Anthropic bid $100 million for the Pentagon drone swarm project, but was not selected; SpaceX and OpenAI camps emerged victorious

**Anthropic Unselected for Pentagon’s $100M Drone Swarm Challenge; OpenAI Secures New Classified DoD AI Deal** Anthropic submitted a proposal earlier this year to compete in the Pentagon’s $100 million Orchestrator Prize Challenge, which aims to develop voice-controlled autonomous drone swarm technology. The challenge—led by the Defense Autonomous Warfare Group (DAWG) under U.S. Special Operations Command (SOCOM) and the Defense Innovation Unit (DIU)—includes five phases: starting with software development, advancing to live-fire tests, and later focusing on “target perception and sharing” and “from launch to effects.” Anthropic’s pitch centered on its Claude AI, which translates commander intent into digital commands to coordinate drone swarms. The plan excluded autonomous targeting or weapon decisions, with human oversight required at all times. The company also proposed a joint research effort with the Pentagon to securely develop and assess autonomous weapon capabilities, arguing its proposal did not cross its “no fully autonomous weapons” red line (humans could monitor and terminate the system at any point). Bloomberg could not confirm why Anthropic was not selected. Chosen proposals include a joint bid from SpaceX and xAI, plus two defense tech firms (one being autonomous military vehicle testing contractor Applied Intuition) that list OpenAI as an AI partner. OpenAI’s technology will support “mission control” by converting voice commands into digital instructions. Hours after the Pentagon announced a ban last Friday on its contractors engaging in commercial activity with Anthropic, OpenAI revealed a new agreement with the Department of Defense (DoD) to use its AI tools on a classified cloud system. Anthropic declined to comment. (Source: Bloomberg)