DFRobot
LattePanda Mu RTX 3060 MXM GPU Carrier Kit
· MPN: KIT0206
This compact carrier kit bridges a LattePanda Mu compute module with a discrete MXM graphics card, giving embedded projects a ready-to-deploy platform for lo...
This compact carrier kit bridges a LattePanda Mu compute module with a discrete MXM graphics card, giving embedded projects a ready-to-deploy platform for local AI inference, 3D rendering and high-density visual output. It is designed for space-constrained builds such as display systems, kiosk controllers, interactive installations and edge AI prototypes.
This RTX 3060 version includes an NVIDIA GeForce RTX 3060 MXM graphics card with 12GB VRAM, pre-assembled with a cooling fan. With video outputs from both the LattePanda Mu iGPU and the MXM dGPU, the kit can drive up to six independent 4K@60Hz displays for video walls, signage and control-room style applications.
The kit includes the LattePanda Mu GPU carrier board, RTX 3060 MXM graphics card, metal enclosure, power adapter, three AC power cords, 12mm metal LED push-button switch, screw pack, CR1220 RTC battery and thermal pad. The LattePanda Mu compute module is required and is not included.
Important: Never connect the DC jack and ATX 8-Pin power supply at the same time, as this will permanently damage the board. Check whether your MXM fan is 5V or 12V and set the PWR_FAN jumper correctly before powering on.
Features:
- Compact-form-factor carrier board for integrating a discrete MXM graphics card with a LattePanda Mu compute module.
- Purpose-built for local AI inference, including LLMs, VLMs and GenAI workloads.
- Suitable for robust 3D rendering and game-engine-based Unity / Unreal interactive installations.
- Supports up to six independent 4K displays.
- 6 video outputs in total: LattePanda Mu 2-channel plus MXM GPU 4-channel.
- Each channel supports 4K@60Hz hardware decoding.
- NVIDIA GeForce RTX 3060 MXM GPU included for high-performance workloads.
- RTX 3060 configuration provides 12GB VRAM.
- Capable of running INT4-quantized LLMs up to around 9B parameters, depending on the framework and runtime configuration.
- Compatible with all LattePanda Mu compute modules.
- Supports Intel N100 or Core i3-N305 LattePanda Mu processor modules.
- Decoupled architecture lets CPU compute scale independently from the MXM GPU selection.
- Suitable for high-density digital signage and video walls.
- Suitable for localised edge AI inference.
- Suitable for interactive media installations.
- Suitable for multi-stream industrial machine vision.
Specifications:
- Compute Module: LattePanda Mu Slot
- GPU Module: MXM 3.0/3.1 (Type A/B Compatible)
- HDMI: 3 x HDMI, Each supports 4K@60Hz (Mu iGPU 2-ch + MXM dGPU 1-ch)
- DisplayPort(DP): 3 x DP, Each supports 4K@60Hz (All from MXM dGPU)
- M.2 M Key: 2230, PCIe 3.0 x1
- M.2 E Key: 2230, PCIe 3.0 x1 + USB2.0
- F_Front: PWRBTN, RESET, PLED (Standard 2.54mm Pitch)
- CPU Fan: PH2.0-4P
- MXM Fan: PH2.0-4P & PH2.5-4P
- PWR_FAN Jumper: 2.54mm Pin Header (5V/12V selectable)
- RTC: CR1220 battery holder
- USB: 2 x USB 3.2 Gen2 + 2 x USB 2.0
- Ethernet: 1 x RJ45, 10/100/1000Mbps Auto-Negotiation
- Carrier Board Size: 146 x 102mm
- Overall Size: 159 x 107.5 x 88.5mm
Use it where a full desktop GPU system is too large, but you still need serious edge compute, multi-display output or local visual AI capability.
Jargon buster
Plain-language definitions for the technical terms used above.
- DisplayPort
- A digital display interface used to connect monitors and other video displays. It matters because the supported DisplayPort version affects display resolution, refresh rate and monitor compatibility.
- HDMI
- HDMI is a common digital video and audio connection used by computers, media players, and many displays. If a display kit has HDMI input, it is usually much easier to test with a single-board computer because it can act like a normal monitor.
- LED
- A light-emitting diode is a small electronic component that lights up when current flows through it in the correct direction. In this kit, LEDs create the flashing effect, so polarity and correct soldering matter for the project to work.
- M.2
- A compact edge-connector format commonly used to plug small modules into a carrier board without soldering. On this product it is the physical connector used by the MicroMod system, so compatibility with the matching processor board is important.
- PCIe 3.0 x1
- A high-speed internal computer expansion connection using one PCI Express data lane. It matters because the expansion board must physically and electrically match the LattePanda IOTA’s PCIe connector to work.
- RJ45
- The common plug and socket style used for wired Ethernet network cables. If a board has an RJ45 connector, you can usually plug it into standard Ethernet cabling without making a custom connector.
- RTC
- A Real-Time Clock keeps track of time even when the main processor is asleep or powered down, usually with a small backup battery. It matters for data logging and tracking projects that need accurate timestamps.
Find this product in
