allocation | Enhance mission planning with data-driven options by quickly crunching thousands of data points and suggesting smart strategies. |
| Counter-UAV (CUAS) | AI-powered systems that detect, track, and neutralise enemy drones in military settings. | Anduril Industries | Drone Defence: Real-time drone detection and neutralisation in sensitive military zones | Protects forward operating bases and high-value fleet assets from drone incursions. |
| AI Ground Stations | Platforms that use AI to analyse satellites, sensors, and other data sources to provide actionable intelligence to military leaders. | Palantir, Anduril Industries | Data Fusion & Analysis: Real-time battlefield intelligence gathering, decision-making support for commanders | Provides commanders with real-time operational picture by combining data from satellites, sensors, drones, and troops. |
| AI Language Models | AI models that understand and generate text, optimised for classified environments to ensure data security and integrity. | Anthropic | Classified Intelligence: Analysing secure communications and intelligence data, ensuring ethical AI deployment | Safely handle top-secret information and avoid leaking. |
Future of Frontier AI
As the Department of Defence shows increased institutional commitment to AI-led capabilities, it can redefine defence policy by performing high-level tasks with precision, speed and autonomy. It is not just an upgradation of technology but a redefinition of how we understand military strategy, in terms of command structure and mission execution, as well as deterrence. Some of the ways in which Frontier AI is poised to transform defence and foreign policy for the U.S. are as follows:
- As technology overtakes intelligence analysis and issues commands, it will decrease the role of human oversight.
- The interoperability between different domains will extend to integration with allies and partners.
- The technological advantage over adversaries would strengthen deterrence.
- This perceived superiority can also trigger an arms race or preemptive unilateral actions by adversaries.
- The legal and ethical imperatives of decision-making technologies raise questions regarding international laws, norms and ethics.
- International community would need to step up in formulating international agreements and laws pertaining to the ethical use of technology in warfighting.
- The dual use of technology for civil and military purposes could be a potential legal gray zone complicating diplomacy and trust.
The biggest challenge facing the U.S. military going forward would be to avoid the pitfalls of technological overreach, arms race and accidental conflict and move towards establishing credible deterrence and the ultimate goal of conflict prevention.
Dr. Shivani Yadav is a non-resident research fellow at the America Program of the Sixteenth Council
Trump Can’t Take Greenland – and the World Knows Why