SBIR/STTR Specific Topic Pre-Release
AF CYBERWORX
SBIR/STTR
Through the Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) programs, America’s Seed Fund provides non-dilutive funding to help small businesses develop innovative technologies and bring them closer to commercialization. AF Cyberworx is hosting three topics in this funding round, offering a unique opportunity for businesses to align their solutions with the program’s objectives. Explore the details below to learn more about our focus areas and how to participate. If you have any questions about eligibility, project alignment, or our capabilities, we encourage you to reach out.
AF CYBERWORX
SBIR/STTR
Through the Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) programs, America’s Seed Fund provides non-dilutive funding to help small businesses develop innovative technologies and bring them closer to commercialization. AF Cyberworx is hosting three topics in this funding round, offering a unique opportunity for businesses to align their solutions with the program’s objectives. Explore the details below to learn more about our focus areas and how to participate. If you have any questions about eligibility, project alignment, or our capabilities, we encourage you to reach out.
AF CYBERWORX
OUR THREE SPECIFIC TOPICS
The SBIR 25.4 / STTR 25.D Pre-Release period runs from May 7 – May 27. During this time, small businesses can review solicitations, ask technical questions directly to topic authors, and gain a deeper understanding of program requirements before the official announcement period begins. Once the announcement opens, direct communication with topic authors will no longer be permitted. We encourage interested businesses to take advantage of this window to gather critical insights. If you need guidance on the application process or want to determine whether your project aligns with our SBIR/STTR focus areas, contact us today or join our specified information sessions!
AI/ML-ENHANCED RISK MANAGEMENT FRAMEWORK
AI/ML-Enhanced Risk Management Framework (SBIR: D2P2 - AF NUMBER: AF254-D0802)
Topic Objective: Develop a software application that employs AI/ML or similar methodologies to automate the Risk Management Framework (RMF) process which is required to achieve Authority To Operate (ATO) for software and hardware products on government networks.
Topic Description: The current RMF process relies heavily on manual efforts and human expertise, which can result in delays, inconsistencies, and potential oversights. As the DoD continues to adopt advanced technologies and faces increasingly sophisticated cyber threats, there is a pressing need to streamline and automate the RMF process to ensure the timely and effective management of risks. AI and ML technologies offer promising solutions to address these challenges by enabling data-driven decision-making, predictive analytics, and automated risk assessment. USAF CIO, USSF, MAJCOM/A6s, and program offices are highly interested in the development of an AI/ML-powered RMF platform that integrates with existing DoD systems and processes. The ideal platform will leverage advanced algorithms and techniques, such as natural language processing, graph analytics, and deep learning, to automate and optimize various aspects of the RMF process.
Topic Phase I Description: It is expected that proposers provide evidence of sufficient prior work and feasibility study to apply AI/ML or similar methodologies to the Risk Management Framework.
Topic Phase II Description: Provide a prototype software application which employs AI/ML or similar methodologies to automate the RMF process. Provide a demonstration of the prototype evaluating an example product which has already been through the manual RMF process within the last two years (achieve TRL 6 maturity).
DETECTION OF UNCREWED AIRCRAFT SYSTEMS IN CLUTTERED ENVIRONMENTS
Detection of Uncrewed Aircraft Systems in Cluttered Environments (STTR P1: AF NUMBER: AF25D-T008)
Topic Objective: Provide an effective sensor technology to detect small uncrewed aircraft systems autonomously flying low to the ground through cluttered environments (obscured in the foreground by trees or buildings).
Topic Description: Military installations are often surrounded by cluttered environments which can now be exploited by autonomous and semi-autonomous Uncrewed Aircraft Systems (UAS). Advances in intelligent navigation systems allow UAS threats to fly low to the ground through trees and between buildings to sneak into unauthorized areas and surveil or attack high value targets on the ground with little to no warning. Existing electro-optical, infrared, radar, and lidar systems cannot effectively detect these threats when flying low to the ground in clutter, especially when obscured in the foreground by clutter. This challenge is further exacerbated by emerging capabilities for small, inexpensive drones to fly together in large, coordinated “swarms” to intelligently avoid detection and overwhelm defenses. Air Combat Command’s J3 Operations Directorate and J4 Logistics Directorate have joint responsibility for Counter-UAS (C-UAS) activities and Air Base Air Defense. ACC is highly interested in novel approaches to detect low flying, relatively small and slow UAS threats in cluttered environments. It is believed that emerging work in acoustics and wireless/RF signal sensors, as well as artificial intelligence and machine learning, may offer a solution to the difficult C-UAS detection problem by “seeing” through trees and buildings. Proposed solutions need to provide a cost-effective method for detection of UAS threats for large military installations with vast perimeters to defend. To be effective, the solutions must be able to distinguish between actual threats and false alarms such as birds. It is acceptable to trade off range by, for example, operating in a mesh network. The new technology will ultimately need to communicate with traditional sensor systems in a layered defense architecture.
Topic Phase I Description: Develop a feasibility study that identifies limitations of available sensor technology to detect small UAS in clutter, evaluates potential novel sensor technologies and methodologies, and proposes an acceptable development path.
AI/ML-GENERATED DECOY NETWORKS
AI/ML-Generated Decoy Networks (SBIR P1: AF NUMBER: AF254-0801)
Topic Objective: Provide a software application that generates decoy networks that are 1) efficient to employ (maximum automation, minimum manual inputs) and 2) realistic enough to deceive a sophisticated state-sponsored hacker. It is expected that recent advancements in machine learning and artificial intelligence will support this objective.
Topic Description: Defensive Cyber Operations (DCO) across the Air Force and DoD face a daily onslaught of state-sponsored expert hackers. Due to the quantity and sophistication of these adversaries, it is insufficient to rely solely on firewalls, anomaly/intrusion detection software, and human monitors. An additional method of defense is to create decoy networks, often referred to as “honey pots” or “honey nets” (in the case of multiple connected decoy networks). These decoys are intended to lure adversaries into wasting time and exposing their tactics, techniques, and procedures (TTPs) in a simulated environment where they can do no harm. While promising, past attempts to create decoy networks have been overly burdensome to create and largely ineffective against expert hackers because they are too easy to identify as fake. Air Force CyberWorx, 16th Air Force, and Air Combat Command are highly interested in novel approaches to create more realistic “digital twin” decoy networks that are dynamic. These networks need to accurately simulate users, infrastructure, data, and data flows. It is believed that emerging work in artificial intelligence, machine learning, expert systems, virtualization, and block chain technologies could dramatically improve realism and assist in counter measures. Proposed solutions could be trained on live networks to mirror characteristics and behaviors then apply algorithms to create the decoy and dynamically change like real networks would and adapt to threat behavior. Additional training of the algorithms could be provided by expert “white hat” cyber operators to improve fidelity. The system should detect, distract, and track the adversary and report activity to authorized defensive cyber operators. Decoy modifications or actions against the threat in real time should be selectable as automated, semi-automated, and/or manual.
Topic Phase I Description: Provide a feasibility study that evaluates potential AI/ML or other similar methodologies and recommend an approach to implement these methodologies in a user-friendly software application that allows defensive cyber operators to generate and manage realistic, dynamic decoy networks and track hacker activity in real-time without the hacker knowing they are being watched or manipulated.
OR
Join Our Ask Me Anything Info Sessions
Our team will be having an open invite information session where we will discuss contract goals and answer FAQ.
May 13, 2025
May 22, 2025
AF CYBERWORX
OUR THREE SPECIFIC TOPICS
The SBIR 25.4 / STTR 25.D Pre-Release period runs from May 7 – May 27. During this time, small businesses can review solicitations, ask technical questions directly to topic authors, and gain a deeper understanding of program requirements before the official announcement period begins. Once the announcement opens, direct communication with topic authors will no longer be permitted. We encourage interested businesses to take advantage of this window to gather critical insights. If you need guidance on the application process or want to determine whether your project aligns with our SBIR/STTR focus areas, contact us today or join our specified information sessions!
AI/ML-ENHANCED RISK MANAGEMENT FRAMEWORK
AI/ML-Enhanced Risk Management Framework (SBIR: D2P2 - AF NUMBER: AF254-D0802)
Topic Objective: Develop a software application that employs AI/ML or similar methodologies to automate the Risk Management Framework (RMF) process which is required to achieve Authority To Operate (ATO) for software and hardware products on government networks.
Topic Description: The current RMF process relies heavily on manual efforts and human expertise, which can result in delays, inconsistencies, and potential oversights. As the DoD continues to adopt advanced technologies and faces increasingly sophisticated cyber threats, there is a pressing need to streamline and automate the RMF process to ensure the timely and effective management of risks. AI and ML technologies offer promising solutions to address these challenges by enabling data-driven decision-making, predictive analytics, and automated risk assessment. USAF CIO, USSF, MAJCOM/A6s, and program offices are highly interested in the development of an AI/ML-powered RMF platform that integrates with existing DoD systems and processes. The ideal platform will leverage advanced algorithms and techniques, such as natural language processing, graph analytics, and deep learning, to automate and optimize various aspects of the RMF process.
Topic Phase I Description: It is expected that proposers provide evidence of sufficient prior work and feasibility study to apply AI/ML or similar methodologies to the Risk Management Framework.
Topic Phase II Description: Provide a prototype software application which employs AI/ML or similar methodologies to automate the RMF process. Provide a demonstration of the prototype evaluating an example product which has already been through the manual RMF process within the last two years (achieve TRL 6 maturity).
DETECTION OF UNCREWED AIRCRAFT SYSTEMS IN CLUTTERED ENVIRONMENTS
Detection of Uncrewed Aircraft Systems in Cluttered Environments (STTR P1: AF NUMBER: AF25D-T008)
Topic Objective: Provide an effective sensor technology to detect small uncrewed aircraft systems autonomously flying low to the ground through cluttered environments (obscured in the foreground by trees or buildings).
Topic Description: Military installations are often surrounded by cluttered environments which can now be exploited by autonomous and semi-autonomous Uncrewed Aircraft Systems (UAS). Advances in intelligent navigation systems allow UAS threats to fly low to the ground through trees and between buildings to sneak into unauthorized areas and surveil or attack high value targets on the ground with little to no warning. Existing electro-optical, infrared, radar, and lidar systems cannot effectively detect these threats when flying low to the ground in clutter, especially when obscured in the foreground by clutter. This challenge is further exacerbated by emerging capabilities for small, inexpensive drones to fly together in large, coordinated “swarms” to intelligently avoid detection and overwhelm defenses. Air Combat Command’s J3 Operations Directorate and J4 Logistics Directorate have joint responsibility for Counter-UAS (C-UAS) activities and Air Base Air Defense. ACC is highly interested in novel approaches to detect low flying, relatively small and slow UAS threats in cluttered environments. It is believed that emerging work in acoustics and wireless/RF signal sensors, as well as artificial intelligence and machine learning, may offer a solution to the difficult C-UAS detection problem by “seeing” through trees and buildings. Proposed solutions need to provide a cost-effective method for detection of UAS threats for large military installations with vast perimeters to defend. To be effective, the solutions must be able to distinguish between actual threats and false alarms such as birds. It is acceptable to trade off range by, for example, operating in a mesh network. The new technology will ultimately need to communicate with traditional sensor systems in a layered defense architecture.
Topic Phase I Description: Develop a feasibility study that identifies limitations of available sensor technology to detect small UAS in clutter, evaluates potential novel sensor technologies and methodologies, and proposes an acceptable development path.
AI/ML-GENERATED DECOY NETWORKS
AI/ML-Generated Decoy Networks (SBIR P1: AF NUMBER: AF254-0801)
Topic Objective: Provide a software application that generates decoy networks that are 1) efficient to employ (maximum automation, minimum manual inputs) and 2) realistic enough to deceive a sophisticated state-sponsored hacker. It is expected that recent advancements in machine learning and artificial intelligence will support this objective.
Topic Description: Defensive Cyber Operations (DCO) across the Air Force and DoD face a daily onslaught of state-sponsored expert hackers. Due to the quantity and sophistication of these adversaries, it is insufficient to rely solely on firewalls, anomaly/intrusion detection software, and human monitors. An additional method of defense is to create decoy networks, often referred to as “honey pots” or “honey nets” (in the case of multiple connected decoy networks). These decoys are intended to lure adversaries into wasting time and exposing their tactics, techniques, and procedures (TTPs) in a simulated environment where they can do no harm. While promising, past attempts to create decoy networks have been overly burdensome to create and largely ineffective against expert hackers because they are too easy to identify as fake. Air Force CyberWorx, 16th Air Force, and Air Combat Command are highly interested in novel approaches to create more realistic “digital twin” decoy networks that are dynamic. These networks need to accurately simulate users, infrastructure, data, and data flows. It is believed that emerging work in artificial intelligence, machine learning, expert systems, virtualization, and block chain technologies could dramatically improve realism and assist in counter measures. Proposed solutions could be trained on live networks to mirror characteristics and behaviors then apply algorithms to create the decoy and dynamically change like real networks would and adapt to threat behavior. Additional training of the algorithms could be provided by expert “white hat” cyber operators to improve fidelity. The system should detect, distract, and track the adversary and report activity to authorized defensive cyber operators. Decoy modifications or actions against the threat in real time should be selectable as automated, semi-automated, and/or manual.
Topic Phase I Description: Provide a feasibility study that evaluates potential AI/ML or other similar methodologies and recommend an approach to implement these methodologies in a user-friendly software application that allows defensive cyber operators to generate and manage realistic, dynamic decoy networks and track hacker activity in real-time without the hacker knowing they are being watched or manipulated.
OR
Join Our Ask Me Anything Info Sessions
Our team will be having an open invite information session where we will discuss contract goals and answer FAQ.