AI-Ready
XDNA2 NPU up to 50 TOPS. Built for local AI inference, Copilot+ workloads, and edge deployment — without a cloud dependency.
AMD Ryzen AI 300 Series. 50 TOPS NPU. 117mm.
XDNA2 NPU up to 50 TOPS. Built for local AI inference, Copilot+ workloads, and edge deployment — without a cloud dependency.
Magnetic tool-less top cover. Swap RAM and storage in seconds — no screwdriver needed. Built to grow with your workload.
117mm chassis. 12 cores. Radeon 890M graphics. VESA-mountable anywhere. The power of a full workstation in your palm.
AI 340
AMD Ryzen AI 5 340
AI 350
AMD Ryzen AI 7 350
AI 370
AMD Ryzen AI 9 HX 370
XDNA2 Architecture
The XDNA2 neural processing unit delivers up to 50 TOPS of dedicated AI compute — fully isolated from your CPU and GPU so your workloads never compete for resources.
Tool-less Design
One magnetic snap and the top cover lifts off. No screwdrivers, no tools — just direct access to every upgrade slot you need.
Connectivity Suite
USB4 at 40Gbps. Dual 4K@144Hz display output. 2.5G wired networking. Full-speed wireless. The Sapphire Edge connects to your entire ecosystem without compromise.
| Specification | AI 340 | AI 350 | AI 370 |
|---|---|---|---|
| CPU | Ryzen AI 5 340 | Ryzen AI 7 350 | Ryzen AI 9 HX 370 |
| Cores / Threads | 6C / 12T | 8C / 16T | 12C / 24T |
| Boost Clock | 4.8 GHz | 5.0 GHz | 5.1 GHz |
| Integrated GPU | Radeon 840M | Radeon 860M | Radeon 890M |
| NPU | 50 TOPS | 50 TOPS | 50 TOPS |
| Memory Type | DDR5 SO-DIMM | DDR5 SO-DIMM | DDR5 SO-DIMM |
| Max Memory | 96 GB | 96 GB | 96 GB |
| Storage Slots | 3× M.2 NVMe | 3× M.2 NVMe | 3× M.2 NVMe |
| Display Output | Dual HDMI 2.1 | Dual HDMI 2.1 | Dual HDMI 2.1 |
| Max Resolution | 4K @ 144Hz | 4K @ 144Hz | 4K @ 144Hz |
| USB | USB4 × 2 + USB-A | USB4 × 2 + USB-A | USB4 × 2 + USB-A |
| Network | 2.5G LAN + WiFi 6E | 2.5G LAN + WiFi 6E | 2.5G LAN + WiFi 6E |
| Bluetooth | 5.3 | 5.3 | 5.3 |
| Dimensions | 117 × 111 × 30mm | 117 × 111 × 30mm | 117 × 111 × 30mm |
| Power Adapter | 120W | 120W | 120W |
Render 4K timelines, AI-upscale footage in real time, and run Stable Diffusion locally — all from a device that fits behind your monitor.
Run local LLMs, containerize AI pipelines, and deploy edge inference models — with 50 TOPS NPU doing the heavy lifting natively.
Slim profile for digital signage, reception kiosks, or a compact hot-desk workstation. VESA-mountable to keep surfaces clear.
Light gaming on Radeon 890M, a compact home setup that respects your desk space, and full Linux compatibility out of the box.
Wishlist is empty.