[ SOVEREIGN EDGE AI · ISR · AUTONOMOUS SYSTEMS ]

From detection
to understanding.

SAPIEN X is a sovereign reasoning layer for ISR. It runs onboard drones and ground vehicles and in mobile command centres, turning raw detections into assessed, geolocated threats ranked for the commander. From ISR to decision, onboard and on the ground — fully offline, with no third-party dependency.

Onboard Fully offline Hardware-agnostic Plug & play Human in the loop

[ THE PROBLEM ]

Detection is not awareness.

Fielded AI draws boxes: a class and a confidence score for each point of interest, frame after frame. It does not relate events, retain what happened earlier in the mission or state what a scene means. Situational awareness still sits in the operator's head.

[ THE PROBLEM ]

Cognitive overload at the operator.

1:8Operator : drones

A single ISR operator now supervises up to eight drones, often alongside UGVs on the same mission. Every feed converges on one station; correlating events and holding the mission timeline are done manually, for the full duration of the mission.

[ THE SOLUTION ]

Scene understanding, onboard.

SAPIEN X runs on the platform. Each frame is re-assessed: every point of interest is identified, described and assessed for threat, with a confidence score, and the scene is summarised. No dependence on the datalink.

[ THE SOLUTION ]

Event correlation over time.

Each new point of interest is checked against earlier detections along the route. Movement, persistence and new activity are kept on a single mission timeline, in timestamped reports per frame and per platform.

[ THE SOLUTION ]

A ranked picture for the commander.

Reports from every platform are aggregated at the operator station and ranked by threat and confidence — nothing is discarded. The command centre cross-analyses them into one picture of the area of operations.

[ OODA ]

Observe · Orient

Automated

Scene understanding, event correlation and temporal analysis run on the platform and at the station. Threat, confidence, position and timeline arrive ranked.

Decide · Act

Human

The operator reviews ranked threats, decides and acts. Authority stays with the operator and the command.

The system qualifies. The human decides.


[ TECHNOLOGY ]

Inside the
reasoning layer.

01 / Scene understanding

What the frame means

Each point of interest is identified, described and assessed for threat, and the scene is summed up with the points that need attention.

02 / Event correlation

Now, linked to before

Every new point of interest is checked against earlier detections along the route and against the other vehicles of the mission.

03 / Temporal analysis

One mission timeline

Change, persistence and appearance are kept across the mission, so a situation is read in its direction, not as a snapshot.

04 / Geolocation

Whatever the navigation

Each point of interest is geolocated with its level of uncertainty, from the vehicle's own navigation source, including in GPS-denied environments.

05 / Deployment

Onboard, station, command

The same core runs in the vehicle, at the operator station and in the command centre. Air-gapped operation, signed updates, zero egress.

01 / 05

[ DEPLOYMENT ]

One cognitive core.
Every vehicle.

01 / Drone · VTOL

Onboard understanding

Scene understanding and collection along the route, on the compute the aircraft already carries, with no dependence on the link to the ground.

02 / UGV

Ground-level awareness

Ground-level observations correlated with the wider mission as it unfolds.

03 / Mobile command centre

One picture of the AO

Every platform fused into a single, ranked picture of the area of operations, updated through the mission.

Plug & play. Our solution is hardware-agnostic: it integrates into partner drones, UGVs and C2 systems, above the detectors already in service.


[ WHY NOW ]

Warfare has changed speed.

01 / Active conflict

Tactics change weekly

Current conflicts produce new tactics faster than doctrine or fielded models can follow. The side that understands first holds the advantage.

02 / Autonomous mass

Fleets are multiplying

Drones, UGVs and loitering systems are fielded by the thousand. Every new platform is one more feed for a human to interpret.

03 / Edge-ready stack

Reasoning fits onboard

Embedded compute and a maturing autonomy stack now allow reasoning models to run inside the vehicle, offline.

The window to build the cognitive layer of autonomous warfare is now.


[ WHY SAPIEN X ]

Built from the
problem up.

  1. Problem-first. We start from the operational problem, not from the model.
  2. Weeks, not years. From requirement to field demonstration in weeks. Models adapt to new tactics in days, not years.
  3. End to end, in-house. Models, edge software and platform integration are built in-house.
  4. Military advisory. We work with a military advisor. Operational feedback shapes every build.

[ TEAM ]

The team.

Ayman Berouaine

Co-founder · CEO & CTO

Built the full stack, end to end. AI, autonomous systems and edge computing.

Aboubakr Benomar

Co-founder · CFO & Operations

Finance, legal, procurement and defence partnerships.

Supported by a military advisor.


[ SHOWCASES ]

Jun 2026Eurosatory · ParisEurosatory Lab · live demonstrations
Oct 2026Marrakech Air ShowMeet the team
Oct 2026Future Forces · PragueMeet the team
Jan 2027IDEX · Abu DhabiMeet the team

Book a briefing.

The system qualifies. The human decides.

A technical briefing with your ISR, C2 or autonomy team: the platforms you operate, where the operator load sits, and what a reasoning layer changes in the sensor-to-decision chain.

Book a briefing