Part of the DoW SBIR/STTR FY27 Release 2 topic list. Browse all 83 topics by industry or service.
Topic MDA27BZ02-DV010 from the DoW SBIR 2027 BAA, Release 2. Pre-release now; proposals due January 6, 2027, 12:00 PM ET.
Quick answer
MDA27BZ02-DV010 is a Missile Defense Agency SBIR topic seeking transfer learning tools that train AI agents in low-fidelity or unclassified environments and then adapt them to high-fidelity and classified Missile Defense System (MDS) simulations. It is Direct-to-Phase II only; Phase I-like proposals will be rejected as nonresponsive. Proposals are due January 6, 2027, 12:00 PM ET.
Solicitation at a glance
- Topic number: MDA27BZ02-DV010
- Official title: Multi-Fidelity Transfer Learning for Defense Applications
- Agency: Missile Defense Agency
- Program: SBIR (DoW SBIR 2027 BAA, Release 2)
- Award type: Direct to Phase II (no Phase I awards)
- Award amount: Up to $2,000,000 (Direct to Phase II). Source: MDA Release 2 instructions in DSIP.
- CMMC level: Level 2 (Self)
- ITAR: Yes. Export-controlled (ITAR/EAR); see topic text
- Technology areas: Battlespace, Human Systems, Information Systems
- Pre-release (direct TPOC contact allowed): November 4 to December 2, 2026, 8:00 AM ET
- Opens in DSIP: December 2, 2026, 8:00 AM ET
- Proposals due: January 6, 2027, 12:00 PM ET
- Official source: DSIP Topics app (search MDA27BZ02-DV010)
BW&CO's take
What the Missile Defense Agency wants
Training AI agents for missile defense is held back by the cost, scarcity, and security restrictions of high-fidelity simulation. MDA wants a staged approach: build core agent behaviors in low-fidelity or unclassified environments where data is plentiful, then transfer and refine those models in high-fidelity environments at higher classification levels. Phase II delivers a full prototype with a low-fidelity training environment, transfer learning methods, high-fidelity MDS simulation representations across classification levels, a human-machine interface to orchestrate training workflows, and a case study showing resource savings versus traditional training. Workflows must integrate with existing MDS agent training systems.
Who is a good fit
- Working transfer learning methods for AI/ML agents, with test data or results showing reduced training resources
- Experience building multi-fidelity simulation environments for aerospace, missile defense, or battle management problems
- Ability to operate across classification levels, including handling classified high-fidelity environments
- Software for orchestrating AI training pipelines, including a usable human-machine interface for analysts
Watch-outs
- This is Direct-to-Phase II. You must submit feasibility documentation (technical reports, test data, prototype designs, performance results) showing a Phase I-equivalent concept was already developed, substantially by your team or PI.
- The feasibility documentation should show a proof of concept, validation in a relevant environment, and/or TRL 5 or higher, and must validate technology development already completed.
- The topic is marked ITAR-restricted in the topic metadata, which generally limits participation by foreign nationals; confirm your team's eligibility.
- The topic requires CMMC Level 2 (Self). Phase II also calls for high-fidelity MDS representations across multiple classification levels, so plan for secure work.
Think your company fits MDA27BZ02-DV010? Talk with BW&CO before the January 6 deadline.
Plan Your MDA27BZ02-DV010 Submission
Official topic text (verbatim from DSIP)
The text in this section is copied word for word from the official DSIP topic as published on November 4, 2026. DSIP is the official source and topics can be amended. Always confirm the current version at dodsbirsttr.mil before you submit. Our commentary above is BW&CO's interpretation, not Government guidance.
Title
Multi-Fidelity Transfer Learning for Defense Applications
Objective
Use new and evolving Artificial Intelligence (AI) training techniques, algorithms, and protocols to rapidly reduce the resources required to integrate AI into the Missile Defense System (MDS).
Description
Rapid Artificial Intelligence (AI) integration into the MDS is required for the United States to properly defend itself against its adversaries. The development of robust Artificial Intelligence/Machine Learning (AI/ML) agents for complex missile defense and aerospace applications is often constrained by the computational cost, limited availability, and security restrictions of high-fidelity simulation environments. Transfer learning provides a critical pathway to overcoming these barriers. By initially training agents on low-fidelity models or unclassified representations—where data is abundant and computational demands are reduced—core behaviors, strategies, and decision-making frameworks could be efficiently developed and refined. These pretrained models could then be adapted and optimized in high-fidelity and/or environments at higher classification levels, allowing the agent to leverage its foundational knowledge while rapidly acquiring the nuanced performance required for operational deployment. This staged approach significantly reduces training time, minimizes resource expenditure, and mitigates the risk of overfitting to scarce high-fidelity/classified data. Moreover, it enables iterative development cycles where low-fidelity environments serve as agile testbeds for innovation, and high-fidelity environments refine and validate for mission-readiness. Leveraging transfer learning in this manner not only accelerates capability delivery but also ensures that AI/ML systems achieve the precision, resilience, and adaptability necessary for real-world mission success.
Phase I
Phase I-like proposals will not be evaluated and will be rejected as nonresponsive. For this topic, the Government expects the small business would have accomplished the following in a Phase I-like effort via some other means, e.g., independent research and development (IRAD) or other source, a concept for a workable prototype or design to address, at a minimum, the basic capabilities of the stated objective above. Proposal must show, as appropriate, a demonstrated technical feasibility or nascent capability. The documentation provided must substantiate the proposer’s development of a preliminary understanding of the technology to be applied in their Phase II proposal in meeting topic objectives.
Documentation should comprise all relevant information including, but not limited to, technical reports, test data, prototype designs/models, and performance goals/results. Feasibility = maturity and what have you already done/validated.
Proposers interested in participating in Direct to Phase II must include in their responses to this topic Phase I feasibility documentation that substantiates the scientific and technical merit and Phase I feasibility described in Phase I above has been met.
(i.e., the small business must have performed a proof of concept like “Phase I” component and/or other validation in a relevant environment, and/or at a much higher Technology Readiness Level (TRL) level (5 or higher) and describe the potential commercialization applications. The documentation provided must validate that the proposer has completed development of technology in previous work or research completed.)
IRAD work: Documentation should include the most relevant information including, but not limited to: technical reports, test data, prototype designs/models, and/or performance goals/results. Work submitted within the feasibility documentation must have been substantially performed by the proposer and/or the principal investigator (PI).
Phase II
Create a full prototype capability implementing the tools and demonstrating the processes required to enable AI transfer learning for MDS applications. Design and conduct a case study using the prototype implementation with representative data and models.
Key Technical Objectives:
-Establish a low fidelity training environment based on limited data for AI agent learning applicable to MDS use cases.
-Develop, apply, and implement transfer learning methodologies to drive transfer learning.
-Develop and integrate high fidelity representations of MDS simulation environment across multiple levels of classification.
-Demonstrate resource savings of transfer learning methods versus traditional training methods.
-Design and implement a human-machine interface to seamlessly orchestrate the AI training workflows.
-Deliver prototype AI training workflows for mission relevant use cases that integrate with existing MDS agent training systems.
Phase III Dual Use Applications
Scale up the capability from the prototype utilizing the new
processes and software technologies developed in Phase II into a mature, fieldable capability. Work with additional DoD or commercial organizations to integrate the technology into a system level testbed for analyst use.
References
Suder, Piotr M., Jason Xu, and David B. Dunson. "Bayesian Transfer Learning." Statistics 2024.2312 (2023).
Reeve, H., Cannings, T., & Samworth, R. (2021). Adaptive transfer learning. Computing Research Repository, 2021(2106).
Cody, T., & Beling, P. (2021). A Systems Theory of Transfer Learning. Computing Research Repository, 2021(2107).
Keywords
Transfer Learning; AI/ML; Data Science; Simulation; Modeling; Applied Statistics; Data Reduction; Battle Management
Frequently asked questions
What is MDA27BZ02-DV010?
It is a Missile Defense Agency SBIR topic titled Multi-Fidelity Transfer Learning for Defense Applications. It seeks AI training techniques that reduce the resources needed to integrate AI into the Missile Defense System by pretraining agents in low-fidelity environments and adapting them in high-fidelity ones.
When is MDA27BZ02-DV010 due?
Proposals are due January 6, 2027, 12:00 PM ET. DSIP pre-release began November 4, 2026, and direct contact with the topic authors is allowed only until the topic opens on December 2, 2026, 8:00 AM ET.
Is this a Phase I or Direct-to-Phase II topic?
It is Direct-to-Phase II. Phase I-like proposals will not be evaluated, and proposers must include Phase I feasibility documentation showing prior proof-of-concept work, validation in a relevant environment, or TRL 5 or higher.
What must the Phase II prototype demonstrate?
It must establish a low-fidelity training environment from limited data, apply transfer learning methods, integrate high-fidelity MDS simulation representations across classification levels, demonstrate resource savings versus traditional training, provide a human-machine interface for training workflows, and deliver workflows that integrate with existing MDS agent training systems.