Edge inference appliances
Compact edge inference boxes and small-form-factor AI servers for retail, smart city, and branch deployments — GPU or NPU options, wired networking, and export-ready chassis programs.
3 representative models — browse below or search by model. Carrier compatibility, SDK, and export MOQ confirmed on inquiry.
Edge AI box PCs
Ready-to-deploy edge inference appliances — integrated SoC, NPU, industrial I/O, and optional M.2 AI expansion for vision, LLM, and smart-city programs.
APC880
NXP i.MX8M Plus edge AI box with onboard 2.3 TOPS NPU and M.2/B2B slots for optional accelerator modules up to 40 TOPS — fanless industrial I/O for factory and transportation vision.
- NXP i.MX8M Plus · 2.3 TOPS onboard NPU
- Optional M.2 or B2B AI module · up to 40 TOPS
- Dual GbE · CAN · RS232/485 · 6× USB
- 4K HDMI · dual MIPI-CSI · fanless enclosure
- Industrial -40°C to +85°C option
- CPU: 4× Cortex-A53 up to 1.8 GHz + Cortex-M7
- Onboard AI: 2.3 TOPS NPU
- Expansion AI: M.2 Key M + B2B · optional module up to 40 TOPS
- Memory / storage: 4 GB RAM (8 GB opt.) · 64 GB eMMC (up to 128 GB opt.)
- Network: 2× Gigabit Ethernet · optional Wi-Fi / 4G
- Operating system: Yocto Linux
- Power: DC 12 V / 3 A
- Dimensions: 238 × 131 × 37 mm
- Operating temperature: 0°C to 60°C commercial · -40°C to 85°C industrial
- Supplier program: Geniatech
- MOQ: Confirm on inquiry
APC3588-AI
Rockchip RK3588 edge AI box with 6 TOPS onboard NPU and rich AV/industrial I/O — local LLM and multi-channel vision for NVR, signage, and control applications.
- RK3588 octa-core · 6 TOPS onboard NPU
- Optional built-in M.2 AI accelerator module
- Triple GbE · dual HDMI out · HDMI in
- RS232/485 · CAN · GPIO · 7×24 watchdog + RTC
- Android 13 or Debian 11
- SoC: Rockchip RK3588
- CPU: Quad Cortex-A76 + quad Cortex-A55 up to 2.4 GHz
- GPU: Arm Mali-G610 MP4
- Onboard AI: 6 TOPS NPU
- Expansion AI: Optional M.2 AI accelerator module
- Memory / storage: 4 GB LPDDR4 (up to 32 GB opt.) · 32 GB eMMC (up to 256 GB opt.)
- Network: 3× Gigabit Ethernet
- Operating system: Android 13 · Debian 11
- Power: AC 220 V 50 Hz
- Dimensions: 247 × 230 × 55 mm
- Operating temperature: 0°C to 60°C commercial · -40°C to 85°C extended
- Supplier program: Geniatech
- MOQ: Confirm on inquiry
APC888
NXP i.MX95 generative AI edge box with integrated NPU and M.2 slot for accelerator modules up to 40 TOPS — scalable edge AI for industrial and commercial deployments.
- NXP i.MX95 · integrated eIQ Neutron NPU
- M.2 AI module slot · up to 40 TOPS expansion
- 4 GB LPDDR5 · 32 GB eMMC (up to 128 GB opt.)
- Dual GbE · rich MIPI/CAN/USB/UART I/O
- Industrial -40°C to +85°C · Yocto 5.0 BSP
- CPU: 6× Cortex-A55 + Cortex-M7 + Cortex-M33
- Onboard AI: NXP eIQ Neutron NPU
- Expansion AI: M.2 slot · e.g. NXP Ara-240 up to 40 TOPS
- Memory / storage: 4 GB LPDDR5 · 32 GB eMMC (64/128 GB opt.)
- Video: 1080p60 H.264/H.265 encode & decode · HDR ISP
- Network: 2× Gigabit Ethernet · optional Wi-Fi / 4G / GPS
- Operating system: Yocto 5.0 Linux
- Power: DC 12 V / 1.5 A
- Dimensions: 203 × 174 × 67 mm
- Operating temperature: 0°C to 60°C commercial · -40°C to 85°C industrial
- Supplier program: Geniatech
- MOQ: Confirm on inquiry
Accelerator tier, I/O map, and enclosure depth — confirm on inquiry.
Priority export program — start with edge box and fanless appliance SKUs; accelerator modules are expansion-tier follow-on programs. Share workload, environment, certification, and MOQ for early matching.