Decentralized Artificial Intelligence through Controlled Emergence (DICE)
🏛 DARPA - Information Processing Technologies Office
Can you apply?
This grant is for organizations and researchers developing decentralized AI systems and multi-agent coordination technologies. Federal contractors, businesses, nonprofits, universities, and research institutions can apply. There are no geographic restrictions; this is open nationwide.
The program funds development of theory, algorithms, and systems for autonomous AI agents operating in contested environments. Projects must demonstrate scalability and resilience while maintaining human control and predictability.
Applicants must be "responsible sources" capable of meeting government needs. Submit proposals through the DARPA BAA process. See the BAA for full eligibility details and technical requirements.
Program description
The DICE program seeks to develop the theory and algorithms for decentralized coordination and local inference control to enable a scalable, adaptive, and resilient collective of heterogeneous AI agents that can autonomously execute sustained long-time-horizon missions in contested environments while remaining under human control. In contrast to small-scale, rigid, and fragile centralized orchestration or the high-risk unpredictable nature of ad hoc compositions of AI agents, DICE aims to harness the scalability and adaptability of self-organizing systems while minimizing risks and ensuring that the collective behavior remains predictable and aligned with intended outcomes. This approach mirrors the principles of decentralized self-organization that underpin the internet’s own scalability and resilience, where robust global behavior emerges from simple, local rules.
Who can apply
Eligible applicants
How to apply
Application links
Required documents
- DARPA BAA submission template (per current BAA)
- Project Narrative (technical approach and innovation)
- Budget and Budget Justification
- Organizational capability documentation
- Resumes of key personnel
- Letters of support (if applicable)
Program contact
- 👤 DARPA - Information Processing Technologies Office
- 📧 DICE@darpa.mil
- 📞 571-218-4375
Funding track record
Recent awards under CFDA 12.910 from the last 3 years — real organizations that won funding through this same program.
Top 10 Largest Recent Awards
-
$44,089,743
-
$36,151,284
-
$33,938,157
-
$29,592,309
-
$28,734,050
-
$26,975,142
-
$26,089,306
-
$25,544,440
-
$25,503,840
-
$23,343,160
Top States by Funding
- CA 22 awards $252.2M
- MA 15 awards $149.7M
- MI 4 awards $75.0M
- NY 7 awards $58.9M
- NC 2 awards $52.5M
Source: USAspending.gov — federal spending transparency. Data covers last 3 years.
Funding history
Annual funding for this program — Federal obligations (CFDA 12.910). How funding has trended year over year.
| 2024 | $465,182,503 | |
| 2025 | $322,093,385 | |
| 2026 est. | $352,000,000 |
FAQ
Who is eligible to apply for DICE funding?
All responsible sources capable of meeting government needs may apply. This includes universities, nonprofits, small businesses, and large contractors.
What types of projects does DICE fund?
Projects developing decentralized AI coordination, multi-agent algorithms, and autonomous systems that maintain human control and predictability in contested environments.
When is the deadline?
The deadline is August 25, 2026. Check the BAA for any rolling submission windows or pre-proposal requirements.
What makes a DICE proposal competitive?
Strong proposals address scalability, adaptability, resilience, and human oversight. Demonstrate clear technical innovation in decentralized coordination and local inference control.
What can be included in my budget?
DARPA research budgets typically include personnel, equipment, travel, and contractor costs. No cost-sharing is required. Review the BAA for specific allowable categories.
💡 Tips for applicants
- Read the full BAA carefully; DARPA programs have specific technical evaluation criteria and page limits not always obvious from summaries.
- Emphasize human control and predictability in your approach; DICE explicitly requires alignment with intended outcomes and minimized risk.
- Include a clear technical approach section explaining your algorithms and theoretical contributions to decentralized coordination.
- Address scalability and resilience directly; show how your system works with heterogeneous agents over long time horizons.
- Plan time for government review; DARPA awards are competitive and evaluation timelines are often longer than typical grants.
⚠️ Common mistakes
Proposals that focus only on centralized orchestration or ignore human control safeguards get rejected. Weak technical depth in multi-agent algorithms or coordination theory signals unreadiness. Vague descriptions of how the system scales or remains resilient in contested environments reduce competitiveness.
Similar grants
- OPEN Rural Community Health Integration2026 — New York State Department of Health
- OPEN FY26 Bureau of Land Management Rangeland Resource Management – Bureau wide — Bureau of Land Management
- OPEN FY26 Bureau of Land Management Cultural and Paleontological Resource Management – Bureau wide — Bureau of Land Management
- OPEN FY26 Bureau of Land Management Youth Conservation Corps – Bureau wide — Bureau of Land Management
- OPEN Infertility Training Center — Office of the Assistant Secretary for Health