MX3500A-SPMXM 3.1 Type B module with NVIDIA RTX 3500 Ada Generation Embedded GPU
- Manufacturer
- Configurations
- 1 available configurations
MX3500A-SP is a graphics module in MXM 3.1 Type B format built on the NVIDIA Ada Lovelace architecture. It integrates into carrier boards and embedded systems where professional graphics and AI inference must fit within an 82.0 x 105.0 mm footprint, not a full-size slot.
Two axes set it apart: the PCI Express 4.0 x16 bus, with faster sequential transfers than Gen 3, and GDDR6 memory on a 192-bit interface with 432 GB/s of bandwidth. API support covers DirectX 12 Ultimate, Vulkan 1.2 and OpenGL 4.6 on Windows 10/11 64-bit, with digital resolution up to 7680 x 4320.
Available configuration: MX3500A-SPA-B1 (MXM-B, NVIDIA RTX 3500 Ada, Gen4.0 x16, 12GB GDDR6, DP, 115 W, 0 to +55 °C).
At a glance
- 23.04 TFLOPS FP32 peak
- GDDR6 memory with ECC support
- Total Graphics Power 115 W
- Optional conformal coating for rugged applications
- CE, FCC and UKCA certifications
Technical specifications
| Graphics | NVIDIA RTX™ 3500 Ada Generation Embedded GPU |
|---|
| Memory | 12GB GDDR6 |
|---|
| Architecture standard | MXM |
|---|---|
| Board function | GPU / Graphics |
| Bus interface | PCIe Gen 4 x16 interface |
| Power consumption | 115 W |
|---|
| Dimensions | 82.0 x 105.0 |
|---|
| Operating temperature | 0–55 °C |
|---|---|
| Storage temperature | -40–85 °C |
| Operating systems | Windows |
|---|
| Form factor | MXM Type B |
|---|
Data extracted from the manufacturers' technical sheets: the official product spec sheet remains the reference for final confirmation.
Family configurations
The order codes of this family: the sheet applies to all of them, the detail of each configuration is in the datasheet.
| Configuration (PN) |
|---|
| MX3500A-SPA-B1 |
Documentation
In the same family
More products from the same series: select up to 3 and compare them side by side.
In the same category
Let's talk about your project
Phone and email go straight to our technical sales team in Turin: the first answer is already a technical one.