# Arc-Edge v1

> A modular carrier board that combines socketed edge compute and AI acceleration, three universal radio slots, and a program-configured Arc software and model load for aircraft integration.

Product descriptor: Mission-Ready Edge Carrier + PACE

Marketing headline: One flight board. Every mission link.

## Mission fit

- Adding onboard autonomy and perception to small unmanned aircraft
- Changing communications pathways without replacing the carrier architecture
- Integrating one edge stack with PX4, ArduPilot, or another MAVLink autopilot through a validated adapter

## Capabilities

- One flight-optimized base board
- Three universal radio-card slots
- Standard Arc-Cortex configuration described at 140+ TOPS
- Mesh, LTE / 5G, and Iridium Certus paths available by configuration
- Arc runtime, mission models, radio profiles, and vehicle interfaces provisioned for the selected integration

## Hardware

- **Carrier board:** Flight-optimized PCB with three universal radio-card slots. Keeps the core aircraft integration stable while radio cards and compute modules change.
- **Standard compute:** Arc-Cortex using a Jetson Orin NX 16 GB-class module; described at 140+ TOPS. Runs perception, mission reasoning, and autonomy locally.
- **Lite compute:** Pin-compatible Jetson Orin Nano on the same carrier PCB. Supports a lower-SWaP configuration while retaining the carrier architecture.
- **Modular sockets:** Socketed compute, AI accelerator, and radio cards. Allows program-specific upgrades and service without replacing the base board.
- **Platform protection:** TPM 2.0, conformal coating, and vibration isolation. Protects keys and electronics in the intended flight environment.
- **Sensor interfaces:** Carrier design described as supporting more than 20 camera inputs. Feeds onboard perception and sensor-processing pipelines.

## Software

- Arc autonomy runtime and mission orchestration
- Program-selected mission models and edge inference pipelines
- PACE policies and radio profiles for the selected loadout
- MAVLink interfaces configured for the target autopilot
- Update and rollback workflow configured with the program

## Integrations

- PX4
- ArduPilot
- Additional MAVLink autopilots through platform-specific adapter validation
- Program-selected cameras and sensors
- Mesh, cellular, and satellite communications by loadout

## Configurations

- **T1 — Command Core (Dev):** Arc-Cortex; primary link: WFB-native link; role: Development and hardware-in-the-loop baseline.
- **Lite — Arc-Lite:** Jetson Orin Nano on the Arc-Edge carrier PCB; primary link: Microhard; additional paths: LTE; role: Lower-SWaP edge autonomy configuration.
- **T2 — PACE Base:** Arc-Cortex; primary link: Microhard mesh; additional paths: LTE / 5G, Iridium Certus; role: Baseline multi-path PACE communications.
- **T3 — PACE Doodle:** Arc-Cortex; primary link: Doodle Labs Defense Helix; additional paths: LTE / 5G, Iridium Certus; role: Doodle Labs-based tactical mesh loadout.
- **T4 — PACE Silvus:** Arc-Cortex; primary link: Silvus LC5200 OEM; additional paths: LTE / 5G, Iridium Certus; role: Silvus-based tactical mesh loadout.
- **T5 — Swarm Command:** Arc-Cortex; primary link: Silvus LC5200 plus Microhard backbone; additional paths: LTE / 5G, Iridium Certus; role: Command and swarm backbone configuration.

## Qualifications

- PACE means primary, alternate, contingency, and emergency; the role of each link is set by the selected mission architecture.
- The public architecture describes sub-20 ms onboard decision loops, but measured latency depends on sensors, models, compute mode, and integration.
- Camera count, radio availability, environmental hardening, and airworthiness evidence are confirmed for each platform and program.

## Canonical references

- Category: Flight-optimized edge AI carrier board
- Canonical page: https://arcedge.ai/arc-edge-v1
- Full public product reference: https://arcedge.ai/llms-full.txt
- Machine-readable catalog: https://arcedge.ai/site-metadata.json
- Publisher: Arc by Rownd (Rownd, Inc.)
- Contact: defense@rownd.ai
- Last reviewed: 2026-08-12