OpenAI and Google both finalized agreements with the Pentagon this month, with Google introducing AI agents to military operations on March 10, 2026. The dual procurement happened as Anthropic CEO Dario Amodei reopened defense contract talks just days after his company filed a lawsuit against the Department of Defense on March 9.
Anthropic's rapid pivot from litigation to negotiation signals how competitive pressure shapes AI vendor decisions. The company stated it seeks "harnessing AI to protect national security" while simultaneously challenging DOD practices in court. This dual-track approach—legal opposition paired with contract pursuit—reflects the financial stakes of declining government work while competitors accept it.
The Pentagon's strategy distributes AI capabilities across multiple vendors rather than consolidating with a single provider. This reduces vendor lock-in risks and creates competitive dynamics that make it costly for AI companies to refuse government contracts. When one major AI lab declines defense work, others can capture that market share and influence over military AI development.
Google's March 10 agent deployment follows years of internal debate over military AI work, including employee protests over Project Maven in 2018. OpenAI previously avoided defense partnerships but shifted its policy stance in 2024. The convergence of these three companies—each with different ethical frameworks—toward Pentagon contracts suggests market forces outweigh internal resistance.
Government and defense AI adoption now relies on competitive procurement rather than exclusive partnerships. This creates a prisoner's dilemma for AI companies: declining defense work on ethical grounds may simply redirect influence and revenue to competitors without changing government AI deployment. The strategy accelerates military AI adoption by ensuring no single vendor refusal can block progress.
The multi-vendor approach also distributes capability development across organizations with different technical strengths. OpenAI's language models, Google's infrastructure scale, and Anthropic's safety research each address different military AI requirements. The Pentagon gains redundancy and negotiating leverage by avoiding dependence on any single AI provider.
Anthropic's lawsuit continuation alongside contract negotiations indicates companies may seek to shape defense AI policy through both legal challenges and direct participation rather than choosing one path exclusively.