Pentagon’s $54.6B autonomy request hinges on bill Senate appropriators have ruled out

On September 9, at a National Defense Industrial Association conference on emerging technologies in Washington, the Army three-star who runs the Pentagon’s Defense Autonomous Warfare Group said out loud what budget officials had been saying more carefully for months. Lt. Gen. Steven Marks told the audience he did not know whether his organization would receive the $53.6 billion the administration had asked Congress to provide outside the regular budget, and that if nothing arrived, the “momentum” built since the group was stood up would fade. The prospect, he said, “truly concerns me.”

Three weeks later Defense Secretary Pete Hegseth stood at Marine Corps Base Quantico and announced that the department would build a four-star Autonomous Warfare Command, to stand up on October 1, 2027. This article follows the money between those two events. It shows that about 98 percent of the largest autonomy request in the department’s history sits in a single research line that depends on a reconciliation bill the Senate’s own Republican appropriators have said will not happen; that the public documents describe what the money would buy by category but not by system or quantity; and that the Pentagon has kept adding organizations on top of a funding base that, as of this week, is a continuing resolution.

From Replicator to one line in the research budget

The group’s lineage runs through Replicator, the initiative announced in August 2023 to field thousands of attritable uncrewed systems within two years. The Congressional Research Service’s running summary of the program records a $300 million reprogramming request in fiscal 2023, $200 million appropriated in fiscal 2024 and a $500 million request for fiscal 2025, the selection of systems from AeroVironment, Anduril and Performance Drone Works, and a former official’s assessment that only hundreds of systems had been fielded by the summer 2025 deadline. CRS also lists, as an issue for Congress, the difficulty lawmakers had in getting information about what was being bought.

In 2025 the new administration split the effort. Replicator’s counter-drone work went to Joint Interagency Task Force 401 under an August 2025 memo, according to CRS. The offensive side was renamed. Breaking Defense reported in December 2025 that Replicator lived on as the Defense Autonomous Warfare Group, or DAWG, with a focus on larger, longer-range one-way attack drones suited to the Pacific, while small first-person-view drones were handled by a separate Drone Dominance program. Defense One has described Marine Gen. Francis Donovan, now head of U.S. Southern Command, as a former leader of the group.

The fiscal 2027 request, released April 21, turned a small office into the department’s largest single autonomy account. The comptroller’s R-1 research exhibit shows program element 0604539D8Z, Defense Autonomous Warfare Group, at $225.9 million in fiscal 2026 and $54.6 billion in fiscal 2027: $1 billion in discretionary funds and $53.6 billion in mandatory funds. DefenseScoop calculated the increase at more than 24,000 percent.

What the documents say the money would buy

The most detailed public account is two pages in the department’s mandatory funding overview. Somewhat confusingly, it files the entire $53.6 billion under the heading “Drone Dominance” while assigning every dollar to the DAWG program element. It divides the sum five ways. The largest piece, $16.85 billion, is for “Asset Purchases and Industrial Demand Signal,” meaning uncrewed systems for air, surface, subsurface and ground use. Counter-unmanned systems take $14.4 billion, to extend protection to what the document calls “250+ sites.” Contested logistics takes $13.53 billion, for leased sites, modular container platforms and commercial repositioning of systems without military lift. Collaborative autonomy software receives $4.5 billion, and “force generation,” the training pipelines and exercises needed to produce operators and planners, $4.32 billion.

Two features of that breakdown deserve attention. First, less than a third of the request would buy uncrewed systems outright; more than half goes to defending bases and to moving and sustaining what is bought. The emphasis on logistics and force generation answers a criticism of Replicator that a Senate staffer put to DefenseScoop in June: that the earlier program bought assets without a coherent way to deliver them to combatant commanders. Second, the whole amount is requested as research, development, test and evaluation money in budget activity 04, advanced component development and prototypes, although the text speaks of “mass procurement.”

That is where the public record stops. The document names no systems, quantities, unit costs or delivery dates, and says additional information is held at a higher classification. At the April 21 budget briefing, Jules Hurst, then the acting comptroller, described DAWG as a “pathfinder” that tests autonomy software with companies, and Lt. Gen. Steven Whitney of the Joint Staff said the intent was to keep developing new capabilities in short cycles instead of buying one baseline in bulk. Those are descriptions of a method. How many airframes, boats or interceptors $53.6 billion represents is not publicly known, and whether an office funded at $226 million this year could obligate roughly 240 times that amount is a question no official has answered on the record.

Why reconciliation, and what the appropriators said about it

The department’s stated reason for putting the money in a reconciliation bill is flexibility. Officials told reporters in April that mandatory funds give more time to obligate and suit fast-changing technology such as DAWG and Golden Dome. The political reason was laid out when the budget topline appeared: a reconciliation bill needs only Republican votes in the Senate, which let the White House propose $350 billion in defense increases without negotiating matching nondefense spending. The 2025 reconciliation law had already delivered roughly $152 billion to the Pentagon that way.

The people who write defense spending bills objected from the start, and not only Democrats. The House Appropriations Committee’s report on its fiscal 2027 bill, submitted by Rep. Ken Calvert, the Republican who chairs its defense subcommittee, calls reconciliation “uncertain, unaligned with the annual appropriations cycle” and outside the committee’s jurisdiction, and says of the two-track request: “This approach is risky and uncoordinated.” The same report complains that justification exhibits for last year’s mandatory money were incomplete or absent, which it says makes oversight harder. In the Senate, Angus King called reconciliation money “essentially a slush fund” at an April 30 hearing where Hegseth acknowledged that $26 billion of the 2025 money was on contract after ten months.

Then the vehicle itself stalled. On June 9, Sen. Mitch McConnell, who chairs the defense appropriations subcommittee, said a third reconciliation bill was “not an option,” and full committee chair Susan Collins agreed. House Republicans in July advanced a version with $60 billion for defense, which Breaking Defense reported tracked the administration’s June war supplemental; that request contained $2.4 billion for drones and $5.1 billion for cybersecurity and autonomy. Even if every one of those dollars went to DAWG, which nothing indicates, it would be a small fraction of the group’s request. By late August, Breaking Defense reported, the Senate majority leader had not found the votes to move it and Senate appropriators had released no defense bill at all.

What DAWG has in hand is therefore modest. The House committee’s funding table recommends exactly the $1 billion discretionary request, and the report “supports the investment strategies and budget requests” for DAWG, JIATF 401 and the Defense Innovation Unit. It also adds a new general provision, section 8141, letting DAWG and JIATF 401 funds be spent across research, procurement and operations, including “initial acquisition of end-items for operational use.” But that bill has not reached the House floor. The government is operating under a continuing resolution through December 11, which holds programs to fiscal 2026 levels, and Breaking Defense counted DAWG among the initiatives likely to be scaled back without a new budget.

A command structure built ahead of the appropriation

While the money stood still, the organization chart changed four times. On April 22, Southern Command created its own Southcom Autonomous Warfare Command. A week later Hegseth told the House Armed Services Committee, “We will shortly announce a sub-unified command for autonomous warfare.” A sub-unified command sits beneath a combatant command and is established with the secretary’s approval; DefenseScoop could not establish whether he meant the Southcom organization, and a Pentagon official declined to clarify. No separate sub-unified command has been announced since.

In June the Senate Armed Services Committee went further than the secretary had. Its summary of the fiscal 2027 authorization bill, approved 18 to 9, describes a provision “permitting the establishment of the Robotic and Autonomous Systems Combatant Command.” The verb matters: the committee’s own text permits a command and does not order one. That bill has not reached the Senate floor.

On June 29, Hegseth signed a memo creating a Direct Reporting Portfolio Manager for unmanned systems, reporting to the deputy secretary, and, as Breaking Defense and DefenseScoop both reported, placed DAWG and JIATF 401 beneath it as deputy offices. Then came Quantico. Hegseth described the planned Autonomous Warfare Command as “a new four-star combatant command with service-like authorities” and renamed the three-month-old portfolio office Project Agincourt, to be led by Owen West, who is taking leave from the Defense Innovation Unit, with a Navy SEAL senior chief. Its job is to prototype a model that joins buying and operating before the command exists.

None of the reporting on the announcement identified a budget for the command, and Breaking Defense noted that giving a combatant command acquisition authority is almost certain to require legislation. The Pentagon memos were not accessible for this article, so their terms are known here only through those accounts. Where DAWG ends up inside a future command, and whether Marks’s group survives as a distinct office, has not been stated publicly.

Who gains, and the case that the gap is smaller than it looks

Because no system is named, any list of corporate beneficiaries is inference. What can be documented is the adjacent market. Under the $1 billion Drone Dominance competition, Defense News reported orders of 14,000 drones from Neros and 8,000 from Perennial Autonomy. Shield AI said in May it had been chosen to put its Hivemind software on the roughly $35,000 LUCAS one-way attack drone, according to Breaking Defense, which did not report the contract’s value. Firms that build one-way attack aircraft, small boats, interceptors and autonomy software are the obvious candidates for the five budget categories, and several have invested ahead of orders. Anduril officials said in September that the company could go no further on a separate Air Force drone production contract without fiscal 2027 procurement dollars, which illustrates how private capital behaves when appropriations slip.

The strongest objection to reading all this as a crisis is that reconciliation was never the only source of autonomy money, and may have been the only route to money on this scale. The department’s budget overview counts $54.0 billion for autonomous and remotely operated systems, of which $39.2 billion is tied to the mandatory Drone Dominance request; the remaining $14.8 billion is spread across other accounts, and $6.2 billion of the $20.6 billion counter-drone total likewise lies outside it. A base of $1 billion for DAWG is still more than four times this year’s level, and House appropriators met it in full and added spending flexibility. Defenders of the approach, among them Sen. John Kennedy, argue that reconciliation is the only way the military gets large sums, and the Pentagon can point to having obligated $142 billion of last year’s $152 billion before the October 1 deadline.

Against that stand two facts. The base-budget figure funds an office, not the mass the department says deterrence requires, and the choice of vehicle was the administration’s. Analysts drew the conclusion early: Todd Harrison of the American Enterprise Institute said in July, as quoted by Breaking Defense, that hopes for a $1.5 trillion budget were “effectively dead.”

What is still undecided, and who decides it

The department has described a fallback in general terms. Under Secretary Emil Michael said in June that if reconciliation failed the Pentagon would make trade-offs against “exquisite” weapons to protect drone purchases. Hurst, in remarks Breaking Defense reported on September 10, said the department could go back to appropriators before they finish the fiscal 2027 bill and ask which priorities should be funded now and which can wait for fiscal 2028, adding that it was too early to name them. No such list has been made public.

The decisions now belong to identifiable people. Collins, McConnell and their House counterparts must decide, in a lame-duck session after the November elections or in the new year, how much of the mandatory request to absorb into a discretionary bill whose topline is itself unsettled. Hurst and Deputy Secretary Stephen Feinberg must decide what to offer up in exchange, and whether to give Congress the system-level justification the House committee says it did not get last time. The armed services committees must decide in conference whether a four-star autonomy command is permitted, required or deferred, and with what acquisition powers. Until those choices are made, Marks’s group is funded at its fiscal 2026 rate through December 11.