1,000+ Opportunities
Find the right grant
Search federal, foundation, and corporate grants with AI — or browse by agency, topic, and state.
TRACTOR (Trustworthy Reasoning for AI Cybersecurity Tasks, Operations, and Resilience) Program is sponsored by Defense Advanced Research Projects Agency (DARPA). Funds research at the intersection of AI safety and cybersecurity, developing AI systems that can perform cybersecurity tasks while maintaining verifiable safety properties.
Awards support work on formal verification of AI security tools, adversarial robustness in cyber AI, and trustworthy autonomous cyber operations.
Get alerted about grants like this
Save a search for “Defense Advanced Research Projects Agency (DARPA)” or related topics and get emailed when new opportunities appear.
Search similar grants →Extracted from the official opportunity page/RFP to help you evaluate fit faster.
DARPA TRACTOR Program Grant (Up to $3M per award · Apply Now) | AI Safety Directory Dernière mise à jour : March 29, 2026 Montant Up to $3M per award Date limite June 15, 2026 Portée géographique United States The DARPA Trustworthy Reasoning for AI Cybersecurity Tasks, Operations, and Resilience (TRACTOR) program funds research on developing AI systems that can perform cybersecurity tasks while maintaining verifiable safety properties.
The program addresses the dual challenge of using AI to enhance cyber defense while ensuring AI-driven security tools do not introduce new vulnerabilities. Awards support work on formal verification of AI security tools, adversarial robustness in cyber AI, and trustworthy autonomous cyber operations.
robustness evaluation secure deployment threat modeling red teaming DARPA's TRACTOR (Trustworthy Reasoning for AI Cybersecurity Tasks, Operations, and Resilience) program funds research at the intersection of AI safety and cybersecurity, addressing the dual challenge of using AI to enhance cyber defense while ensuring that AI-driven security tools do not introduce new vulnerabilities.
Awards of up to three million dollars support research on formal verification of AI security tools, adversarial robustness in cyber AI, and trustworthy autonomous cyber operations. The program recognizes that AI systems used for cybersecurity must meet exceptionally high safety standards.
TRACTOR's focus on cybersecurity applications of AI safety is particularly relevant as AI-driven security tools become increasingly prevalent in both military and civilian contexts. The program addresses fundamental questions about how to trust AI systems that operate in adversarial environments where attackers actively try to deceive or manipulate them.
Research funded by TRACTOR develops methods for verifying that AI security tools behave correctly even under adversarial pressure, for detecting when AI systems are being manipulated, and for building resilient AI architectures that maintain safe operation despite active attacks. TRACTOR proposals are submitted through DARPA's standard procurement process via SAM. gov. The current deadline is June 15, 2026.
Proposals should describe the technical approach to building trustworthy AI for cybersecurity, the safety guarantees the system would provide, and the team's qualifications in both AI and cybersecurity. DARPA expects detailed technical plans with clear milestones and demonstration scenarios relevant to cyber operations. The proposal process involves submitting technical, management, and cost volumes.
DARPA values teams that combine expertise in machine learning, formal methods, and cybersecurity. Both academic institutions and defense industry partners are eligible, and teaming arrangements that combine academic research with operational cybersecurity experience are encouraged. Prospective applicants should review the TRACTOR Broad Agency Announcement carefully for specific technical requirements.
What Makes a Strong Application Strong TRACTOR proposals demonstrate deep expertise at the intersection of AI and cybersecurity, with credible approaches to building AI systems that maintain verifiable safety properties in adversarial environments. DARPA values proposals that address real operational cybersecurity scenarios and that propose formal methods for verifying AI behavior under adversarial pressure.
Experience with red-teaming, penetration testing, or formal verification of security properties strengthens proposals. The most competitive proposals tackle the fundamental challenge of trusting AI in environments where adversaries actively try to deceive the system.
Research that develops novel approaches to adversarial robustness, intrusion-tolerant AI architectures, or formal verification of AI security decision-making is particularly valued. Proposals should include clear demonstration scenarios that illustrate how the proposed approach would perform in realistic cyber operations contexts. Frequently Asked Questions What is the TRACTOR program's relationship to other DARPA AI programs?
TRACTOR specifically targets the intersection of AI safety and cybersecurity, complementing other DARPA programs focused on general AI assurance or AI capabilities. TRACTOR addresses the unique challenges of deploying AI in adversarial cyber environments where safety and security are closely intertwined. Is cybersecurity experience required for TRACTOR proposals?
Cybersecurity expertise is strongly valued but can be provided through teaming arrangements. Proposals that combine AI safety researchers with cybersecurity practitioners are competitive. DARPA encourages collaboration between these communities to address the program's interdisciplinary challenges.
What is the TRACTOR proposal deadline? The current proposal deadline is June 15, 2026. Awards are up to three million dollars.
Prospective applicants should review the Broad Agency Announcement on SAM. gov for detailed requirements and begin preparing proposals well in advance of the deadline. DARPA AI Assurance Programs Defense research contracts for AI assurance, testing, and evaluation in high-stakes environments.
DARPA program funding verifiable, explainable AI systems through compositional learning and reasoning. Behavioral AI platform protecting against AI-generated phishing, business email compromise, and social engineering. AI red-teaming company specializing in adversarial attacks, jailbreaks, and continuous LLM security testing.
Adversarial Robustness Toolbox Open-source library by IBM Research for defending ML models against adversarial attacks. Security testing and red-teaming platform for AI agents and multi-agent system vulnerabilities. Enterprise adversarial ML platform for automated red-teaming and security testing of AI systems.
Adversarial Robustness Evaluation Test Automated evaluation framework for assessing RAG system reliability, faithfulness, and context relevance. Adversarial Robustness Toolbox IBM's comprehensive library for defending ML models against adversarial attacks across multiple frameworks. Open-source security scanner for agentic AI workflows with OWASP mapping and prompt hardening.
Blue Teaming en sécurité de l'IA : stratégie, outils et bonnes pratiques Un guide complet sur le blue teaming en IA — la fonction d'opérations défensives pour la surveillance, la détection et la réponse aux menaces de sécurité contre les systèmes d'IA.
Red Teaming de l'IA : le guide complet pour tester les systèmes d'IA (2026) Un guide complet sur le red teaming de l'IA — méthodologies, outils, cadres de référence et études de cas pour tester systématiquement les systèmes d'IA afin d'identifier les vulnérabilités et les...
Détection de deepfakes : technologies, outils et bonnes pratiques (2026) Un guide complet sur la détection des deepfakes — comment fonctionnent les deepfakes, les technologies de détection, les outils disponibles, le cadre juridique et les bonnes pratiques pour les... Recevez les mises à jour hebdomadaires sur la sécurité et la sûreté de l'IA Nouveaux outils, cadres réglementaires et recherches livrés dans votre boîte de réception.
← Retour à toutes les bourses
Based on current listing details, eligibility includes: US research institutions, Defense contractors, University research labs. Cybersecurity expertise is strongly valued but can be provided through teaming arrangements. Applicants should confirm final requirements in the official notice before submission.
Current published award information indicates Up to $3,000,000 per award Always verify allowable costs, matching requirements, and funding caps directly in the sponsor documentation.
The current target date is June 15, 2026. Build your timeline backwards from this date to cover registrations, approvals, attachments, and final submission checks.
Federal grant success rates typically range from 10-30%, varying by agency and program. Build a strong proposal with clear objectives, measurable outcomes, and a well-justified budget to improve your chances.
Requirements vary by sponsor, but typically include a project narrative, budget justification, organizational capability statement, and key personnel CVs. Check the official notice for the complete list of required attachments.
Yes — AI tools like Granted can help research funders, draft proposal sections, and check compliance. However, always review and customize AI-generated content to reflect your organization's unique strengths and the specific requirements of the solicitation.
Review timelines vary by funder. Federal agencies typically take 3-6 months from submission to award notification. Foundation grants may be faster, often 1-3 months. Check the program's timeline in the official solicitation for specific dates.
Many federal programs offer multi-year funding or allow competitive renewals. Check the official solicitation for continuation and renewal policies. Non-competing continuation applications are common for multi-year awards.
Past winners and funding trends for this program
SBIR/STTR Programs (Defense Health Agency) is sponsored by Department of Defense (DOD) - Defense Health Agency (DHA). The DHA SBIR/STTR Programs fund biomedical and health-focused technologies that enhance medical readiness, clinical care delivery, force health protection, operational medicine, and military healthcare modernization. Priority research domains include digital health systems, AI-enabled triage, and physiological analytics.
Broad Agency Announcement (BAA) Call N0001425SBC03 For Office of Naval Research (ONR) Global Opportunity: GlobalX Innovation Joint Challenge: AI for Advancing Maritime Security is sponsored by Office of Naval Research (ONR) Global. This BAA Call seeks proposals for the GlobalX Innovation Joint Challenge: AI for Advancing Maritime Security. It funds the development of artificial intelligence solutions for maritime security applications, focusing on innovative AI algorithms for challenging maritime scene perception scenarios using real-world or synthetic data from UxV platforms. The program aims to accelerate the traditional knowledge generation cycle, leading to revolutionary dual-use capability for the U.S. Navy and Marine Corps and the commercial marketplace. White papers are highly encouraged and due May 23, 2025, with full proposals due June 23, 2025.
Operation Stonegarden (OPSG) is a federal grant program administered by FEMA through the Office of the Governor's Public Safety Office that funds enhanced border security cooperation among Customs and Border Protection (CBP), U.S. Border Patrol, and state, local, tribal, and territorial law enforcement agencies. The program supports joint operations to secure land and water border routes, improve intelligence sharing, and expand 287(g) screening operations within correctional facilities. In 2025, the national priority is Supporting Border Crisis Response and Enforcement, covering training, operational coordination, and risk management. Eligible expenses include operational overtime costs, staffing support for screening activities, and training programs in immigration law, civil rights protections, and 287(g) procedures.