Mercury I represents a new class of connected intelligence designed for edge environments that demand secure, low-latency processing. This platform integrates multi-modal sensing, adaptive compute, and encrypted communication into a compact module.
Engineered for industrial IoT, smart infrastructure, and advanced driver-assistance scenarios, Mercury I balances strict power budgets with enterprise-grade reliability and real-time analytics.
| Dimension | Specification | Value | Notes |
|---|---|---|---|
| Form Factor | Module Type | PCIe-based COMe 3.0 | Compatible with SMARC 2.1 carrier boards |
| Processor | CPU Cores | 8 | ARM Cortex-A78AE, up to 2.4 GHz |
| AI Acceleration | NPU Throughput | 32 TOPS INT8 | Supports sparsity and mixed-precision |
| Memory | LPDDR5 Capacity | 32 GB | Error-corrected, bandwidth-optimized |
| Storage | eMMC / Flash | 64 GB | Expandable via PCIe NVMe |
| Security | Features | Secure Boot, TPM 2.0, HSM | Hardware root of trust |
| Connectivity | Interfaces | 5G, Wi-Fi 6E, BLE 5.3, 2x 10G Ethernet | Time-sensitive networking support |
| Sensors | On-board | IMU, Barometer, GNSS | Industrial temperature range |
| Power | Range | 12–48 VDC, typical 25 W | Dynamic power capping supported |
Architecture and Hardware Design
Compute Subsystem
Mercury I leverages a heterogeneous compute fabric with a primary cluster of ARM Cortex-A78AE cores and a dedicated real-time co-processor for safety-critical tasks. The design ensures deterministic latency for control loops and high-throughput data paths.
Edge AI Capabilities
The integrated NPU with 32 TOPS INT8 acceleration enables on-device inference for anomaly detection, computer vision, and predictive maintenance, minimizing cloud dependency and data exposure.
Connectivity and Networking
Built-in multimodal radios allow Mercury I to serve as a gateway across 5G, private LTE, and wired Ethernet, while time-sensitive networking guarantees prioritized traffic for industrial protocols.
Dual-band Wi-Fi 6E and BLE 5.3 support dense device deployments, and optional GNSS deliver global positioning for mobile and outdoor applications without external modules.
Security and Lifecycle Management
Mercury I implements a layered security model rooted in hardware, with secure boot, measured boot, and firmware authentication. A dedicated HSM safeguards keys used for over-the-air updates and encrypted storage.
Remote provisioning and attestation enable organizations to enforce compliance policies, rotate credentials, and retire devices through a centralized management console aligned with zero-trust principles.
Deployment and Integration
Mounting brackets, pre-certified carrier boards, and reference designs simplify integration into gateways, robots, and edge servers. Extensive diagnostics and health monitoring reduce field troubleshooting time.
Recommendations and Next Steps
- Evaluate carrier boards with redundant power paths for high-availability deployments.
- Leverage the security HSM to manage device identities and encrypted over-the-air updates.
- Use the on-board sensors for time-synchronized logging in condition monitoring scenarios.
- Plan network segmentation to isolate control traffic and telemetry using VLANs and QoS policies.
- Test edge AI models under real-world latency and thermal constraints before production rollout.
FAQ
Reader questions
What workloads is Mercury I optimized for at the edge?
Mercury I excels at video analytics, predictive maintenance, autonomous navigation, and distributed control, where low latency, high-throughput AI, and robust connectivity are essential.
How does Mercury I ensure data privacy in regulated industries?
On-device processing, encrypted storage, and hardware-rooted trust minimize data movement, while fine-grained access controls and audit logs support compliance with GDPR, HIPAA, and sector-specific mandates.
Can Mercury I operate in harsh environmental conditions?
Yes, the module is qualified for extended industrial temperature ranges, with conformal coating options and validated vibration and shock profiles for demanding deployments.
What tools are available for developers building on Mercury I?
Comprehensive SDKs, container runtimes, and a simulation framework provide reproducible builds, secure pipelines, and remote debugging, enabling rapid iteration and continuous delivery at the edge.