Get your free personalized podcast brief
We scan new podcasts and send you the top 5 insights daily.
To prevent vendor lock-in, defense programs should keep data ownership, data governance, and pipeline architectures strictly in government hands while outsourcing approximately 90% of operational work to commercial industry. Internal government engineers should develop enough capability to maintain intellectual parity with the state of the art, using a modular acquisition framework that rotates competitive industry partners across data labeling, training, and deployment.
Related Insights
The Pentagon expects to buy AI with full control, just as it buys an F-35 jet from Lockheed, without the manufacturer dictating its use. AI firms like Anthropic see their product as an evolving service requiring ongoing involvement, creating a fundamental paradigm clash in government contracting.
As AI gets embedded in core workflows, the key strategic question becomes who owns the resulting intelligence. Enterprises are wary of outsourcing their core logic to model providers who have explicitly stated they will compete in their customers' industries, making ownership of these learnings paramount.
According to the Under Secretary, the foundational mistake that led to the Anthropic conflict was a previous administration's decision to rely on one AI provider. This created a monopolistic scenario, giving the vendor outsized leverage. The current strategy mandates a multi-vendor ecosystem to ensure competition and balance power.
The US nuclear weapons industry operates as a hybrid: the government owns the IP and facilities, but private contractors like Honeywell and Boeing operate them and build delivery systems. This established public-private partnership model could be applied to manage the risks of powerful, privately-developed AI.
The Department of War's top AI priority is "applied AI." It consciously avoids building its own foundation models, recognizing it cannot compete with private sector investment. Instead, its strategy is to adapt commercial AI for specific defense use cases.
Current military cybersecurity frameworks, such as the Risk Management Framework, treat all software identically, requiring tedious full package reaccreditation for every change. Patel argues the Department of Defense should separate AI models and data from base software infrastructure. This architectural decoupling allows defense teams to rapidly push frequent model updates to operational edge units without triggering an entire software re-authorization cycle.
Sending proprietary enterprise data to external foundational models is a critical mistake that 'leeches' value and intellectual property. The correct, secure approach is to bring AI models into a company's own air-gapped or on-premise environment to maintain data sovereignty and control.
Over-reliance on bespoke, defense-only suppliers creates a fragile industrial base. The Department of Defense now champions a "commercial first" strategy, prioritizing designs with commercial applicability to build resilience, reduce costs, and avoid self-inflicted vulnerabilities from single-source dependencies.
The U.S. government cannot develop leading AI in-house primarily because it lacks the technical talent. Crucially, it also cannot compete with the massive private capital mobilized for building data centers and training models. The commercial applications are so vast that they dwarf the defense sector's budget and influence.
For many companies, 'AI sovereignty' is less about building their own models and more about strategic resilience. It means having multiple model providers to benchmark, avoid vendor lock-in, and ensure continuous access if one service is cut off or becomes too expensive.
