A four-star command for autonomous warfare, and the statute it needs

FACT (primary document, read in full). On 2026-09-30, the Secretary of War delivered a "State of the Force" address at Marine Corps Base Quantico, Virginia, to an audience of active-duty E-6s and O-3s, and used it to announce six initiatives. One of the six is the Autonomous Warfare Command.

FACT — the department's own words. The official release, "Secretary Hegseth Announces Six Major Initiatives at Quantico Speech" (Sept. 30, 2026):

"The establishment of the Autonomous Warfare Command represents the fastest peacetime operational shift in modern military history, designed to scale autonomous warfare across the Joint Force as a 4-star functional combatant command by October 1, 2027. Building on the success of the Department's Drone Dominance program, which has delivered thousands of robotic systems to the Middle East in support of Operation Epic Fury, the Secretary is directing DRPM UxS to lead Project Agincourt."

"To lead this historic transition, Secretary Hegseth appointed Owen West, the current Director of DIU, and Navy SEAL Senior Chief Max Strasiser, the first enlisted SEAL to graduate from Navy Test Pilot School, to direct DRPM UxS under a unique CEO-COO partnership."

FACT — Hegseth's own description of the authorities. Reported verbatim and identically by Military Times, DefenseScoop, Breaking Defense and Air & Space Forces Magazine, from the speech itself:

"service-like authorities built to scale autonomous and robotic capabilities across the joint force in the fastest peacetime shift in modern military history"

FACT — and why the word matters. The same primary release contains a pointed line about why a command rather than another office:

"The autonomous and robotic revolution has enormous implications for equipment, doctrine, organization, and the assumptions that have governed warfare for decades. Civilian leaders can provide direction and urgency, but lasting change must be anchored in military command structures and operationalized by leaders across our joint force." (Hegseth, at Quantico)

FACT — Hegseth's framing of the problem. "The pace of war is changing faster than the process to support it." He argued the department had spent forty years "buy[ing] fewer, increasingly sophisticated weapons," and that "advanced computing now enables precise, low-cost weapons to be produced at massive scale."

FACT — Reuters framing. Reuters reported the same day that Hegseth "announced the creation of a new military command that would be dedicated to building and providing autonomous and robotic capabilities across the US military"; that "the last one was US Space Command in 2019"; and that the US military "currently has 11 combatant commands."

FACT — the interim mechanism. The same day Hegseth signed a memorandum, "Project Agincourt — Pathway to the Autonomous Warfare Command," addressed to "SENIOR PENTAGON LEADERSHIP / COMMANDERS OF THE COMBATANT COMMANDS / DEFENSE AGENCY AND DOW FIELD ACTIVITY DIRECTORS." It directs the Direct Reporting Portfolio Manager for Unmanned Systems (DRPM-UxS) to lead Project Agincourt, with two missions:

"Establish the pathway to the Autonomous Warfare Command, a 4-star functional combatant command that will lead the Joint Force in unmanned and autonomous warfare. This command will possess directed manpower, budget, acquisition authorities and create dedicated military career pathways for Officers and Enlisted personnel. DRPM-UxS will drive the immediate organizational shifts required and collaborate directly with Congress to secure necessary legislation with a target establishment date of October 1, 2027."

"Execute the existing DRPM-UxS mission while prototyping a new acquisition model for unmanned and autonomous systems. Teaming operators with innovators and competitive manufacturers, DRPM-UxS will push decisions and dollars to the edge of our operational forces. This must yield a repeatable procurement system whose output is not a particular drone, but the ability to adapt in combat, on the right side of the cost exchange."

FACT — the charter sentence that defines the command's role. "Autonomous Warfare Command will define joint needs and integrate Service-provided forces in support of geographic combatant commanders. It will test equipment and force designs with realistic opposition in demanding exercises and rugged environments, modify rapidly, and field what works."

FACT — accountability already attached. The memorandum's attachment ("Accountability and Implementation Requirements") sets: a 30-day plan of action and milestones (POAM) due from DRPM-UxS with CJCS support (i.e. around 30 October 2026); a requirement that the Secretaries of the Military Departments, coordinated by the Under Secretary of War for Personnel and Readiness, produce "phased manning plans and proposals to establish or adapt military occupational specialties (MOSs), ratings, or equivalent career fields" for unmanned and autonomous systems; and monthly reporting by DRPM-UxS to the Secretary on "operational results, Service adoption, progress against the plan, and decisions required."

FACT — leadership of the interim team. Owen West and Max Strasiser serve "as the Chief Executive Officer and Chief Operating Officer, respectively, of the DRPM-UxS." Hegseth described the split: West on "strategic risk, resources, and scale," Strasiser on "operational problem-solving and the realities of employment in combat."

FACT — no commander named. "We will soon have a four-star commander to raise the AutoWarCom flag [and] prepare the way." (Hegseth.) No individual was named on 2026-09-30.

INDEPENDENT EVIDENCE — immediate organisational consequence (2026-10-02, inside the window). DefenseScoop reported that Owen West is taking a "leave of absence" from his post leading the Defense Innovation Unit to serve as CEO of the DRPM-UxS, that his deputy Travis Metz was elevated to acting DIU director, and that both will report directly to Hegseth — West in his new capacity, Metz as a principal staff assistant.

The change is institutional, not technical. No new weapon, algorithm or capability was announced. What changed is who is accountable, who holds authority, and who is paid from which budget line.

3.1 Naming — settled only in the long form, and this matters for anyone building tooling or tracking the entity

• Official: "Autonomous Warfare Command." The department release and the memorandum use only this. No acronym.
• Same-day variants in circulation: "AutoWarCom" (Military Times, DefenseScoop), "AUTOWARCOM" (Breaking Defense), "Autowarcom" (Reuters headline and body). GlobalSecurity notes Bloomberg used "Autowarcom" and the Washington Examiner "AUTOWARCOM".
• INTERPRETATION: any downstream system — a vendor CRM record, a contract-tracker entity, a sanctions/registry screen, an OSINT query — should key on the long form plus the document mark OSD010441-26/CM DO 13361-26, and treat the short forms as aliases that were never standardised.

3.2 The department's own name — also unsettled

FACT. The release is published on war.gov under the masthead "U.S. Department of War", is signed "SECRETARY OF WAR", carries the tags "secretary of war" and "service", and the memorandum's address block reads "SECRETARY OF WAR / 1000 DEFENSE PENTAGON".

FACT — the press did not converge. Reuters, Military Times, DefenseScoop, Breaking Defense and Air & Space Forces Magazine all used "Defense Secretary" / "Secretary of Defense" for the same person on the same day.

INDEPENDENT EVIDENCE. DefenseScoop, 2026-10-02: "President Donald Trump signed an executive order rebranding DOD as the Department of War last year, but officially changing the agency's name requires an act of Congress." Separately, Breaking Defense reported that the FY2027 Senate defence bill contained a department-endorsed amendment to change the name, which Senator Tim Kaine called "juvenile" and "embarrassing."

INTERPRETATION: the story is therefore reported correctly as a Pentagon / Defense Department event by the trade press, and correctly as a Department of War action by the primary documents. Both are accurate descriptions of a live, legally unresolved branding split.

3.3 The single most consequential change: from an acquisition office to a command

FACT. The immediately preceding structure, created roughly three months earlier, was an acquisition office — the DRPM-UxS, established by a 29 June 2026 memorandum (released 1 July 2026), reporting to the Deputy Secretary. Military Times reported that it "assumed oversight of existing efforts, including Joint Interagency Task Force 401, the Defense Innovation Unit and the Defense Autonomous Warfare Group."

FACT. Its own founding instrument drew a line that the September announcement crosses. As characterised in reporting, the portfolio manager's authority "does not extend to the operational employment of military forces."

INTERPRETATION. This is the real delta. An acquisition office buys and sets standards; a combatant command can be assigned forces and employ them. Hegseth explicitly chose the heavier instrument, and said why: "lasting change must be anchored in military command structures." Everything else in the announcement — the acquisition reform, the MOS frameworks, the executive titles — follows from that choice.

3.4 The acquisition model, named

FACT. Hegseth called the new model "warfighting acquisition":

"This model fuses operators with entrepreneurs in rapid adaptation cycles while distributing decisions and dollars closer to frontline units to combatant commands." (via Military Times)

Military Times reported that "realistic training exercises — and, when possible, combat — will be used to test autonomous systems and determine which technologies should be rapidly expanded."

FACT. The doctrine of quantity was made explicit — the "high-low mix." Hegseth described high-end weapons as the lightning and massed autonomous systems as "the wind and rain" of the same storm: "Today, we need both quality and quantity." Reuters quoted the same line.

FACT — the Secretary's own causal claim (COMPANY CLAIM, not verified). "Cheap compute, superintelligence, and advanced commercial manufacturing have enabled the proliferation of low-cost, high-precision strike." (Reuters)

FACT — the portfolio, as reported. Air & Space Forces Magazine: the command's "portfolio will include drones, artificial intelligence, and command and control systems," and it "will drive the design, development, and fielding of autonomous and semi-autonomous capabilities and systems." The release itself does not enumerate a portfolio.

FACT — the vocabulary shifted live. Hegseth used the President's coinage: "Drone warfare supercharged by SI-enabled targeting is the biggest battlefield revolution in generations." Air & Space Forces Magazine glossed "SI" as "so-called super intelligence, President Donald Trump's favored term for AI." GlobalSecurity notes that the department's release and the Agincourt memorandum, as published, do not mention the separate "AI Force" the President announced on 2026-09-19, and that neither the President nor the Secretary has explained how the two relate.

COMPANY CLAIM — and it deserves to be labelled as such. The department's release calls this "the fastest peacetime operational shift in modern military history." That is a presentational framing, not a measurable finding, and this research does not verify it.

INTERPRETATION — the structural point. Notice what the After column does not say. It does not say the command will employ forces in its own theatre, and it does not say it gets a new domain. The memo's sentence — "define joint needs and integrate Service-provided forces in support of geographic combatant commanders" — describes a force provider and integrator on the pattern of U.S. Special Operations Command, not a new geographic combatant command. GlobalSecurity makes the same reading and notes this is the second time in 2026 that the organisation chart ran ahead of the appropriations.

Mechanically, the announcement sets up a two-phase construct: an interim authority that already has powers, and a statutory command that does not yet exist.

Phase 1 — now, under existing authority ("Project Agincourt", ~12 months).

1. DRPM-UxS leads. The 29 June 2026 acquisition office, reporting to the Deputy Secretary, is repurposed as the engine for standing up the command. It absorbs oversight of DIU, JIATF-401 and DAWG.
2. A CEO–COO pair runs it. This is the genuinely unusual mechanic. The memorandum assigns West the "Chief Executive Officer" title and Strasiser (a Navy SEAL senior chief, E-8) the "Chief Operating Officer" title — Hegseth's phrase: "a CEO-COO partnership." Breaking Defense and GlobalSecurity both note these are business titles with no counterpart in ordinary defence-acquisition vocabulary.
3. A 30-day POAM is the gate. Within 30 days of the memorandum, DRPM-UxS — "with the support of the Chairman of the Joint Chiefs of Staff" — must submit a plan of action and milestones for the Secretary's approval. The POAM must cover priority operational efforts, Service force contributions, force-generation responsibilities, command relationships, DRPM-UxS absorption, test-and-update support, resources, required authorities, milestones, responsible officials, initial operating capability, and "measures of adaptation and Service adoption." Where choices are unresolved, the POAM must present alternatives and a recommended decision.
4. The Joint Staff gets a counterpart. "Concurrently, the CJCS will designate a senior Joint Staff planning counterpart to support this effort."
5. The personnel track runs in parallel. The Secretaries of the Military Departments, coordinated by the Under Secretary of War for Personnel and Readiness, produce phased manning plans and MOS/rating proposals "building on existing specialties and identifying gaps."
6. Monthly accountability. DRPM-UxS reports monthly on "operational results, Service adoption, progress against the plan, and decisions required," escalating unresolved matters to the Deputy Secretary or the Secretary.

Phase 2 — the target state (on or about 2027-10-01).

• A 4-star functional combatant command that "will lead the Joint Force in unmanned and autonomous warfare."
• Authorities: "directed manpower, budget, acquisition authorities" plus "dedicated military career pathways for Officers and Enlisted personnel."
• Method: "test equipment and force designs with realistic opposition in demanding exercises and rugged environments, modify rapidly, and field what works."
• Role: define joint needs; integrate Service-provided forces; support the geographic combatant commanders.
• Stated output of the acquisition experiment: "a repeatable procurement system whose output is not a particular drone, but the ability to adapt in combat, on the right side of the cost exchange."

How it differs from the alternatives. GlobalSecurity's reference analysis draws the distinction that most coverage blurs: Special Operations Command buys its own equipment and carries its own budget because the 1986 Nunn-Cohen amendment wrote those powers into statute; Cyber Command's acquisition authority rests on a later statute; and the work of recruiting, organising, training and equipping forces is assigned by statute to the military departments. That is precisely why the memorandum says "collaborate directly with Congress." It also notes that a four-star headquarters could be created by presidential directive under 10 U.S.C. §161 and the Unified Command Plan — as Space Command was in 2019 — but that the manpower/budget/acquisition package is the part only an act of Congress can supply.

INTERPRETATION. Read the announcement as three separate bets, because they can fail independently: (a) the org chart — likely, because an executive can build it; (b) the statute — uncertain, because it needs both chambers and a Senate confirmation for the commander; (c) the money — the weakest link, see §12.

FACT — the Department gave the reason itself. Hegseth: "The autonomous and robotic revolution has enormous implications for equipment, doctrine, organization, and the assumptions that have governed warfare for decades."

INTERPRETATION — the institutional precondition for scale. Discovery's framing ("autonomous systems move from programme-level procurement into a dedicated command structure with its own budget, acquisition authority and acquisition-path authority") is right in direction but overstates the present. On announcement day the command had none of those three. It has an intention to acquire them, expressed in a memorandum that explicitly instructs its staff to go and lobby Congress. The honest version is: autonomous systems have moved from programme-level procurement into a command-shaped ambition with a 12-month interim and a legislative dependency. That is a real shift in the centre of gravity — the Secretary has put his own career weight behind a single accountable owner — but it is not yet a shift in authority.

INTERPRETATION — a second, largely closed demand pool for embodied AI. Two things follow. First, procurement volume and industrial-policy attention are now concentrated in one place, which is what makes a "Gauntlet"-style competitive fly-off model scalable as a policy, not just a pilot. Second, this demand pool is structurally separate from civilian commercial AI: different security regimes, different procurement pathways, different liability posture, and — per the Department's own framing — a closed, sovereign industrial base.

FACT — a live counter-pressure that this announcement did not resolve. Air & Space Forces Magazine, same day: "It is not yet clear how the new four-star command will be structured, who will lead it, or how it will work with the autonomy efforts already underway within the different services." GlobalSecurity records that the Air & Space Forces Association had formally asked the Senate Armed Services Committee to strike the comparable provision from its FY2027 bill in a 30 June 2026 letter, arguing it "would greatly complicate the development, acquisition and operational use of" unmanned systems across all services — an argument that is now aimed at a live proposal rather than a committee bill.

FACT — how big the demand pool is claimed to be. Reuters, same day, citing estimates: drones "are estimated to account for about 70% of Russian casualties." Hegseth cited the Drone Dominance programme as having "delivered thousands of robotic systems to the Middle East in support of Operation Epic Fury" (COMPANY CLAIM as to the delivery count and its effect).

FACT — where autonomy actually is in the fleet, as reported the same day. Air & Space Forces Magazine quotes Air Force Chief of Staff Gen. Ken Wilsbach saying true autonomous aircraft are "a few years off," while the FQ-42 and FQ-44 Collaborative Combat Aircraft are framed as autonomous in the sense that "you just send the aircraft off, and it does its mission because it's automated."

FACT — capability that was previously impractical is now someone's named deliverable.

1. A single joint requirements voice for drones, AI and C2. Before this, each service defined its own unmanned requirements. Air & Space Forces Magazine reports the new portfolio "will include drones, artificial intelligence, and command and control systems" — meaning a component like a targeting stack could now be specified once for the joint force rather than per-service.
2. A procurement experiment with a stated success metric that is not a product. The memorandum asks for "a repeatable procurement system whose output is not a particular drone, but the ability to adapt in combat, on the right side of the cost exchange." That is a testable, transferable claim.
3. Testing against realistic opposition as the gate for scale. "test equipment and force designs with realistic opposition in demanding exercises and rugged environments" — with Military Times reporting that exercises and, where possible, combat will decide what gets expanded.
4. Career paths for autonomy practitioners inside the services. The MOS/rating work creates a durable personnel pipeline and, with it, institutional memory — the thing Hegseth says the military lacks.
5. An actual decision right in 30 days. The POAM must produce named responsible officials, an initial operating capability definition, and recommendations on unresolved choices. That is a dated, checkable artefact.
6. A citable counterparty. For any company, a four-star command with its own acquisition authority is a different customer from a service programme office.

INTERPRETATION. The most under-sold item on that list is #5. Within about a month of the speech, the public record should contain a plan that names who owns what. Whatever else happens, that artefact will make the organisation concrete or reveal that it is not.

Institutional (the ones that decide whether this is real):

1. Will a "4-star functional combatant command" be established on 1 October 2027, and under what statute?
2. Does the final legislation grant directed manpower, budget and acquisition authority, or only the "limited acquisition authorities" a Senate staffer described in June?
3. Does the Senate's permissive version survive, does the House add its own, or does the FY2027 NDAA die and the command have to wait for FY2028?
4. Who is confirmed as the first commander, and does the Senate Armed Services Committee hold a confirmation hearing on a command that does not yet exist in statute?
5. What budget line does it get, and does it inherit the Defense Autonomous Warfare Group's request or a fraction of it?
6. What happens to the Service programmes it is meant to coordinate — are they realigned to the command, or does the command only define joint requirements?

Operational / technical: 7. What is the initial operating capability the POAM defines? 8. What is "realistic opposition" in the testing clause — red teams, opposing forces, or a specific exercise serial? 9. Does the command absorb JIATF-401 (defensive) and DAWG (offensive) into one integrator, and who arbitrates when defensive priorities conflict with offensive ones? 10. Does "Maven Smart System" become the mandated integration target for all autonomous systems under this portfolio, as an end user asserts?

Personnel and doctrine: 11. What MOS / rating frameworks are being created, in which services, and by when? (First observable signal: the Service proposals inside the 30-day POAM.) 12. Hegseth said the new command would create "a new generation of autonomous warfighters." What does that career actually look like — and does it attract the software talent the acquisition model assumes?

Politics and governance: 13. Does the "AI Force" the President announced on 19 September 2026 relate to this command? Neither the release nor the memorandum mentions it, and neither the President nor the Secretary has explained the relationship. 14. Where does Project Meridian (report due 2027-01-28) sit relative to AutoWarCom's first doctrine — does the command receive its mission from Meridian, or the reverse? 15. Does the department's legal name change this year? The announcement was issued under "Department of War" branding; every major outlet called it the Defense Department. 16. What is the observable evidence of success at the 30-day mark, and what would failure look like?

Absence of sourcing (recorded, not speculated): This research found no source claiming the command has already been staffed, funded, sited or given a commander. Any such claim appearing in circulation should be treated as unverified.

Dated, checkable, and sourced where it exists.

• By ~30 October 2026 — FACT (from the memorandum). DRPM-UxS, with CJCS support, submits a plan of action and milestones for the Secretary's approval. It must cover force contributions, force-generation responsibilities, command relationships, required authorities, resources, milestones, responsible officials, an initial operating capability definition, and "measures of adaptation and Service adoption." The CJCS designates a senior Joint Staff counterpart. The military departments submit phased manning plans and MOS/rating proposals.
• From October 2026 — FACT. DRPM-UxS reports monthly to the Secretary on operational results, Service adoption, progress and decisions required, escalating unresolved matters.
• Already in motion — FACT (DefenseScoop, 2026-10-02). Owen West on "leave of absence" from DIU to serve as DRPM-UxS CEO; Travis Metz acting DIU director, reporting to the Secretary as a principal staff assistant.
• Near-term — PREDICTION (low confidence on the actor, high on the mechanism). The four-star nomination is the single most likely next confirming event, and it will arrive before the statute — because a combatant command commander requires Senate confirmation regardless of the authorisation path, and the Secretary said a commander would be named "soon." Expect the nominee's hearing to be the first place the command's charter is argued in public detail.
• Through FY2027 — PREDICTION. The command's existence will be decided by the FY2027 defence authorisation process and then by appropriations. Treat the 2027-10-01 date as a target that requires two independent legislative victories, not a date.
• By 2027-01-28 — FACT (Breaking Defense). Project Meridian's report on the trajectory of future warfare and critical capability gaps is due. This is the document most likely to supply the command's first mission statement.
• Structural watch item — PREDICTION. If Service programmes are realigned to the new integrator, Service leadership resistance will surface in the POAM or in public comment from the affected services. The AFA's June 2026 letter is the template.
• What would falsify the story's central claim. If the FY2027 bill stalls, the October 2027 date slips and the entity remains an acquisition office with a four-star aspiration, then the correct headline two years out is "Pentagon created a drone acquisition office," not "created a command." That is the outcome to watch for, and it is entirely plausible.

The Pentagon's most consequential AI announcement of the week was not a model, a benchmark or a product. It was an org-chart decision — and the interesting part is that the announcement itself admits it is not yet a decision at all.

Two documents, read together, tell the whole story. The release says the Autonomous Warfare Command will exist "by October 1, 2027." The memorandum signed the same morning says the target date is October 1, 2027 and, in the same paragraph, orders its staff to "collaborate directly with Congress to secure necessary legislation." Between those two sentences sits the entire story: an executive creating an institution that only Congress can fully create.

That is the right way to read the whole AI-in-military complex right now. The velocity people keep citing as proof of decisiveness — an initiative in 2023, a task force in 2025, a working group, an acquisition office in June 2026, a theater command in April, a four-star command in September — is not evidence of a settled strategy. It is evidence of a strategy being renegotiated roughly monthly. GlobalSecurity's chronology is the most useful corrective here: the same concept has been re-containered at least six times in about a year, each time at a higher tier. That is escalation, but escalation is not the same as resolution.

The second thing worth saying plainly is that the bottleneck in military AI is institutional, not technical. Hegseth put it in one line — "The pace of war is changing faster than the process to support it" — and then acted on it by drawing a box on the org chart. Whether that works depends on things no speech can deliver: a statute, a budget line, a confirmed commander, and Services that are willing to hand over programmes. Two of those four were missing on announcement day.

And the third is a warning about vocabulary. In the space of two weeks the same administration announced an "AI Force" with an AI czar, rebranded the Department as the Department of War, pushed "super intelligence" as a replacement for "artificial intelligence," and created a four-star combatant command — while the department's own release and its own memorandum mention neither the AI Force nor the new terminology. Four institutional frames for one problem in a fortnight, with no stated relationship between them. Anyone trying to build a durable picture here should anchor on plain names, document marks and dates — "Autonomous Warfare Command," OSD010441-26/CM DO 13361-26, 2026-09-30, target 2027-10-01 — and treat every acronym and every superlative as unsettled.

The one-line version for a busy reader: on 2026-09-30 the Pentagon announced a four-star functional combatant command for autonomous warfare, with an interim team and a one-year deadline — and its own paperwork admits the command does not exist until Congress says so. The date to watch is ~30 October 2026, when the plan of action is due; the date that matters is whether Congress writes the statute.