NVIDIA Data Center GPU Tesla V100 SXM2 32GB
The NVIDIA Tesla V100 SXM2 32GB was the pioneer enterprise deep learning accelerator based on the Volta GV100 architecture, introducing the world to Tensor Cores. Upgraded from the original 16GB model to 32GB of ultra-fast HBM2 memory delivering 900 GB/s bandwidth, it features 5,120 CUDA cores and 640 1st-generation Tensor Cores. Connected via 2nd-generation NVLink with 300 GB/s interconnect throughput on SXM2 baseboards (such as NVIDIA DGX-1 and HGX-1), it powered the first major wave of commercial deep learning and large-scale NLP research. In the refurbished market, it is widely traded as an economical accelerator for homelabs and budget research servers. Buyers should inspect SXM2 gold pad integrity.
Also written as: NVIDIA Tesla V100 32GB SXM2, V100 SXM2, Tesla V100 32G
Specifications
| Brand | NVIDIA |
|---|---|
| Family | Data Center GPU |
| Form factor | SXM2 module |
| Released | Mar 2018 |
| End of sale | Sep 2022 |
| Gpu memory (GB) | 32 |
| Tdp (W) | 300 |
Generations
- Next generation: NVIDIA Data Center GPU A100 SXM4 80GB
Reference prices (last 14 days, updated 2026-10-27)
Refurbished: USD 1,850 (1 sighting, Oct 2026)
Used: USD 1,250 (1 sighting, Oct 2026)
All sightings
| Seller | Condition | Price | Note | Seen |
|---|---|---|---|---|
| ebay.com | used | USD 1,250 | Tested pull from DGX-1, 100% functional | 2026-10-27 |
| serversupply.com | refurbished | USD 1,850 | Refurbished SXM2 module with 90-day warranty | 2026-10-27 |
Prices seen on public listings on the dates shown, before tax and shipping. Not an offer.
Buying a used Tesla V100 SXM2 32GB: what to check
Inspect the SXM2 surface-mount connector array for damage or dust, check HBM2 memory integrity with DCGM, verify that the 300W cooling heatsink or water block matches mounting specifications, and ensure NVLink bridges negotiate correctly.
FAQ
What was groundbreaking about the NVIDIA V100?
The V100 introduced hardware Tensor Cores, multiplying deep learning matrix computation speed by up to 5x over Pascal.
What form factor does the V100 SXM2 use?
It uses the high-density SXM2 mezzanine form factor designed for DGX-1 and HGX-1 8-GPU chassis.
What succeeded the Tesla V100?
The NVIDIA A100 Ampere accelerator succeeded the V100, bringing HBM2e memory and MIG.
Buying or selling Tesla V100 SXM2 32GB?
New, refurbished or used. Tell us the quantity and condition and we connect you with a vetted reseller in your region, usually within one working day.
Sources
- NVIDIA Tesla V100 Architecture Whitepaper, checked 2026-10-27