Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
AI Overview
This SBIR seeks a wideband digital recording system capable of capturing high-volume sensor data at the tactical edge to support AI/ML algorithm development. The device must fit within compact rack space while meeting strict security and environmental standards for deployment on military platforms.
This summary is AI-generated from the official solicitation.
Key Details
Official Description
In today’s environment, emphasis is put on how Artificial Intelligence/Machine Learning (AI/ML) can solve most of the Department of War’s (DOW) problems as long as the AI/ML algorithms are trained correctly. This training requires vast amounts of relevant data. Unlike commercial websites where the algorithm developers can have the public train them based on security selection images, the DOW does not have vast stores of relevant data sets much less a global community to train the algorithms. Unf...
Change History
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
# Q&A Changes Summary **New Answers Added:** - **A2**: Confirmed requirement is ≥400 GigaBytes/sec (not Gigabits/sec) for real-time recording - **A3**: Clarified data labeling requirements—must support multiple data types, use Zulu time in filenames, enable querying/searching by data type, and maintain synchronization during playback - **A4**: Expanded throughput clarification—aggregate streams acceptable (e.g., 8×50 GB/sec simultaneously), continuous streaming required with zero gaps/data loss, playback must support real-time plus variable rate options, offline data transfer acceptable, storage requirement is aggregate capacity **Key Technical Clarifications:** - Throughput measured as aggregate across simultaneous streams, not per-interface - Continuous operation mandatory; buffering permitted only if no data loss occurs - Playback must support variable rates (faster/slower than recorded) - System must adapt to multiple networking infrastructures (InfiniBand, NVLINK, PCIe, Ethernet, etc.)
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
# Q&A Changes Summary **New Question Added:** Q4 expanded significantly with 13 detailed sub-questions addressing throughput measurement points (network ingress vs. end-to-end), port distribution, dual-interface requirements, protocol specifics (RoCEv2, RDMA, VITA-49), buffer staging, data retention, timing/synchronization interfaces, storage capacity definition, encryption standards, physical form factor constraints, and playback functionality interpretation. **Clarification Confirmed:** Q8 answer reaffirmed that the 400 GB/sec requirement is **GigaBytes (not Gigabits)**. **Key Impact:** The expanded Q4 addresses critical technical ambiguities that directly affect feasibility and cost estimates, particularly regarding whether throughput is measured at different system points and whether storage write performance can use buffering strategies.
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
**Summary of Q&A Changes:** Added 5 new questions clarifying technical requirements: (1-2, 4-5) Protocol specifications for sensor suite integration and software requirements for recording/playback/search functions; (3) Confirmation that 400GB/s (not 400Gb/s) is the actual requirement despite extreme technical difficulty. Q6 repeats original clarification that requirement is 400GigBytes/sec.
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
Status changed from Pre-Release to Open
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
Added answer to Q1 clarifying that the 400 GB/sec data transfer rate requirement refers to Gigabytes per second (not Gigabits per second).
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
# Summary This Q&A clarifies a technical specification ambiguity regarding data transfer rates in the funding opportunity, asking whether the stated 400 GB/sec refers to gigabits or gigabytes per second.
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
Close Date changed from 2026-04-22 to 2026-06-03
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
Open Date changed from 2026-03-25 to 2026-05-06
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
Status changed from Removed to Pre-Release
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
Opportunity DON26BZ01-NV022 no longer available
Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
New opportunity: Extremely Wide Band Digital Recording System for Artificial Intelligence/Machine Learning Development
Get Alerts for Opportunities Like This
RallyProp monitors thousands of funding sources and sends instant alerts when opportunities match your technology and capabilities.
Start Free Trial