DARPA DICE: The Doctrine Gap in Pentagon's $54B | Drone Intelligence
AI Analysis
DARPA is initiating the DICE program to address the operator bottleneck in drone warfare through decentralized AI and autonomous systems. A parallel RFI seeks to embed computing directly into drone hardware, enabling 'constellation scale' autonomous operations. However, funding for doctrine, training, and force design lags significantly behind technology procurement, potentially hindering effective implementation.
Key Takeaways
- DARPA’s DICE program (DARPA-SN-26-72) focuses on decentralized AI for drone coordination and control, aiming for scalable, resilient, and long-duration missions.
- A related RFI (DARPA-SN-26-76) explores embedding compute power directly into drone actuators and sensors, reducing reliance on centralized processing.
- The Pentagon’s FY27 autonomous warfare budget allocates only ~2% to doctrine, training, and force design despite a $54 billion total request.
- DICE emphasizes 'controlled emergence' – complex behaviors arising from simple, predictable local rules within the AI agents.
- Proposers Day for DICE is scheduled for May 29, 2026, in Arlington, Virginia, with initial responses due May 19, 2026.
Why It Matters
The DICE program represents a significant shift towards more autonomous drone warfare, potentially offering a decisive advantage. The lack of corresponding investment in doctrine and training creates a critical risk of fielding advanced technology without the ability to effectively utilize it. This imbalance could limit the program’s potential and create vulnerabilities.
DARPA DICE: The Doctrine Gap in Pentagon's $54B | Drone Intelligence
SIGNAL DOSSIER/VOL. 02-N
The Doctrine Gap: DARPA's DICE Programme and the $54 Billion Question.
STRATEGIC INTELLIGENCE UNIT, Published Q2 2026
DRONE INTELLIGENCE EDITORIAL TEAM|Q2 2026|10 PRIMARY SOURCES
EXECUTIVE SIGNAL
DARPA's Decentralized Artificial Intelligence through Controlled Emergence (DICE) programme is the most explicit public statement to date of how the Pentagon intends to dissolve the operator-bottleneck problem that has constrained American drone warfare since the Predator era. The Special Notice (DARPA-SN-26-72) was published 28 April 2026, with a response deadline of 19 May 2026 and a Proposers Day scheduled for 29 May 2026 in Arlington, Virginia. Programme manager Susmit Jha at DARPA's Information Innovation Office is seeking theory and algorithms for decentralised coordination and local inference control: a scalable, adaptive, resilient collective of heterogeneous AI agents executing sustained long-time-horizon missions in contested environments while remaining under human control.
A companion RFI from DARPA's Microsystems Technology Office, Materials for Physical Compute in Untethered Robotics (DARPA-SN-26-76, response deadline 27 May 2026), completes the technical envelope by asking industry how to embed compute directly into actuators and sensors rather than through centralised processors. Together the two solicitations articulate a coherent technical vision for autonomous warfare at constellation scale.
The institutional response, however, is lagging the technical procurement. According to a commentary by retired Gen. David Petraeus and Isaac Flanagan published in The Hill, less than 2 percent of the Pentagon's $54 billion FY27 autonomous warfare request is allocated to the doctrine, training, and force design that would convert the technology into capability. The architecture is being procured faster than the institutional framework to govern it.
SIGNAL 01: DICE AND THE DECENTRALISED COORDINATION SPEC
DARPA's Information Innovation Office published Special Notice DARPA-SN-26-72 on 28 April 2026, formally inviting industry concepts for the Decentralized Artificial Intelligence through Controlled Emergence programme. The Special Notice response deadline closes today, 19 May 2026, with a Proposers Day scheduled for 29 May 2026 at the Executive Conference Center in Arlington, Virginia. Programme manager Susmit Jha, in his published programme synopsis, frames the technical objective in three sentences: develop the theory and algorithms for decentralised coordination and local inference control; enable a scalable, adaptive, resilient collective of heterogeneous AI agents; sustain long-time-horizon missions in contested environments while remaining under human control.
The conceptual centre of the programme is what DARPA terms controlled emergence: complex system-level behaviour arising from simple local rules, but constrained and predictable. E