Explore This Research Award in CaptureIQ
Research the technology, evaluate the recipient and agency, identify related opportunities, and explore potential follow-on work with CaptureIQ.
Award details
- Awarding agency
- Department of Defense
- Branch
- Army
- Recipient company
- PROTOPIA AI INC
- Proposal award date
- 2025-07-17
- Contract end date
- 2027-01-27
- UEI
- Y3WETN67KTA5
- Solicitation number
- 24.4
- Topic code
- A244-P017
Research abstract
The rapid adoption of artificial intelligence (AI) is a strategic imperative for the United States Army to maintain its competitive edge against adversaries, notably the People’s Republic of China (PRC), which has emphasized AI as a cornerstone of its military modernization. However, the Army faces significant barriers in securely deploying AI due to the risks of exposing sensitive data, including medical records, financial information, and weapon system capabilities, to inference systems. Current methodologies, reliant on plain-text data decryption, leave critical information vulnerable to compromise during AI processing. Traditional data protection solutions introduce prohibitive latency, require costly specialized hardware, or necessitate fundamental alterations to AI workflows, rendering them impractical at scale. Protopia AI’s Stained Glass Transform (SGT) addresses these challenges with a scalable, privacy-preserving data protection technology. SGT enables secure AI processing by converting sensitive data into stochastic, randomized representations comprehensible only to authorized AI models. This one-way transformation is irreversible while preserving model response accuracy and imposing negligible latency. The SGT framework seamlessly integrates into existing AI lifecycles as a post-training optimization step and operates efficiently on commodity hardware without requiring specialized infrastructure. Key Innovations of SGT:1. Minimal Latency Impact: SGT maintains near-zero latency during inference, critical for real-time military applications.2. Hardware Agnosticism: SGT requires no specialized hardware, facilitating broad deployment across existing and emerging platforms.3. Workflow Compatibility: No modifications to AI workflows are necessary, ensuring straightforward integration with Army systems. The Stained Glass Engine (SGE), the core product behind SGT, generates tailored transformations for specific AI models. This process preserves data utility for training, validation, and inference while rendering the data unintelligible to unauthorized users. By mitigating risks associated with data exposure, SGT supports a variety of Army use cases, including:• Secure inference and prediction in operational environments.• Privacy-preserving verification and validation during AI model testing.• Enabling data obfuscation for large-scale AI training without exposing sensitive content.
Technology keywords
Not reported
Other awards by this recipient
Verified by the same UEI. These are other public SBIR/STTR award pages, not a merged company record.
Awards with the same topic code
Verified by an exact topic code match. This is not a text-similarity match.
Research funding history
2 SBIR/STTR records share this UEI (2025). Reported award amounts total $3,244,043. Reported amounts are not federal obligations.
Phases: Phase II – Direct to Phase II
Agencies: Department of Defense
Agency intelligence
Department of Defense
Back to SBIR/STTR Research Awards