Modern orbital satellite payloads generate extremely high-rate imagery and sensor data in a second. Traditionally, this raw data is transmitted to Earth stations for post processing but this process is limited by channel bandwidth, power, and latency constraints between ground station contact. As a result, these high data rate of unprocessed terabytes of raw data from satellites often send, while only a small portion is truly mission-critical.
On-orbit AI/ML processing is fundamentally overcoming these constraints and enabling new processing architectures that significantly reduce computational encumbrances. With radiation-tolerant AMD XQRVE2302 Versal AI Edge SoCs, satellites can perform real-time inference, feature extraction, and data triage directly in space. By leveraging AI Engines, programmable logic, and heterogeneous compute resources, these devices enable onboard classification, compression, and event detection reducing downlink volume by orders of magnitude while minimizing latency for time-critical missions.
This article explores how the Versal AI Edge SOM is accelerating autonomous and intelligent space missions and why it’s emerging as a game-changer for modern satellite systems.
Key challenges in traditional satellite data pipelines:
- Limited downlink bandwidth restricts how much data a satellite can transmit.
- High latency between data capture and interpretation leads to delayed actions.
- Power and thermal constraints make large compute architectures unsuitable.
- Growing complexity of Earth observation, ISR, climate monitoring, and lunar missions.
Onboard AI solves these challenges by enabling:
- Data filtering and compression at the source
- Real-time insights without waiting for ground processing
- Autonomous, mission-driven decision-making
- Efficient bandwidth usage by sending only relevant information
Versal AI Edge System on Module delivers all these benefits in a radiation-tolerant, space-qualified Adaptive SoC, making it ideal for next-generation LEO, GEO, and lunar missions.
The Versal AI Edge XQRVE2302 SOM is a radiation-tolerant variant designed specifically for aerospace and defense. It integrates AI Engines for high-performance ML inference, adaptable logic for sensor-specific acceleration, scalar processing for control and decision tasks, high-speed I/O for multi-sensor payloads, and radiation mitigation features (SEFI protection, EDAC, configuration scrubbing, etc.)
Key Advantages for Space Systems
- Low-latency AI inference for image analytics and signal processing
- Adaptive acceleration for mission updates through reconfigurable logic
- Reduced power consumption compared to GPU-based compute
- Space-grade reliability with radiation-tolerant device architecture
- High throughput to handle large sensors (optical, multispectral, SAR)
While the XQR Versal AI Edge SoC provides the computational foundation for AI-driven in-orbit processing, our radiation-tolerant SOM transforms this capability into a fully engineered, space-ready platform. The module integrates the XQRVE2302 device with space grade DDR memory, high-speed SERDES interfaces, robust power management, and advanced thermal control ensuring reliable operation in space environments. By encompassing radiation-tolerant components, fault-management circuitry and standardized interfaces, the SOM offers a turnkey compute solution for applications such as real-time SAR image formation, hyperspectral classification, object detection, and autonomous satellite decision-making.
In addition to providing a complete platform built around the XQR Versal AI Edge architecture, the SOM is engineered for stringent space-grade reliability. It withstands a Total Ionizing Dose (TID) of up to 30 krad (Si), offers Single-Event Latchup (SEL) immunity beyond 43 MeV·cm²/mg, and demonstrates Single-Event Upset (SEU) robustness above 60 MeV·cm²/mg. The module is qualified to MIL-STD-883E and ensuring resilience against radiation-induced degradation, latchup, and transient fault mechanisms encountered in LEO, MEO, and deep-space environments.
Versal Portfolio from iWave
iWave offers one of the industry’s most extensive product portfolios built on AMD’s Versal Adaptive SoC architecture. Our Versal System on Modules are engineered to address applications across AI inference, embedded compute, high-speed connectivity, advanced signal processing, and space/defense workloads.
Versal AI Edge Series SOM: Optimized for edge intelligence with AI Engines, low-latency processing, and energy-efficient compute.
Ideal for: Autonomous systems, edge analytics, satellite onboard processing, robotics.
Versal AI Edge Gen 2 Series SOM: Next-generation AI acceleration with enhanced ML density, higher performance-per-watt, advanced processing unit with multiple cores and improved sensor fusion capabilities. SOM incorporates four higher bandwidth LPDDR5x memory, high performance Ios and high-speed transceiver. Supporting multiple high-speed interfaces such as PCIe GEN5, 100G ethernet and HDMI through high speed SerDes.
Ideal for: Advanced satellite imaging, automotive ADAS, intelligent vision subsystems.
Versal Prime Series SOM: Designed for general-purpose computing with balanced DSP and logic resources, targeting compute-intensive applications and 100G Ethernet interfaces.
Ideal for: Industrial automation, networking, embedded computer platforms.
Versal Prime Gen 2 Series SOM: Higher throughput, improved I/O bandwidth, and extended feature set over Gen1 Prime devices. Targeting on embedded system with General-purpose compute, real-time sensor processing, and high-performance with LPDDR5x memory. The SOM facilitate high speed transceivers to support PCIe generation5 and high-speed networking interfaces.
Ideal for: Communications infrastructure, mid-range edge processing, flexible control systems.
Versal Premium Series SOM: Versal Premium SOM from iWave facilitates the ultra–high-bandwidth interfaces including 600G Ethernet fabrics and PCIe Gen5 interconnects by leveraging the VP1502/VP1802 class devices. It supports exceptional high-speed I/O capability through integrated 112G PAM4 GTM transceivers, hardened protocol accelerators, and advanced networking subsystems, enabling next-generation data-plane throughput and latency-optimized processing.
Ideal for: Data center acceleration, 5G/6G transport, high-bandwidth satellite communication.
Versal AI Core Series SOM: AI Engines paired with balanced programmable logic for ML acceleration + DSP workloads.
Ideal for: Radar processing, sensor fusion, high-speed signal intelligence.
Versal RF Series SOM: Integrated RF-ADC and RF-DAC chains for direct-RF sampling and real-time signal processing.
Ideal for: SATCOM payloads, ground-based feeder links, EW/ISR missions, high-speed RF chains.
As satellite missions become more complex, onboard intelligence is no longer optional, it is essential. The Versal AI Edge XQRVE2302 Adaptive SoC, combined with iWave’s rugged System on Modules, enables smarter satellites, reduced bandwidth usage, faster mission responsiveness, AI acceleration directly in orbit, and reconfigurable capabilities for evolving mission needs. The future of space is autonomous, AI-driven, and adaptive and Versal AI Edge is at the heart of this transformation.
iWave is a global leader in the design and manufacturing of FPGA System on Modules and ODM Design Services. With over 26 years of diverse experience in the FPGA domain and a strong design-to-deployment competence, iWave strives to transform your ideas into time-to-market products with reliability, cost, and performance balance.
Looking for more insights? Contact us at mktg@iwave-global.com
