Search  for anything...
HP

HPE NVIDIA Tesla V100 32GB HBM2 PCIe 3.0 x16 Passive GPU Computational Accelerator for AI Machine Learning HPC Deep Learning 699-2G500-0216-400 (Renewed)

  • Based on 16 reviews
Condition: Refurbished - Excellent
Checking for the best price...
$859.00 Why this price?
Save $195.96 was $1,054.96

Buy Now, Pay Later

As low as / mo
  • – Up to 36-month term if approved
  • – No impact on credit to apply
  • – Instant approval decision
  • – Secure and straightforward checkout

Ready to go? Add this product to your cart and select a plan during checkout.

Payment plans are offered through our trusted finance partners Klarna, Affirm, Afterpay, Zip, Apple Pay, and Google Pay. No-credit-needed leasing options through Acima may also be available at checkout.

Learn more about financing & leasing here.

Free shipping on this product

FREE 90-day refund/replacement

To qualify for a full refund, items must be returned in their original, unused condition. If an item is returned in a used, damaged, or materially different state, you may be granted a partial refund.

To initiate a return, please visit our Returns Center.

View our full returns policy here.

Availability: Only 10 left in stock, order soon!
Fulfilled by Amazon
Arrives Oct 30 – Nov 4
Order within 10 hours and 31 minutes
Available payment plans shown during checkout
Protection Plan Protect Your Purchase
Checking for protection plans...

Recently Viewed

Features

  • NVIDIA Volta GV100 Architecture 4,608 CUDA Cores, 640 1st-Gen Tensor Cores delivering 14 TFLOPS FP32 and 112 TFLOPS deep learning performance for AI training, inference, HPC, and scientific computing workloads
  • 32GB HBM2 ECC Memory 900 GB/s Bandwidth High-bandwidth memory on a 4096-bit bus with ECC error correction provides the memory capacity and throughput required for the largest AI models, simulations, and datasets
  • PCIe 3.0 x16 Interface 250W TDP Standard PCIe Gen3 connectivity with passive cooling designed for enterprise rack server deployment in HPE ProLiant, Dell PowerEdge, and Supermicro platforms with adequate chassis airflow
  • NVLink Scale to 96GB Unified Memory Connect two V100 GPUs via NVLink at 300 GB/s bi-directional bandwidth to scale GPU memory from 32GB to 96GB for larger AI training and HPC workloads
  • Multi-Precision Computing Supports FP64 (7 TFLOPS), FP32 (14 TFLOPS), FP16 (112 TFLOPS) and INT8 precision modes for flexible deployment across training, inference, and scientific simulation workloads
  • HPE OEM Validated P44861-001 Also known as HPE Part No P44861-001 (Spare No P44899-001), validated for deployment in HPE ProLiant servers including DL380 Gen10, DL360 Gen10, and other PCIe Gen3 x16 compatible platforms

Description

The HPE NVIDIA Tesla V100 PCIe 32GB is an enterprise-grade computational accelerator purpose-built for artificial intelligence, machine learning, deep learning, and high-performance computing workloads. Powered by NVIDIA's Volta GV100 architecture with 4,608 CUDA cores and 640 first-generation Tensor Cores, the V100 delivers 14 TFLOPS of FP32 performance and 112 TFLOPS of deep learning performance — making it one of the most capable data center GPUs available for AI training and inference. Equipped with 32GB of high-bandwidth HBM2 memory on a 4096-bit bus delivering 900 GB/s of memory bandwidth, the V100 handles the largest AI models and datasets with ease. The PCIe 3.0 x16 interface and passive cooling design make it compatible with a wide range of enterprise rack servers including HPE ProLiant, Dell PowerEdge, and Supermicro platforms. NVLink support enables two V100 GPUs to be connected for 96GB of unified memory at 300 GB/s bi-directional bandwidth for even larger workloads. Backed by NVIDIA's proven Volta architecture and HPE enterprise validation for mission-critical data center deployments requiring maximum AI compute density and reliability.

Graphics Coprocessor: NVIDIA Tesla v100


Brand: HP


Graphics Ram Size: 32 GB


GPU Clock Speed: 1380 megahertz


Graphics Processor Manufacturer: NVIDIA


Graphics Coprocessor: NVIDIA Tesla v100


Graphics Card Ram: 32 GB


GPU Clock Speed: 1380 megahertz


Graphics Ram Type: HBM2


Compatible Devices: Enterprise Rack Server


Graphics Card Interface: PCI-Express x16


Memory Clock Speed: 877 MHz


Item Dimensions L x W: 10.5"L x 1.4"W


Brand: HP


Video Processor: NVIDIA


Built-In Media: 1


Model Name: Tesla V100


Graphics Description: NVIDIA Tesla v100-32GB


Unit Count: 1 Count


Number of Packs: 1


Manufacturer: HP


Number of Items: 1


UPC: 728434281566


Item Weight: 1.1 pounds


Mfr Part Number: 699-2G500-0216-400


Warranty Description: 90 days limited warranty


Frequently asked questions

If you place your order now, the estimated arrival date for this product is: Oct 30 – Nov 4

Yes, absolutely! You may return this product for a full refund within 90 days of receiving it.

To initiate a return, please visit our Returns Center.

View our full returns policy here.

  • Klarna Financing
  • Affirm Pay in 4
  • Affirm Financing
  • Afterpay Financing
  • Zip Pay in 4
  • Financing through Apple Pay
  • Financing through Google Pay
Leasing options through Acima may also be available during checkout.

Learn more about financing & leasing here.

Similar Products

Top Amazon Reviews

    • Great for inference workload
This GPU works great with Ollama. You need to run older driver 580, legacy (properiety, not open) kernel module. So no mixing Tesla gpu and newer GPUs like Ada Lovelace and Blackwell. Note: newer driver and open driver only support Volta and later so it wont recognize this Tesla GPU.
Reviewed in the United States on September 22, 2026 by neofob
    • Works great for AI for cheap.
Review Image by Amazon Customer
A++ But I had a difficult time getting this to work with OCuLink under Windows with my Ryzen7 8845HS mini-pc. I am using Nvidia data center drivers version 551.78. I had read that most people give up and use Ubuntu instead. I completely understand, as it took me 4 days to get this sorted out. The key to getting this working, assuming your BIOS has support for Large BAR reporting, is to set the default registry value of Computer\HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\nvlddmkm\Parameters\DmaRemappingCompatible to 0 in order for Device Manager to successfully install the driver and not complain about a code 10 (Insufficient resources). I am using Qwen3.8-Flash-Next Coder on a 32GB system memory PC with Strata. This combination is hand's down the best quality, best reasoning, most efficient, and fastest I have found for the type of work I want to do. The power draw with this model is consistently only 150W, where other models, like Qwen 3.8-27B peaks near 250W and most of the time around 215W and is also slower. This means easier cooling/less fan noise. I am getting 80+ tokens/second on average, with prefill speeds approximately 900 tokens/second. And this stays relatively consistent even when context usage gets high. You can have 128K context with this model and this card. All the experts stay in VRAM like 99.8% of the time. By far, this is the best bang for your buck if you want to experiment with a local LLM and have reasonable speed without going broke. The card I got looked in great condition, and I haven't had any issues with it. ... show more
Reviewed in the United States on October 6, 2026 by Amazon Customer
    • Defective
Review Image by daniel
The GPU appears to be defective. When I install the Tesla V100 as a second or third GPU, the computer completely fails to power on. The system works normally when the Tesla V100 is removed. This suggests a possible hardware-level fault with the GPU, potentially involving the PCIe interface, power circuitry, or another component on the card. I cannot determine whether the CUDA cores or HBM2 memory are defective because the system cannot boot with the GPU installed, so I am unable to run CUDA or memory diagnostics. I have also verified that the computer has sufficient power capacity. The main issue is that installing the Tesla V100 prevents the computer from starting at all, it work with other gpus. ... show more
Reviewed in the United States on October 3, 2026 by daniel
    • Simply the BEST
Seller is awesome! Card works great. I'm running 2: Qwen3-Coder-Next 80B replacement — VERIFIED COMPLETE Official Qwen3-Coder-Next 80B A3B Q4_K_M is installed and activated as the only curated local model on server 1.6.33, split 1:1 across the two 32 GiB V100s. Qualified context options are 32,768 and 258,000 tokens. The model is text-only. Codex's default provider remains OpenAI and default model remains gpt-6.1-sol. ## Verified results | Check | 32K / 150W per card | 258K / 250W per card | |---|---:|---:| | Actual long input | 28,167 tokens | 249,861 tokens | | Exact retrieval | 6 / 6 | 6 / 6 | | Generated Python function cases | 5 / 5 | 5 / 5 | | Exact synthetic tool call | Passed | Passed | | Prompt processing | 492.75 tokens/sec | 354.95 tokens/sec | | Long-input decoding | 67.83 tokens/sec | 31.98 tokens/sec | | Long request elapsed | 59.22 sec | 708.57 sec | | Peak GPU temperature | 46 C | 50 C | | Peak aggregate VRAM | 46.43 GiB | 50.46 GiB | | OOM, thermal, latency gates | Passed | Passed | | API restored / boot unchanged | Verified | Verified ... show more
Reviewed in the United States on October 7, 2026 by Jamie H.
    • Do your homework
I'm still struggling to get it to work. Even if things tell you it will, it's not that straightforward on a Dell 5820. You need power cables, a 3-D printed fan shroud, a fan, and the know how.
Reviewed in the United States on September 23, 2026 by Tyson Kreger
    • Cheapest way to 32 GB LLM support
At present these are going for $700, so they are the cheapest way to support a local LLM of a reasonable size. You have to run the 580 stack of CUDA though, so you are stuck with an older version. If Intel brought the cost of the Intel Arc Pro B70 back down to the initial price, I would probably say go that way at current Nvidia markups. Running Qwen 3.8 on Ollama well and recently Qwen 3.8 Next Swift hosted on Strata at good TPS. Draw back is adding noisy add on fans but the card stays cool with no additional heat sink. So like $750 all in. ... show more
Reviewed in the United States on October 4, 2026 by James H.
    • Great for advanced Starters in AI
This was a great starter, moved up to something faster but it was 4x the cost. Easy to use on older equipment (PCIE 3), make sure you have power supply with extra 8port cpu power for 300w of power and easy connection.
Reviewed in the United States on August 10, 2026 by Patrick Chatelle

Can't find a product?

Find it on Amazon first, then paste the link below.
Checking for best price...