UBB Store

NVIDIA Tesla V100 16G/32G SXM2 GPU, AI LLM Training Deep Learning Compute Card For Sale

When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.

NVIDIA Tesla V100 16G/32G SXM2 GPU, AI LLM Training Deep Learning Compute Card

Price: CN $349.00

NVIDIA Tesla Referral V100 16G/32G SXM2 GPU NVIDIA Tesla Referral V100 16G/32G SXM2 GPU AI LLM Training Deep Learning Compute Card Why Buy This ✓ Custom finished Tesla Referral V100 SXM2 16GB /32GB with Wurth 4090 heatsink + adapter board ✓ Excellent air cooling, full load 300W keeps temp ~50°C, low noise ✓ SXM2 to PCIe conversion, Dual 8Pin power, for AI training & inference ✓ Fully tested, 1-month warranty, real item photos Product Overview

This custom finished NVIDIA Tesla Referral V100 SXM2 compute card includes V100 GPU core, Wurth 4090 style cooling heatsink and SXM2 to PCIe adapter board. Optimized air cooling design can maintain temperature around 50°C under full 300W load with quiet operation, great for home AI LLM training, deep learning and model inference. This card has NO video output, must use iGPU or separate boot card for display. Download official NVIDIA driver directly after installation.

Key Features

GPU Model: NVIDIA Tesla Referral V100 SXM2, available in 16GB /32GB HBM2
Assembly: V100 core + Wurth 4090 radiator + SXM2 to PCIe adapter board
TDP: 300W, Power: Dual 8Pin PCIe power connector
Cooling performance: ~50℃ at full load, ultra quiet, perfect for home workstation
Application: LLM training, deep learning, AI inference, HPC computing
No display output, requires iGPU or separate boot GPU for boot & display

Specifications Item Detail GPU Tesla Referral V100 SXM2 16GB /32GB Dimension Length:35cm, Thickness:6.5~6.8cm, occupies 3.5~3.8 PCIe slots Power 300W, Dual 8Pin Interface SXM2 converted to PCIe Motherboard Requirement Support PCIe3.0 & Above4G Decoding Important Pre-Purchase Notes ⚠ Please read carefully before purchasing:

✅ This custom Tesla Referral V100 card has NO video output. You need iGPU or separate boot card for display.
✅ Leave minimum 2cm gap between multi GPUs for proper cooling, insufficient space will cause overheating or performance loss.
✅ Intel 11th~14th Gen CPUs may encounter driver drop issue, remote technical support available to resolve.
✅ On Windows system, do NOT use Intel iGPU + NVIDIA discrete card combination, this will cause driver crash.
✅ Requires motherboard to support PCIe3.0 and Above4G Decoding in BIOS.
✅ Official driver available from NVIDIA website, remote setup guidance offered if needed.
✅ This is custom modified compute GPU, photos show actual item.

Item Condition Used SXM2 GPU card, fully tested, 1 month warranty, for AI model training & inference.

Used custom assembled Tesla Referral V100 SXM2 16G/32G compute card
Full function tested before packing, 1 month functional warranty
Custom air cooling assembly, low noise for home workstation
Compute-only card without video output port

Package Includes

1 × Custom Tesla Referral V100 16G/32G finished GPU card (V100 core + heatsink + adapter board)
Power cables, motherboard, boot card NOT included

Warranty & Returns Warranty 1-month functional warranty Returns Accepted only for verified hardware defects Pre-Shipment Check Full function test before dispatch Shipping