DIU seeks AI engine for real-time space missile threat fusion under 5s latency

The solicitation targets gaps in existing tools for distinguishing closely spaced objects and maintaining current threat models amid high-speed, multi-domain attacks. It specifies fusion of live video, radar, satellite imagery, geospatial data, and classified reports to generate confidence-scored outputs for human-in-the-loop or automated C2 workflows. Contractors must prove measurable gains in detection speed and attribution accuracy over baselines.

Procurement records show this aligns with DIU's pattern of rapid prototyping for contested environments, following Northrop Grumman's July Camgian partnership and BAE-Scale AI collaborations for integrated air and missile defense. Open-source requirements with machine-to-machine APIs introduce tension with classified data handling, a recurring friction in defense AI programs where audit trails and model provenance remain under-documented in public filings.

Operational pressure stems from interceptor scarcity documented in Ukraine and Red Sea engagements, forcing faster allocation decisions. The five-second latency threshold and burst throughput specs indicate intent to close the sensor-to-shooter loop beyond current manual processes, with implications for escalation dynamics if automated workflows reduce human veto points.

Next steps center on September 24 proposals, followed by prototype evaluation against live test data. Contract awards will likely favor teams already embedded in service sensor architectures.