GPU/GPU Stack/NVIDIA A40
Inference Optimized

NVIDIA A40

The versatile data center GPU for AI inference, visualization, and virtual workstations. 48GB GDDR6 memory with ray tracing acceleration for multi-purpose deployments.

Key Specifications

ArchitectureAmpere
VRAM48 GB GDDR6
Memory Bandwidth696 GB/s
CUDA Cores10,752
Tensor Cores336 (3rd Gen)
FP16 Performance150 TFLOPS

Technical Specifications

ArchitectureAmpere
VRAM48 GB GDDR6
Memory Bandwidth696 GB/s
CUDA Cores10,752
Tensor Cores336 (3rd Gen)
FP16 Performance150 TFLOPS
RT Cores84 (2nd Gen)
TDP300W
InterconnectNVLink Bridge (112.5 GB/s)
PCIeGen4 x16
Form FactorPCIe

Pricing Plans

Flexible pricing options to match your workload requirements.

On-Demand

Pay as you go with no commitment

120/hour
  • 1x NVIDIA A40 GPU
  • 16 vCPUs
  • 128 GB RAM
  • 500 GB NVMe SSD
  • No minimum commitment
  • Start/stop anytime
Most Popular

Reserved 1 Month

Save 15% with monthly commitment

51,000/month
  • 1x NVIDIA A40 GPU
  • 16 vCPUs
  • 128 GB RAM
  • 500 GB NVMe SSD
  • 15% discount
  • Priority support

Reserved 1 Year

Maximum savings with annual commitment

36,000/month
  • 1x NVIDIA A40 GPU
  • 16 vCPUs
  • 128 GB RAM
  • 500 GB NVMe SSD
  • 40% discount
  • Dedicated support

Why Choose NVIDIA A40

Versatile Workloads

Combines AI inference, professional graphics, and virtualization in one GPU.

Ray Tracing

2nd generation RT cores for real-time ray tracing and path tracing.

48GB VRAM

Large memory pool for complex models and high-resolution rendering.

vGPU Support

Certified for NVIDIA vGPU software for virtual desktop infrastructure.

Use Cases

AI Inference

Deploy trained models for production inference at optimal cost.

Virtual Workstations

Power remote professional graphics workstations with GPU acceleration.

Rendering & Visualization

Real-time ray tracing with 2nd generation RT cores.

Video Processing

Accelerate video encoding, decoding, and transcoding workflows.

Ready to Deploy NVIDIA A40?

Optimized for AI inference and visualization