Warfighting Acquisition: What Project Agincourt's Open Letter Demands from Defense Companies

BLUF — Bottom Line Up Front

On October 21, Project Agincourt co-leads Owen West and Senior Chief Max Strasiser released an open letter to the Joint Force and defense industry that defines one key measure of program success: how quickly an operational problem receives a fielded solution that can be evaluated for scale. Systems will be judged by competitions, exercises, and combat performance — not documentation milestones. A week earlier, on October 14, Hegseth addressed the Project Agincourt panel at the Association of the U.S. Army's annual conference and announced a $50 billion one-time reconciliation request; the FY2027 budget request separately includes $53.6 billion for the Drone Dominance initiative with explicit line-item priorities. Acting Army Secretary Adam Telle had stated two days before the panel that open architecture is now non-negotiable for all future Army equipment purchases. For any company working in autonomous systems, the budget tells you where the money is, the letter tells you how you'll be evaluated, and the open architecture mandate defines the technical baseline. The Agincourt team has tied the model to existing DRPM-UxS authority; the letter does not specify when the first structured evaluations begin.

What the Open Letter Actually Says

The letter's diagnosis comes early:

"Low-cost drones have changed the character, speed and economics of war."

"Since the Cold War, American power projection has rested on a small number of sophisticated, expensive weapons. That model has been disrupted."

The gap the letter is closing is between that reality and the acquisition infrastructure built for the previous model. West and Strasiser state Project Agincourt has two missions: stand up AUTOWARCOM by October 1, 2027, and — continuing DRPM-UxS's work — "pioneering a new acquisition model that pairs autonomous and robotics companies directly with operational forces at the edge. The goal is not a particular drone. It is a system built to adapt at scale."

The phrase that will matter most to contractors is buried mid-letter: "Evidence will decide outcomes. Competitions, exercises and combat performance will determine which systems scale and which are cut. The power of the purse demands strict stewardship."

This is the evaluation standard. Not requirements documents. Not CDR/PDR milestones. Not Milestone B. Competition performance, exercise results, and combat record.

Two other lines define the character of the model. "Urgency does not excuse imprecision" — which reads as a guard against the trajectory of fast-fielding programs that expanded quickly and delivered expensive, underperforming capabilities. And "We will not replace one bureaucracy with another" — which signals that Agincourt is not an organizational expansion play. The letter calls on Agincourt to "quickly remove the barriers that constrain the initiative of our service members by fusing operations and development."

The evaluation language for what passes muster: "These new systems will not be declared ready on paper. Combatant commanders must be able to wield them in combat. The new model must strengthen the Services while empowering each theater build its own mix."

That last clause matters: the model doesn't require a single solution. It requires each combatant commander to be able to configure an appropriate mix. Pacific requirements differ from European requirements. Vendors with rigid platform configurations don't fit this model; vendors who have designed for configuration flexibility do.

The Shift in What "Program Success" Means

Prior acquisition models measured success against development milestones. Contract award, critical design review, initial operational capability, a program of record designation. Those checkboxes were designed for high-complexity, long-cycle platforms where specification compliance and cost control were the dominant risks.

The Agincourt model redefines program success as operational problem resolution. The feedback loop runs from warfighter to company and back, measured in operational cycles — not development phases.

This creates a new primary business-development variable: access to operational units before contracts are structured. In the previous model, demonstration was a marketing activity that happened after a company won a contract. In Agincourt's model, demonstration in realistic conditions *is* the qualification step.

The letter's framing of combatant commander flexibility reinforces this. West and Strasiser describe a future where commanders configure different mixes based on theater — the letter is direct: "The ideal Pacific portfolio is not right for Europe." INDOPACOM, EUCOM, and CENTCOM will potentially operate different combinations. Companies that position for one theater's requirements may or may not carry over to another. The unit of evaluation is operational performance per theater, not universal technical certification.

The Budget Architecture

Hegseth addressed AUSA's Project Agincourt panel on October 14 and tied the program to specific dollar figures. The $50 billion reconciliation request is described as a one-time allocation — not a recurring annual appropriation — intended to scale production from thousands to hundreds of thousands to millions. Hegseth's framing: "That's how you get from 1,000 drones to 100,000 to a million."

The FY2027 Drone Dominance budget request sits at $53.6 billion, broken out as:

$16.9 billion — unmanned-system procurement
$14.4 billion — counter-drone systems
$13.5 billion — contested logistics
$4.5 billion — collaborative autonomy
$4.3 billion — institutional systems and personnel

The relative weighting is informative. Counter-drone nearly equals the unmanned procurement line — confirming that both offense and defense are simultaneous priorities, not sequential. Contested logistics, covering uncrewed resupply and distribution under denied conditions, is larger than the collaborative autonomy and institutional lines combined.

Two caveats. First, as of October 14, the reconciliation request had no clear path to Congressional approval. Pentagon Comptroller Jay Hurst has noted a contingency: if reconciliation fails, the request could shift into the base budget, which would likely require cuts in other areas. Under Secretary Emil Michael separately said the Pentagon would weigh giving up pricier conventional weapon systems to keep funding autonomous systems. The budget proposal establishes the *architecture of intent*; appropriations are still pending. Second, Defense News noted that the $53.6 billion figure and the $50 billion reconciliation request are not confirmed to be the same money, and the relationship between the two is not fully established in public DoD budget documentation.

These caveats don't change what the budget tells you about priorities. They do mean that new awards dependent on reconciliation or FY2027 appropriations may slip until the funding picture resolves.

The Open Architecture Mandate

At the AUSA annual meeting's opening ceremony on October 12, Acting Army Secretary Adam Telle stated that open architecture is now a prerequisite for all future Army equipment purchases: "Everything we buy from here on out will have open architecture. It's non-negotiable," per MeriTalk and Military Embedded Systems coverage of the event.

This connects directly to the Agincourt model's requirement for theater-configurable mixes. If combatant commanders are selecting capabilities from different vendors, those capabilities must interoperate. Open architecture is the technical precondition for the operational flexibility Agincourt describes.

Operation Jailbreak, the Army's effort to connect legacy systems for interoperability without requiring complete redevelopment, provides one indication of what the open architecture mandate looks like in practice.

For vendors, Telle's open architecture mandate means proprietary integration designs are now a compliance risk for new Army equipment purchases. Agincourt has not yet published definitive interface standards or technical architecture, but open data exchange is the stated direction: systems will be expected to interoperate, not operate as closed stacks.

For companies that have treated proprietary integration stacks as a competitive moat, this is a direct challenge. For companies whose architectures were built around open data exchange, it is a technical alignment with stated DoD requirements.

What Has Not Been Resolved

The October 11 analysis on this site examined the Congressional constraints in detail — specifically, the Senate Armed Services Committee's S. 4784 language that would permit rather than require a four-star Robotic and Autonomous Systems Combatant Command, and the unresolved legislative gap between the command's October 1, 2027 target and statutory authority for service-like acquisition powers.

Two things have clarified since then, and one has not.

Clarified: Hegseth's AUSA appearance made explicit that AUTOWARCOM's acquisition ambitions are tied to appropriated dollars that haven't been allocated. The command's service-like authorities require both authorization legislation and appropriated funding. Neither is in place. Comptroller Hurst's contingency about shifting the request into the base budget underscores that the funding path has options, but each option requires its own legislative negotiation.

Clarified: The open letter ties the model's authority to the existing DRPM-UxS mission, which means the acquisition framework has organizational standing regardless of whether Congress creates AUTOWARCOM. Companies that wait for the command's statutory establishment before engaging are waiting past the window when requirements are being shaped.

Unresolved: The final acquisition authority structure. Whether AUTOWARCOM receives authorities analogous to SOCOM's 10 U.S.C. § 167 — its own program executive offices, budget title, and direct procurement authority — or a narrower coordination authority with interoperability standards and shared T&E depends on the conference outcome between House and Senate NDAA positions. The letter doesn't resolve this; it describes a model that works under either scenario, which is itself informative about the design intent.

Where to Engage

The open letter's evaluation criteria are more specific than general guidance about building relationships.

Competitions and exercises are the primary qualification channel. The letter names competitions, exercises, and combat performance as what determines which systems scale. Structured exercises — such as Scarlet Dragon and Project Sandhills — are the kind of venue where this evaluation model operates. Demonstration in realistic conditions is a prerequisite for serious contract consideration, not a showcase activity after requirements are set.

Theater-specific positioning matters. Combatant commanders are being given flexibility to configure their own mixes based on operational environment. Pacific and European requirements differ. Vendors who understand which theaters they're most suited for, and who have mapped their capabilities to those specific operational requirements, are better positioned for the commander-driven evaluation model than vendors pitching universal solutions.

West's organizational role defines the access point. DRPM-UxS is the home of the model's current implementation. West took a leave of absence from DIU to serve as CEO of Project Agincourt and DRPM-UxS (Travis Metz is acting DIU director); his leadership in Agincourt makes DRPM-UxS the direct access point. Engaging now — before the command structure is finalized — is the window when requirements are being shaped, not just implemented.

Open architecture is a threshold, not a differentiator. Telle's open architecture mandate defines the technical direction: open data exchange must be designed in before engagement with operational units, not patched in after contract award.

The Agincourt model creates a specific verification problem the defense industry has not consistently solved: how do you demonstrate that an autonomous system behaves reliably in a degraded, adversarial, or contested environment — not as a controlled test result, but as an operationally reproducible claim that a combatant commander can stake operational decisions on? That is the "evidence will decide outcomes" requirement, operationalized. The core of the problem is behavioral confidence under varied conditions: when does a system's output reflect reliable capability, and when does it reflect configuration or environmental overfitting? Spartan X is developing Arbiter, an AI assurance platform aimed at surfacing AI behavior that falls outside expected parameters across varied inputs — the class of problem that Agincourt's evidence standard, applied to AI-enabled systems, will need to address. The verification gap the letter identifies is one the field is still building the tools to close.

What the Open Letter Actually Says

The Shift in What "Program Success" Means

The Budget Architecture

The Open Architecture Mandate

What Has Not Been Resolved

Where to Engage

Sources

More from Signal

AutoWarCom and the Acquisition Question: What 'Service-Like Authorities' Actually Mean for Autonomous Systems

Army TITAN: The Integration and Support Behind an AI-Enabled Ground Station

Operators in the Development Loop: What CCA Is Changing About Acquisition

Ready to Solve Hard Problems?