Disruption through Intelligent Strategies, Counter Options, and Resilient Defenses (DISCORD)

Solicitation / RFP / RFQ

This solicitation, DARPA-PS-26-27, from the Defense Advanced Research Projects Agency (DARPA) Strategic Technology Office (STO) seeks innovative approaches for an Artificial Intelligence (AI)-native tactics engine. The program, named "Disruption through Intelligent Strategies, Counter Options, and Resilient Defenses (DISCORD)", aims to rapidly generate adaptable tactical options using live sensor data and high-fidelity simulation. The goal is to provide commanders with a diverse portfolio of strategies that can dynamically pivot in response to evolving conditions, enhancing strategic flexibility and enabling commanders to exercise tactical judgment in ways adversaries cannot predict.

• Development of an AI-native tactics engine.

• Integration of live sensor data and high-fidelity simulation.

• Rapid generation of tactical options.

• Continuous adaptation to changes in the operational environment.

• Discovery of a wide range of superior and diverse strategies.

• Program Duration: 24 months, organized into two 12-month phases.

• Anticipated Awards: Multiple awards are anticipated.

• Instrument Type: Other Transaction Authority (OTA) for Prototype.

• CUI Addendum Request Deadline: May 26, 2026, 3:00 PM ET.

• Abstracts Due Date: June 11, 2026, 3:00 PM ET.

• Final Questions Due Date: June 30, 2026, 3:00 PM ET.

• Written Proposals Due Date: July 9, 2026, 3:00 PM ET.

• Abstracts will be evaluated using criteria detailed in the CUI addendum.

• Proposals, if invited, will be evaluated using criteria detailed in the CUI addendum.

• Not specified in this document.

• This effort implements Ender's Foundry, a warfighting Pace-Setting Project (PSP).

• The program involves Controlled Unclassified Information (CUI). A formal request for the CUI addendum is required via Appendix A and email to DISCORD@darpa.mil.

• Publication and presentation of developed information will require release review through the DARPA DISTAR process; a fundamental research exemption will not be provided.

Other / Unclassified

This document is a presentation detailing the DARPA DISCORD program, which aims to develop an AI-native process for collapsing planning and execution to out-pace and out-think adaptive adversaries in real-time. It outlines the program's context within the Department of War AI Acceleration Strategy, its goals, technical areas, challenges, and approaches.

• Program Goals: DISCORD seeks to ensure the US stays ahead of AI-enabled adversaries by accelerating M&S environments, generating a tactics "factory," developing counter-UxS capabilities, and exploiting adversarial vulnerabilities.
• Technical Areas: The program is divided into three technical areas:
  • TA 1: Asymmetric Advantage Engine (focuses on reconstructing adversary dynamics using nonlinear time series techniques).
  • TA 2: Human-Machine Teaming (aims to create an AI that acts as a cognitive partner, improving commander decision-making speed and intent translation).
  • TA 3: Live Adaptation Experimentation (integrates TA 1 and TA 2 into real-world capabilities for validation).
• Schedule & Milestones: The presentation includes a program schedule with key milestones such as Integration Test (Month 3), Midterm Evaluations (Months 6 & 18), and Final Evaluations (Months 12 & 24).
• Metrics: Specific metrics are defined for each technical area, including Advantage Score, Architectural Insight, Speed of Command, Intent Translation Accuracy, Time to Effective Response, and Plan Resilience Under Fire, with targets for Phase 1 and Phase 2.
• Government Team: Details the DISCORD government team's role in steering, validating, and implementing scenarios, involving organizations like IDA and MITRE.