Musk's companies join Pentagon competition for voice-controlled autonomous drone swarms

Elon Musk's companies are participating in a Pentagon competition to develop voice-controlled, autonomous drone swarm technology.

The $100 million competition was launched in January as a joint effort between the Defense Innovation Unit and the newly established Defense Autonomous Warfare Group (DAWG), which operates under US Special Operations Command. DAWG partly continues the work of the Biden-era Replicator initiative, which aimed to produce thousands of autonomous single-use drones. The US military had already invested $1 billion in Switchblade kamikaze drones two years earlier.

The six-month competition is structured in five phases, from software development through real-world testing. The goal is software that converts voice commands into digital instructions and coordinates drone swarms across multiple domains like air and sea. Later phases also include target acquisition and ultimately engagement, according to the Pentagon's description. A defense official said at the announcement that human-machine interaction would directly affect the lethality of these systems.

Musk signed an open letter from AI and robotics researchers in 2015 warning about the dangers of autonomous weapons and spoke out against developing new tools for killing. While SpaceX is an established defense contractor, the company has focused on rockets and satellites - not offensive weapons software. SpaceX and xAI had only announced their merger in early February, valued at $1.25 trillion.

xAI is building a systematic Pentagon presence

xAI is actively recruiting engineers with US security clearances at the "Secret" or "Top Secret" level to work with the Department of Defense. The company has already secured a $200 million contract with the Pentagon to integrate xAI into military systems and announced contracts in December to embed its Grok chatbot in government systems.

Several defense officials told Bloomberg they have concerns about integrating generative AI into weapons platforms. They stressed the importance of limiting generative AI to translating commands rather than letting it control drone behavior. Still, the Pentagon's new AI acceleration strategy from January aims to deploy AI agents broadly across the battlefield.

OpenAI draws a line at weapons integration

OpenAI is also participating in the competition through partner Applied Intuition, but according to Bloomberg, the company is explicitly limiting its contribution to functions like voice control. OpenAI's technology will not be used for operating drone swarms, weapons integration, or target selection. SpaceX and xAI, by contrast, are reportedly collaborating across the entire project.

OpenAI has been active in the defense sector since 2024. In December 2024, the company announced a strategic partnership with Anduril Industries to develop AI-powered counter-drone systems.

The partnership sparked internal criticism at OpenAI. According to the Washington Post, employees raised concerns about military use of the technology and questioned whether deployment could permanently be limited to purely defensive purposes. Management maintained the collaboration was restricted to defense systems, even though Anduril itself manufactures armed drones.

In June 2025, OpenAI received its first official Department of Defense contract worth $200 million. The one-year deal covers support for military healthcare, program data analysis, and cyber defense. All applications must comply with OpenAI's usage policies, according to the Pentagon.

Pentagon pushes for unrestricted access to commercial AI models

In parallel, the Pentagon ramped up efforts in 2026 to make leading AI models available on classified networks without the usual usage restrictions. OpenAI reached an agreement for the open network genai.mil that lifted many usage restrictions while keeping some safeguards in place. Expanding to classified networks would require a new agreement, according to OpenAI.

Anthropic, meanwhile, is locked in a dispute with the Department of Defense over the company's demand for guarantees that its AI won't be used for autonomous weapons control without adequate human oversight or for domestic surveillance. The Pentagon has rejected those restrictions, insisting it can deploy commercial AI technology regardless of manufacturer guidelines as long as US law is followed.