This press release is provided by GlobeNewswire and is published as received. VANCOUVER, British Columbia, Sept. 01, 2026 (GLOBE NEWSWIRE) -- SPARC AI Inc. (CSE: SPAI; OTCQB: SPAIF; Frankfurt: 5OV0) (“SPARC AI” or the “Company”) a defense technology company delivering positioning for drones in GPS-denied environments, today announced the successful completion of a series of flight tests in which the Overwatch Positioning Network accurately calculated target locations while the drone’s camera was physically blindfolded and covered. With the camera lens fully obscured for the duration of the tests, the aircraft flew a structured test pattern and the Overwatch API returned target coordinates computed entirely from flight telemetry. The results confirm a deliberate design principle of the Overwatch architecture: the positioning solve uses no image recognition software and no pre-loaded reference imagery of any kind. Most alternative approaches to GPS-denied positioning and targeting rely on computer vision with matching camera imagery against stored maps or terrain features. Those systems degrade or fail at night, in fog, smoke, or dust, over water, snow, desert, and featureless terrain, and they require imagery of the operating area to be collected and loaded in advance. Overwatch is telemetry-only and none of these limitations apply. The successful blindfold tests demonstrate that Overwatch performs identically in zero-visibility conditions, and that the onboard camera serves purely as an operator aiming aid and not as a sensor input to the positioning calculation. The vision-free architecture also underpins Overwatch’s integration model. The system requires no hardware or software installed on the aircraft. Its inputs are telemetry that every commercial flight controller already produces, and its output is standard latitude, longitude, and time — indistinguishable to an autopilot from a conventional GPS fix. This means any drone OEM can integrate the SPARC AI API into an existing workflow without redesigning the