DeepNight

DeepNight develops a software layer that applies deep neural networks to imagery from commoditized electro-optical cameras. The system aggregates photons over time to reconstruct visible scenes in moonless, ultra-low-light conditions. The company states this approach exceeds the performance of analog Gen 3 image intensifiers while costing less than thermal cameras.

DeepNight has not publicly disclosed its founders, leadership team, organizational structure, or headquarters location. The firm operates through its website and defense-contractor press releases without identifying any named principals.

Documented deployments include sensor integration for U.S. military site protection with Picogrid, announced in November 2025, and a counter-UAS deployment across U.S. Air Force installations with Circle Optics in April 2026. All known engagements are in the defense and critical infrastructure protection sector.

The firm claims its neural-network approach produces higher resolution than thermal cameras at lower cost, while recovering true scene information in starlight-only conditions that conventional digital sensors cannot capture. It explicitly benchmarks against Gen 3 image intensifier tubes, asserting superior performance in darkness without the light signature of infrared illumination.

DeepNight's capital structure, funding sources, and investment backing are not publicly disclosed. The firm does not identify any venture capital partners, family office sponsors, or strategic investors on its website, and no regulatory filings revealing its ownership are available (Altss research).