Enterprise AI Computing GPU NVIDIA Tesla Referral V100 16GB PNY Triple-Fan GPU Volta Architecture — 5,120 CUDA Cores, 640 Tensor Cores, 16GB HBM2 — AI & Deep Learning Accelerator Why Buy This ✓ 16GB HBM2 with 900 GB/s bandwidth — handle large models and datasets without VRAM shortage ✓ Triple-fan cooling — PNY active cooling variant, ideal for standalone / workstation use ✓ Nearly new— tested & fully functional ✓ Cost-effective — enterprise-grade AI acceleration at a fraction of retail price What You'll Get

An enterprise-grade GPU accelerator built by NVIDIA for serious AI, deep learning and high-performance computing workloads. With 16GB of HBM2 memory at 900 GB/s bandwidth, you can train and run medium-to-large neural networks, accelerate inference, and handle scientific computing — all at a competitive price.

This PNY triple-fan version features active cooling, making it more suitable than the standard passive data-center variant for standalone workstations and server chassis with limited airflow. It has been kept in nearly new condition.

Powered by the NVIDIA Volta architecture with 5,120 CUDA cores and 640 Tensor cores, it delivers up to 14 TFLOPS (FP32) and 112 TFLOPS (Tensor) of compute performance. Standard PCIe 3.0 x16 interface works with mainstream x86 servers and workstations, and supports Windows and Linux.

Key Features

• NVIDIA Tesla Referral V100 with 16GB HBM2 high-bandwidth memory
• Volta architecture — 5,120 CUDA cores & 640 Tensor cores
• 900 GB/s memory bandwidth on a 4096-bit interface, with ECC support
• PNY triple-fan active cooling — better airflow & thermal performance than passive variants
• Standard PCIe 3.0 x16 interface, fits servers & workstations
• Ideal for AI training, deep learning, scientific computing & demanding compute tasks

Specifications Item Detail GPU NVIDIA Tesla Referral V100 (PNY triple-fan version) Architecture NVIDIA Volta (GV100) CUDA / Tensor Cores 5,120 CUDA cores / 640 Tensor cores Memory 16 GB HBM2, 4096-bit bus, 900 GB/s bandwidth Performance Up to 14 TFLOPS (FP32), 112 TFLOPS (Tensor), 7 TFLOPS (FP64) ECC Supported Interface PCIe 3.0 x16 Power Connectors 2× 8-pin PCIe Max Power 250 W (PCIe version) Cooling PNY triple-fan active cooling OS Support Windows Server / 10 / 11, Linux (Ubuntu, CentOS, RHEL) Condition Used — nearly new Condition Used — nearly new condition. This graphics card is in excellent, nearly brand-new cosmetic condition. It is equipped with three fans, and has been fully tested — 100% functional.

• 100% passed power-on, nvidia-smi verification & load testing
• Excellent condition — clean, minimal to no usage marks
• Three fans intact and spinning smoothly, no bearing noise
• Original packaging included
• Gold pins intact, no bending

⚠ Important — Read Before Buying:
This is a high-power data-center GPU. It requires a PCIe 3.0 x16 slot, two 8-pin PCIe power connectors, and a sufficient PSU (750W+ recommended depending on your build). It uses a triple-fan active cooler — verify physical clearance in your chassis before ordering. This is a compute accelerator designed primarily for AI / deep-learning / scientific workloads; it is not intended as a consumer gaming or display graphics card. Verify your PSU wattage, chassis clearance and driver support before ordering. We are not responsible for incompatibility caused by insufficient power, space limits, or misunderstanding of its function. Package Includes

1 × NVIDIA Tesla Referral V100 16GB PNY Triple-Fan GPU

Warranty & Returns Warranty 30-day functional warranty from delivery date Returns 30-day returns accepted (buyer pays return shipping if not DOA) Testing Every unit tested before dispatch for stable performance Shipping

• Dispatched within 1–3 business days after payment
• Tracking number provided for every order
• Standard & Expedited options at checkout
• International shipping via Global Shipping Program

Why Buy From Us

• Real stock — what you see is what ships
• Tested inventory — no dead-on-arrival surprises
• Fast response — message us anytime before or after purchase
• Trusted seller — serving AI developers & labs worldwide

FAQ

Q: Is this GPU new or used?

A: It is in nearly new condition. It has been fully tested and is 100% functional.

Q: Can I use this for gaming?

A: This is an enterprise compute accelerator designed primarily for AI, deep learning and scientific computing. It is not intended as a consumer gaming or display graphics card.

Q: What power and space do I need?

A: It needs a PCIe 3.0 x16 slot, two 8-pin PCIe power connectors, and a sufficient PSU (750W+ recommended). Verify chassis clearance for the triple-fan cooler before ordering.

Q: What is the memory capacity and bandwidth?

A: 16GB of HBM2 memory on a 4096-bit bus, delivering 900 GB/s of memory bandwidth with ECC support.

Questions about compatibility or power? Message us before ordering — we reply fast.