The Pentagon's 12th combatant command: what AUTOWARCOM actually orders, and what it doesn't

On September 30, 2026, Defense Secretary Pete Hegseth announced the creation of an Autonomous Warfare Command — a four-star combatant command with 'service-like authorities' aimed at scaling drones and robotic systems across the joint force, with a target establishment date of October 1, 2027. The announcement came with a signed memo, an interim program called Project Agincourt, and a set of new military occupational specialties. It also came without any published framework for human control of lethal autonomous systems.

Every few decades the Pentagon stands up a new combatant command, and the last one was US Space Command in 2019. On September 30, 2026, Defense Secretary Pete Hegseth announced the twelfth: the Autonomous Warfare Command, or AUTOWARCOM, a four-star functional command intended to lead unmanned and autonomous warfare across the joint force, with a target establishment date of October 1, 2027.

Why a new command at all

Hegseth's stated rationale is institutional, and it is a fair one. In his words: "The pace of war is changing faster than the process to support it." And: "Cheap compute, superintelligence, and advanced commercial manufacturing have enabled the proliferation of low-cost, high-precision strike." Reuters reported that drones are estimated to account for roughly 70% of Russian casualties in Ukraine — a figure Hegseth's framing leans on — and that the US military's traditional answer to that shift, buying fewer and ever-more-sophisticated weapons platforms, is being read as unsustainable inside the building.

The battlefield argument is not contested. The interesting question is whether a combatant command is the right instrument. A combatant command is a geographic and functional authority: it defines joint requirements, integrates service-provided forces, and supports geographic commanders. Hegseth's memo says exactly that — AUTOWARCOM "will define joint needs and integrate Service-provided forces in support of geographic combatant commanders." That is a requirements-and-integration body, not a targeting authority, and the distinction will matter a great deal.

Project Agincourt: the twelve-month bridge

The command is a year away, so Hegseth paired the announcement with an interim effort. The Project Agincourt memo directs the existing Direct Reporting Portfolio Manager for Unmanned Systems (DRPM-UxS) to run the transition. Military Times and AeroTime both trace that office back to a June 29, 2026 consolidation that made DRPM-UxS the single joint integrator for unmanned and autonomous systems, reporting to the Deputy Secretary; it absorbed oversight of Joint Interagency Task Force 401, the Defense Innovation Unit and the Defense Autonomous Warfare Group.

The memo's mechanics are unusually concrete. DRPM-UxS gets 30 days to submit a plan of action and milestones, supported by the Chairman of the Joint Chiefs of Staff. It must collaborate directly with Congress to secure the legislation needed to stand up a four-star command — which is the real constraint, since Congress has to authorise the command structure and its authorities. Owen West, the Defense Innovation Unit director, and Navy SEAL Senior Chief Max Strasiser serve as the office's CEO and COO respectively, which Hegseth described as a partnership between strategic scale and operational problem-solving. The memo also directs the military departments to build new military occupational specialties, ratings or career fields for unmanned and autonomous systems, and requires DRPM-UxS to report monthly on operational results and Service adoption.

The acquisition idea is the part worth stealing for any reader who builds systems for a fast-moving market. Hegseth called it "warfighting acquisition": fusing operators with engineers and competitive manufacturers "in rapid adaptation cycles while distributing decisions and dollars closer to frontline units." The memo's phrasing of the goal is unusually good — the output is not a particular drone, but the ability to adapt in combat, on the right side of the cost exchange. He also described a "high-low mix": expensive high-end systems as the lightning, massed cheap autonomous systems as the wind and rain of the thunderstorm. Defense One reported that a companion panel, Project Meridian, was commissioned with a four-month deadline and is co-led by Elon Musk, Palmer Luckey and Newt Gingrich.

What is missing

This is the part most commentary skipped. Read the memo against what a reader would expect from an announcement about scaling lethal autonomous systems, and three documents are conspicuously absent: any framework for human control of autonomous weapons in the field; any statement of who holds targeting authority; and any escalation-control mechanism. The memo's own words describe a testing and fielding mandate — "It will test equipment and force designs with realistic opposition in demanding exercises and rugged environments, modify rapidly, and field what works" — and an organisational one. Training realistic exercises with opposition forces is not the same thing as defining rules of engagement for autonomous strike.

Military Times reported that Hegseth said realistic exercises, and "when possible, combat," will be used to test autonomous systems. Read literally, that sentence is doing a lot of work. It is a testing statement and a fielding statement; it is not a statement about who presses the trigger, and the published announcement documents do not supply one.

The absence is not unique to this announcement. Reuters noted that security experts from both the US and China — the two major drivers of cutting-edge AI development — have been at loggerheads over AI policy and have called for guardrails and consensus-building on military AI, including for nuclear weapons. An institution created specifically to accelerate autonomous capability will, presumably, address governance eventually — but the announcement as it stands does not, and a command structure that is stood up in twelve months will have its doctrine written before the debate is settled.

What this means for you

• If you sell into defence or public sector: the practical signal is the career-path language. New military occupational specialties and a four-star command with directed acquisition authority is a budget line and a requisition pathway that did not exist before. DRPM-UxS reporting monthly to the secretary means the reporting rhythm to expect is monthly, not annual.
• If you build AI agents or autonomy software: the "high-low mix" is the demand signal. The stated strategic priority is cheap, high-volume autonomous systems on the right side of the cost exchange, not exquisite platforms. That is a volume-and-price-pressure thesis, and it favours small, cheap, adaptable systems over large expensive ones.
• If you follow AI governance: the gap in this announcement is the story to track. Watch for the plan of action and milestones, due within 30 days of the memo, for whether it contains any human-control or targeting-authority language at all.

Bottom line: the Pentagon has created its twelfth combatant command to buy and field autonomous systems faster, and the memo behind it is unusually concrete. What it does not contain is any rule for the part that matters most — who decides when an autonomous system is allowed to kill. That absence is not an oversight in reporting; it is an absence in the published documents, and the plan of action due within 30 days is where to look for whether it gets filled.

Prices and benchmark figures are standard published rates as of October 2, 2026 and can change without notice. Some outbound links may be affiliate links — see our disclosure. Found an error? Tell us and we'll correct it in public.