The Pentagon is seeking to acquire AI-powered coding tools that can function at the edge and assist in complex engineering tasks with minimal human involvement. This initiative is outlined in a recent call for solutions.
In collaboration with the Army, the Defense Department’s Chief Digital and AI Office (CDAO) aims to provide military and civilian developers with advanced commercial-grade AI resources to improve the speed and quality of software development for contemporary operations.
According to the solicitation, the Department of Defense (DOD) currently faces a lack of standardized, enterprise-wide access to AI-enabled coding tools that are now standard in the commercial market. This deficiency “restricts developer productivity, delays the rollout of mission-critical software, and puts the department at a disadvantage compared to the rapid pace of private sector innovation,” officials have stated.
The Pentagon is interested in products that support AI-driven code generation, optimization, debugging, and refinement directly at the point where data is generated or utilized.
Applied AI capabilities are a key priority technology area for DOD during the second Trump administration. While the department has invested significantly and issued directives for personnel to utilize AI models in their daily operations recently, efforts related to software coding and engineering have not been prominently highlighted as a major focus area.
Representatives from both the Army and CDAO have not yet responded to DefenseScoop’s inquiry regarding further details about the call for solutions.
Through this initiative, the Pentagon is “seeking solutions that reduce time to delivery while enhancing capability and interoperability with existing systems,” the solicitation explains.
The AI code will be provided through two main delivery methods: IDE-based coding assistance, which integrates with developers’ current code editors to offer code completion and chat-based assistance, or CLI-based agentic coding, which functions in the terminal or command line to carry out multi-step processes.
The desired capabilities must be available to the department’s development workforce across a range of computing environments. They also need to work within the DOD’s security and compliance framework and be capable of large-scale deployment.
“The solution or infrastructure on which it operates must either hold or be on track to achieve FedRAMP High authorization and Defense Information Systems Agency (DISA) IL5 Provisional Authorization (PA),” officials stated, referencing the security compliance approvals necessary for cloud service providers that host sensitive Controlled Unclassified Information (CUI) and National Security Systems (NSS) data.
In addition to standard software-as-a-service (SaaS) options, the Pentagon desires the flexibility to provide these AI-enabled coding tools within customer-managed cloud environments, on-premise infrastructures, and air-gapped or disconnected networks.
“The solution must be capable of deployment to a sizable developer workforce (tens of thousands of users) across various computing environments, including desktop, virtual desktop, and web-based development settings,” the officials added.
Additionally, the officials request that the products include built-in attribution and traceability mechanisms to identify or credit AI-generated code within development workflows.
The government’s submission process for this project will occur in three iterative phases, which are detailed in the call for solutions. The deadline for interested parties to submit their solution briefs is March 6, with questions regarding Phase 1 due by Friday.
