One of the harshest illusions of high-tech warfare in recent years has been faith in invulnerable GPS and constant connectivity. Reality turned out differently: electronic warfare rapidly turned satellite navigation and communication channels into a weak link. And it is precisely at this point that a new class of military AI systems emerged—autonomous, “blind-resilient,” capable of operating where digital infrastructure is effectively destroyed.
In late 2025, the American company Safe Pro Group Inc. introduced updated algorithms for its SPOTD platform (Safe Pro Object Threat Detection), enabling drones to detect mines and kamikaze drones without GPS or external communications. This isn’t a lab theory: the solutions were refined based on real combat conditions and the requests of operators working in Ukraine—where EW became the norm rather than the exception.
The culmination will be a demonstration of the technology during U.S. Army exercises, the Concept Focused Warfighter Experiment 2026 at Fort Hood. Formally, it’s another field trial. In practice, it’s a test of a future battlefield logic—one where autonomy matters more than satellite precision.
AI that navigates without coordinates
SPOTD works differently from classic navigation and reconnaissance systems. Instead of relying on GPS, the platform analyzes drone video feeds, matching visual features of terrain, threats, and objects. On that basis, it builds 2D and 3D models of space suitable for route planning for both ground and aerial unmanned systems.
The key point is architectural flexibility. SPOTD can run in the Amazon Web Services cloud or locally on edge devices—critical under communications suppression. In rapid mapping mode, the system can function even without active AI detection, providing a baseline navigation and reconnaissance picture.
The numbers behind the technology explain why it’s being taken seriously. More than 2.2 million analyzed drone images, over 41,400 detected threats, and coverage of more than 11,400 hectares—this is not a synthetic dataset, but accumulated battlefield experience turned into a training base.
Why this changes the logic of war itself
The main shift here isn’t even mine detection or kamikaze drone spotting per se. The real pivot is abandoning dependence on global navigation systems. If a drone can orient itself, analyze threats, and build a map autonomously, EW stops being the “red button” that simply switches off the opponent’s technology.
This implies a transition from a war of networks to a war of agents. Each drone becomes an independent unit of reconnaissance and decision-making. Loss of connectivity no longer means loss of effectiveness—only a limitation on coordination scale.
That is why such solutions matter beyond the military. The same logic of autonomous perception without GPS applies to robots in destroyed urban environments, disaster response, and underground or industrial zones where navigation has always been unstable.
2026: autonomy as the standard
U.S. Army trials in 2026 will be an indicator of a much broader trend. Future robots and drones will be designed under the assumption that connectivity can vanish at any moment. Autonomy stops being an advantage—it becomes a baseline requirement.
In that sense, SPOTD is not just one company’s product, but a symptom of an entire industry’s maturity. Robotics is leaving the era of “connected devices” and entering a phase of independent systems that don’t ask infrastructure for permission.
And that may be the most important technological lesson of late 2025: in a world of constant interference, the survivor is not the one who is better connected, but the one who can see and understand reality alone.
Source and developer:
Safe Pro Group Inc. — https://safeprogroup.com
