OpenAI and Google finalized Pentagon agreements this month, with Google deploying AI agents to military operations on March 10, 2026. The procurement comes as global military AI spending reaches $18 billion annually, with China's defense ministry allocating $2.7 billion for autonomous systems in 2025.
Anthropic reversed course on defense work days after filing a lawsuit against the Department of Defense on March 9. CEO Dario Amodei reopened contract negotiations while maintaining the legal challenge, pursuing what the company calls "harnessing AI to protect national security." This dual approach reflects how market dynamics override internal resistance when competitors capture government contracts worth hundreds of millions.
The Pentagon's multi-vendor strategy mirrors procurement patterns in Europe, where the EU Defense Innovation Scheme distributes AI contracts across member states to avoid single-vendor dependence. Britain, France, and Germany each signed similar deals with multiple AI providers in 2025. The approach creates competitive pressure that makes refusing defense work financially costly for AI companies.
Google's deployment follows years of internal debate over military AI, including 2018 employee protests against Project Maven that prompted the company to limit defense partnerships. OpenAI shifted from avoiding military work to accepting Pentagon contracts in 2024. The convergence of all three major U.S. AI labs toward defense work suggests competitive forces outweigh ethical objections.
This competitive procurement model creates a prisoner's dilemma for AI companies: declining defense contracts redirects revenue and influence to competitors without preventing military AI deployment. Israel's defense forces, South Korea's military, and NATO all adopted similar multi-vendor AI strategies in 2025, accelerating adoption by ensuring no single company's refusal blocks progress.
The Pentagon gains technical diversity by distributing contracts across providers with different capabilities. OpenAI contributes language models, Google provides infrastructure scale, and Anthropic offers safety research. This reduces vendor lock-in risks while maintaining negotiating leverage.
Anthropic's simultaneous lawsuit and contract pursuit indicates companies may shape defense AI policy through both legal challenges and direct participation rather than choosing one path exclusively.
Sources:
1 Yahoo Finance, "Buying A Home Just Got Cheaper: FHFA Eases Insurance Rules, Cuts Costs For Borrowers" (March 22, 2026)
2 Yahoo Finance, "Mortgage Rates Hit Three-Month High as Iran War Rattles Spring Housing Market" (March 22, 2026)
3 Yahoo Finance, "FutureGen Industries Announces Open Market Investments" (March 23, 2026)
4 News Report, "Markets on edge as U.S.-Iran conflict deepens, energy risks mount" (March 22, 2026)
5 Yahoo Finance, "Boeing’s Reduced Artemis Role And What It Could Mean For NYSE BA Investors" (March 21, 2026)