Diplomatic Analysis: The proliferation of accessible AI capabilities necessitates a shift in US defense acquisition strategy, favouring internal software development over wholesale commercial purchasing.
Overview
The US Department of Defense (DoD) is undergoing a significant re-evaluation of its acquisition strategies, driven by rapid advancements in Artificial Intelligence (AI). A recent policy shift towards “commercial-first” procurement aimed to accelerate the delivery of capabilities to warfighters. However, the exponential growth in AI, particularly in Large Language Models (LLMs) and “agentic” AI tooling, is fundamentally altering the cost-benefit analysis of this approach. This analysis examines the evolving landscape, detailing the key actors and their positions, assessing the implications of these changes, and outlining the likely trajectory of US defense software procurement. The shift indicates a move away from simply buying solutions to building them internally, leveraging commercial infrastructure and AI access rather than complete systems.
Historical Context
For decades, the DoD has grappled with lengthy and expensive acquisition processes. Traditional defense contracting is characterised by complex requirements, bespoke development, and substantial bureaucratic overhead. The ‘commercial-first’ policy, announced in late 2026, was an attempt to circumvent these bottlenecks by prioritising off-the-shelf commercial solutions wherever possible. This approach aimed to reduce development timelines and costs, acknowledging that the lengthy processes of custom development often resulted in solutions being obsolete by the time they were deployed. However, this strategy was predicated on the assumption that commercially available software offered a significant advantage in speed and cost-effectiveness. The recent surge in AI capabilities, in particular the emergence of accessible and powerful agentic AI tools, has challenged that calculation. The traditional limitations of internal government software development – cost, time to market, and the need for specialised skills – are being rapidly eroded by these new technologies.
Key Actors & Positions
The primary actor is the US Department of Defense, particularly the Office of the Secretary of Defense and the individual military services (Army, Navy, Air Force, Marines). Within the DoD, there is a growing awareness, as highlighted in the recent analysis, that the initial “commercial-first” strategy requires recalibration. The Army, in particular, through initiatives like the Automation and Data Integration Office, is leading the experimentation with agentic AI and in-house development.
The commercial defense industry is a key stakeholder. While generally supportive of continued government spending, the industry faces a potential shift in funding allocation. A move towards internal development could reduce demand for large, system-level contracts, requiring them to adapt to providing infrastructure, AI access, and specialized tooling rather than complete software solutions.
Finally, open-source AI communities and companies providing cloud infrastructure and AI access (e.g., those offering API access to LLMs) gain prominence, potentially becoming key partners in the new ecosystem. This creates opportunities for greater competition and potentially lower costs.
Analysis
The core argument presented is that, while the DoD should continue to procure commercial infrastructure and access to AI models, it should significantly expand its internal software development capabilities. Agentic AI – AI capable of independently planning and executing tasks – dramatically lowers the barrier to software development. As demonstrated by the author’s pilot program, complex projects that previously required months and substantial funding can now be completed by individuals with modest coding skills in a matter of days, and at a fraction of the cost.
This presents several implications. First, it reduces the DoD’s dependence on potentially vulnerable proprietary systems and ensures full government ownership of critical software. Second, it increases responsiveness and adaptability, allowing for rapid iteration and customisation of software solutions to meet evolving battlefield needs. Third, it fosters innovation by empowering developers within the services to experiment and develop solutions tailored to specific challenges.
However, challenges remain. Implementing this shift requires significant investment in training the existing workforce, establishing secure and accredited development environments, and revising acquisition procedures. Concerns surrounding the security of open-source software and the potential for misuse of AI necessitate robust governance frameworks and rigorous testing procedures. The potential for vendor lock-in, even with commercial infrastructure, must be addressed through diversification and the adoption of open standards like the Model Context Protocol. The relatively nascent nature of agentic AI also introduces risks related to reliability and unforeseen consequences.
Outlook
The DoD is highly likely to refine its acquisition strategy in the coming years, moving towards a hybrid model. Commercial solutions will continue to be pursued for infrastructure and foundational AI capabilities, but an increasing proportion of application development will be brought in-house. This will involve greater emphasis on internal training programmes, standardised development environments, and internal competition to incentivize efficiency and innovation.
The adoption of open standards like the Model Context Protocol will accelerate, reducing reliance on proprietary data platforms and facilitating interoperability. Securing the software supply chain, particularly in relation to open-source components, will be a paramount concern. The performance of early pilot programs and the speed with which internal development teams demonstrate success will be crucial in driving further investment and adoption of this new approach. The shift will not be immediate, but the pressures of rapidly evolving technology and the demonstrated cost savings and control offered by in-house development make a recalibration of the “make-or-buy” line inevitable.
Sources
Green, Jacob A. “The Make-or-Buy Line has Moved.” War on the Rocks, July 20, 2026. [https://warontherocks.com/2026/07/the-make-or-buy-line-has-moved/](https://warontherocks.com/2026/07/the-make-or-buy-line-has-moved/)
