DeGirum

Orca M.2 AI Accelerator Module

The DeGirum Orca M.2 module delivers high-performance edge AI acceleration in a compact PCIe® form factor

Image of DeGirum Orca M.2 AI Accelerator ModuleThe DeGirum Orca M.2 accelerator module empowers developers to build advanced edge AI solutions with cloud-like performance. Designed in the popular M.2 2280 form factor, Orca provides powerful neural network acceleration while maintaining low power consumption and high efficiency.

With support for 1 GB LPDDR4 DRAM, PCIe Gen3 x4 connectivity, and pruned model optimization, Orca ensures fast, reliable inference across multiple models simultaneously. Its highly efficient architecture enables real-time operation for multi-camera systems and complex AI workloads. Developers can also seamlessly integrate with DeGirum PySDK and AI Hub, ensuring a plug-and-play experience for rapid deployment.

Features
  • M.2 M-Key 2280 form factor with PCIe Gen3 x4 interface
  • Typical power consumption: 2.8 W to 3.5 W (max. 5 W workload-dependent)
  • Supports pruned models and near-lossless multi-model switching
  • Integrated 1 GB LPDDR4 DRAM for complex workloads
  • Operates across -10°C to +70°C with thermal management options
  • Scalable to multi-chip configurations for demanding applications
Applications
  • AI at the edge for embedded PCs and industrial systems
  • Multi-camera vision analytics
  • Robotics and autonomous systems
  • Smart retail and IoT gateways
  • Low-power real-time AI inference at the network edge
發佈日期： 2025-09-18