Specifications
Product
AICore DX-M1M
SoC
DEEPX DX-M1M
AI Performance
Up to 25 TOPS (INT8)
Form Factor
M.2 M + B Key, 2242 (22 × 42 mm)
Host Interface
PCIe Gen3 ×2 (host support up to Gen3 ×4, backward compatible)
Memory
1 GB LPDDR4X, up to 4266 MT/s
Storage
QSPI 1 Gbit NAND / NOR flash
Power Consumption
≈3 W typical
Framework Support
TensorFlow, ONNX, Keras, PyTorch via DX-COM compiler
Compiler / SDK
DXNN SDK (DEEPX)
Mechanical
22 mm × 42 mm
Thermal
Heatsink option
OS Support
Windows 11/10, Debian-based Linux (Ubuntu 24.04/22.04/20.04 LTS), Yocto, Docker
Target Applications
AIoT, smart surveillance, robotics, industrial automation
Advantage vs 2280
Short 2242 length fits compact embedded carriers that cannot take a 80 mm card
Sibling Modules
DX-M1 (M.2 2280, 4 GB LPDDR5), DX-M1+ (dual-chip 50 TOPS / quad-chip 100 TOPS)
Compliance
RoHS
Lead Time
4 weeks
MOQ
1 unit
Overview
The AICore DX-M1M packages the DEEPX DX-M1M SoC into an M.2 2242 card — 22 × 42 mm, roughly half the length of a standard M.2 2280 module. That short length is the point: many compact embedded carriers and industrial SBCs simply have no room for an 80 mm accelerator.
It delivers up to 25 TOPS of INT8 throughput over PCIe Gen3 ×2 at around 3 W typical, carrying 1 GB of LPDDR4X and 1 Gbit of QSPI flash on-board. Models are compiled through the same DEEPX DXNN toolchain as the DX-M1 family, supporting TensorFlow, ONNX, Keras and PyTorch.
For integrators already standardising on DEEPX silicon, this is the density variant: identical performance class and software stack, shorter card.
Key Benefits
M.2 2242 short form factor for compact embedded carriers; 25 TOPS at ~3 W for fanless designs; PCIe Gen3 ×2 host interface with Gen3 ×4 host support; DXNN SDK compatibility with the DX-M1 2280 family.
Applications
Compact AIoT gateways, smart surveillance cameras and recorders, robotics perception modules, industrial automation controllers, and any embedded carrier with a 2242-length M.2 slot.
Request a Quote — AICORE DX-M1M — 25 TOPS M.2 2242 EDGE AI MODULE ON DEEPX DX-M1M
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →