QNAP Edge AI Accelerator NPU 1GB M.2 PCIe Gen 2 QAI-M100

The QNAP QAI-M100 is a compact, high-efficiency M.2 2280 Edge AI Accelerator designed to easily upgrade the AI computing capabilities of QNAP NAS systems. Outfitted with an integrated Neural Processing Unit (NPU) and 1GB of dedicated memory, this card drops directly into internal M.2 slots to handle complex AI computations. By offloading intensive processing from the host CPU, it enables budget-conscious small-to-medium businesses and home power users to maximize performance for smart facial recognition, intelligent surveillance, and automated multimedia classification.

• Dedicated NPU Offloading: Provides up to 3 TOPS of dedicated AI processing power, significantly decreasing NAS CPU utilization and freeing system resources for core storage tasks.
• Intelligent Media Acceleration: Drastically boosts QNAP QuMagie image recognition processing speed, delivering up to a 36% improvement in object recognition and up to 22% in facial recognition.
• Scalable AI Architecture: Supports multi-card configurations, allowing users to deploy up to two QAI-M100 units simultaneously inside compatible NAS systems to double total AI capacity to 6 TOPS.
• Advanced Graphene Cooling: Shipped with a pre-installed premium graphene heatsink and thermal pad to guarantee efficient heat dissipation and prevent thermal throttling during continuous AI operations.
• Flexible Resource Allocation: Integrates seamlessly into the QTS/QuTS hero ecosystems, allowing administrators to prioritize and allocate NPU computing resources to specific mission-critical applications.

Specifications

• Controller: Rockchip RK1808 (Dual-core Cortex-A35, 1.6GHz)
• AI Processing Power: 3 TOPS NPU
• Onboard Memory: 1 GB
• Host Interface: M.2 2280 PCIe Gen 2 x1 (M+B Key)
• Form Factor: M.2 2280 internal expansion card
• Thermal Management: Pre-installed graphene heatsink and thermal pad included
• System Requirements (OS): QTS 5.2.1.2930 build 20241025 (or later) / QuTS hero 5.2.1.2929 build 20241025 (or later)
• Application Prerequisites: QNAP AI Core v3.5.0, Multimedia Console v2.7.0, QuMagie v1.5.1 (or later)