Pentagon’s AUTOWARCOM bets on modular autonomy, open standards, and trust

Key facts

• AUTOWARCOM is planned to launch by October next year, aiming to accelerate deployment of autonomous and robotic systems focused on combat adaptability.
• The article argues for modular, open-standards autonomy and for separating autonomy software (“brains”) from platforms (“body”), citing NATO CMRE experimentation in Portugal and a referenced STANAG 4817 standard.
• It highlights EW-denied operating assumptions (GPS/comms degraded) and warns that trust and reliability remain central risks, referencing Patriot fratricide analysis and recent AI-agent misalignment incidents.

3 minute read

US defense leadership is preparing to launch an Autonomous Warfare Command (AUTOWARCOM) by October next year, intended to compress the cycle from concept to deployment for autonomous and robotic systems and to prioritise combat adaptation over bespoke platforms. A Sept. 30 memo from Defense Secretary Pete Hegseth positions the initiative around “the ability to adapt in combat” against “realistic opposition,” arguing that contemporary battlefields impose week-scale adaptation demands that outpace traditional defence-industrial update and patch cycles.

The article stresses that past successes point to a mission-led approach. Maven Smart System—described as widely adopted across the US military and NATO—features as the exemplar: a “clear, urgent, but broad problem definition” allowed the programme to align with user needs and field capability quickly. Operational pull is illustrated through remarks by Gen. Frank Donovan, commander of U.S. Southern Command and former leader of the Defense Autonomous Warfare Group, who emphasised domain awareness and matching robotic modalities to environment (“it swims, it flies, it has feet”) and framed autonomy less as “robots” than as a data environment that forward users can tap via common devices.

Resilience under electronic warfare is treated as non-negotiable. A former senior US defence official is quoted warning that GPS and communications will be denied or degraded in any future conflict—conditions already observed in Ukraine and the Middle East—driving demand for higher degrees of autonomy that can function under intermittent or absent connectivity.

Procurement implications are explicit: the Pentagon should “buy the brains and body separately,” mirroring the automotive model in which vehicle manufacturers partner with specialist autonomy firms, and mandate modular autonomy packages that can be upgraded as threats evolve. Open standards are presented as the scaling mechanism across a force and across alliances. NATO’s Centre for Maritime Research and Experimentation (CMRE) experiment in Portugal is cited as demonstrating software-driven tasking of drones, enabled not by a new sensor but by partner agreement to adopt a standard (referencing STANAG 4817 for multi-domain control of unmanned platforms).

The limiting factor remains trust and high reliability in complex settings. Drawing on a 2017 study of Patriot autonomy failures—including the 2003 shootdown of an F-18 in Iraq—the piece argues that intent misunderstanding and human-machine interaction failures can be decisive, and that newer AI-agent “misalignment” incidents reinforce the risk that autonomy can execute the wrong interpretation of an operator’s goal. For European stakeholders, the message is twofold: interoperability via NATO standards can accelerate coalition autonomy adoption, but certification, training, and human factors engineering will likely become the pacing items for fielding and delegated-authority concepts.

Source: Defense One