Specifications
APC3588-AI
RK3588 + Hailo-8 M.2 · on-premise local LLM · triple GbE · -40°C to +85°C
APC3588 / APC3588Q
Rockchip RK3588, 8-core 64-bit up to 2.4GHz, Mali-G610 MP4, 6 TOPS NPU, LPDDR up to 32GB, eMMC to 256-512GB, 8K decode, quad HDMI (Q variant)
APC885
NXP i.MX95 generative AI edge box with M.2 M-key AI accelerator slot · -40°C to +85°C
APC880
NXP i.MX8M Plus on-device AI edge computer with modular expansion · M.2 M-key
APC880 Mini
NXP i.MX8M Plus industrial edge AI inference box
APC880E
NXP i.MX8M Plus compact edge AI system with M.2 AI module
Video
32 channels 1080p30 decode; 8K encode/decode
Networking
Triple Gigabit Ethernet; Wi-Fi 6; 5G/4G LTE expansion
I/O
RS232, RS485, CAN, TTL, GPIO, USB 3.0/2.0, HDMI in/out, Type-C DP
Operating Temp
-40°C to +85°C; fanless aluminium enclosure
OS
Android 13, Debian 11, Buildroot, RTLinux
Overview
Geniatech's Edge AI Box family covers ARM-based industrial edge inference from entry-level NXP i.MX8M Plus nodes up to Rockchip RK3588 systems with a plug-in NPU accelerator. The headline model, the APC3588-AI, pairs an octa-core RK3588 with a Hailo-8 M.2 accelerator to run on-premise local LLM workloads with triple Gigabit networking in a fanless, wide-temperature (-40°C to +85°C) chassis. The APC885 moves to NXP's i.MX95 generative-AI SoC with an M.2 M-key AI accelerator slot, while the APC880, APC880 Mini and APC880E cover i.MX8M Plus inference at the low-power end. The industrial APC3588 and quad-HDMI APC3588Q provide the general-purpose embedded computing base, with up to 32GB LPDDR, 512GB eMMC and 32-channel 1080p30 video decode.
Key Benefits
Local LLM without the cloud: the APC3588-AI's Hailo-8 accelerator plus RK3588 NPU keeps generative inference and 32-channel video analytics on-premise, which matters for privacy, latency and sites with intermittent connectivity. One vendor across the ARM ladder: Geniatech spans NXP i.MX8M Plus, NXP i.MX95 and Rockchip RK3588 in the same industrial form-factor family, so a design can move up or down the performance ladder without changing enclosure, mounting or thermal approach. Industrial hardening: fanless aluminium construction, -40°C to +85°C operation, M.2 M-key accelerator expansion, RS232/RS485/CAN/TTL fieldbus I/O, Wi-Fi 6 and 5G/4G options. Accelerator-agnostic: the M.2 slot accepts Hailo, MemryX MX3, NXP Ara-240 or DEEPX accelerator modules, so the NPU can be swapped as model requirements change.Applications
Machine vision and industrial inspection, industrial automation and control, transportation and roadside analytics, retail and campus analytics, local LLM and private GPT appliances, smart healthcare, and multi-display digital signage.
Request a Quote — GENIATECH EDGE AI BOX SERIES — RK3588
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →