IRIS for Defense

A soldier on a ridge at dusk, locked by a swarm of drones

The challenge

The modern soldier is exposed.

Autonomous threats define modern battlefields. Detection, decision and response remain fragmented at the operator level.

60-80%

of battlefield reconnaissance is drone-enabled.¹

90%+

of counter-UAS systems are built for fixed or vehicle defense, not individual operators.²

5-15 sec

or less. The detection and response window against low-altitude threats.³

Exposure is lethal.

¹ RUSI, Tactical Developments in the Third Year of the Russo-Ukrainian War, Feb 2025.

² Congressional Research Service, R48477, DoD Counter-UAS, 2025; U.S. Army FY25 C-UAS budget submission.

³ Ukraine and IDF counter-FPV field reporting, 2024 to 2025; based on FPV speeds of about 130 km/h across typical detection ranges.

Simulated interface

Every system. One operator interface.

Detect

Identify threats in real time.

Decide

Validate and assign the response.

Execute

Deploy autonomous interception.

Operators designate. IRIS executes.

Explore IRIS

Operators on a ridge with drones, vehicles and a city linked to the OMNI X mark by red lines

The solution

IRIS. Interoperable Response and Intelligence System.

One orchestrator where detection, decision and response are connected. Real-time coordination across drones, sensors and operators.

The architecture

One orchestrator. Every system connected.

Data sets, drones, robotics and sensors connected through an agentic command layer.

MCP

Modular. Coordination across mission systems.

A2A

Agentic. Systems negotiate resources with each other.

NANDA

Autonomous. Decision-making in dynamic battlespace.

Core LLM core / Memory / RAG

Decode. Normalize. Authenticate. Rate-limit. Audit.

Simulated IRIS view through the goggles: drone lock

Simulated interface

What the operator sees.

Threats, assets and system status in the operator's field of view. Designate by gaze, gesture or voice. Authorize. The interceptor does the rest while the operator moves to the next target.

Simulated interface

Built for the real world

A human authorizes. Every time.

Human in the loop

Every action requires operator authority.

Degraded-mode resilient

Built to operate in low-connectivity and contested environments.

Fully auditable

Every decision and every action is logged for review.

Modular and open

Works across a wide range of drones, sensors and systems.

The category we are creating

Not AI analytics. Not drone hardware. Not fixed-site defense.

Every vendor owns a layer. OMNI X runs above them: operator-level orchestration at the tactical edge, where the action happens.

Vertical wins inside one stack. IRIS works across stacks.

Visibility

Action

Command center

Enterprise intelligence

Autonomous platforms

Tactical edge

Fixed-site counter-UAS

OMNI X / IRIS

Operator-level orchestration.

An operator at a command table with drones and a satellite linked by red lines

Why now

Autonomous threats are evolving faster than defense systems can respond.

The gap is real-time coordination at the operator level.

01
Drone warfare is proven at scale. Cheap autonomous systems now decide engagements.
02
Budgets are accelerating. Counter-UAS and autonomy budgets are accelerating, and procurement vehicles such as SBIR, OTA and TACFI reward integrators, not only platform builders.
03
Agentic AI has crossed the threshold. Real-time multi-system orchestration is now buildable by a small team.
04
Every vendor is shipping a closed ecosystem. The window to become the neutral orchestration layer is open now, and it closes as lock-in hardens.