Leopard DPU - On-Board Data Processing Unit

Leopard – On-Board AI Processing for Micro and Mini Satellites

Leopard is a high-performance Data Processing Unit (DPU) built for on-board data processing in micro and mini satellite missions. Designed in a compact PC/104 form factor and compatible with CubeSats (<1U), Leopard brings the power of edge AI directly into space. Instead of downlinking vast volumes of raw data, Leopard enables satellites to process and interpret data in orbit – delivering only the most valuable insights to Earth. This approach dramatically reduces bandwidth usage and mission costs while accelerating response times.

Key features include:

• High-Performance Processing – Capable of up to 3 TOPS (Tera Operations Per Second), enabling real-time, low-latency analysis of complex datasets.

• AI-Driven Insights – Supports deep learning models onboard, allowing for intelligent decision-making and automated data interpretation directly in orbit.

• Compact & Flexible Design – Space-optimized, CubeSat-compatible (<1U), and easily integrated with various micro and mini satellite platforms.

Applications include:

• Earth Observation – Real-time image segmentation, classification, and object detection.

• Space Situational Awareness (SSA) – Onboard analysis to support collision avoidance, debris tracking, and in-orbit diagnostics.

Part of the Smart Mission Ecosystem – comprehensive solutions, including a fully integrated package of hardware, software, and in-house developed algorithms for onboard satellite data processing.

The Leopard DPU can be combined with a Simera Sense imager. By combining Hyperscape100 hyperspectral imaging with Leopard DPU, this integrated solution enables faster, smarter, and more efficient Earth Observation - with key data acquired, processed and stored directly onboard the satellite.