How does NVIDIA A40 compare on raw specs to A100 and H100?
💡 Answer
NVIDIA A40 is a Ampere card with 48 GB of GDDR6 memory, 696 GB/s of bandwidth, and 150 TFLOPS of FP16 compute. It launched in 2020 and uses up to 300W of power.
On real workloads the memory subsystem is typically the most important factor: 696 GB/s of bandwidth is enough to sustain most transformer decoding tasks without starving the compute units, while 48 GB of VRAM supports a wide range of model sizes, especially once quantisation is applied.
Rent NVIDIA A40 today from Vast.ai, RunPod, or Massed Compute — compare live pricing and spin up in minutes.
More FAQs about NVIDIA A40
Vast.ai vs RunPod vs Massed Compute vs Cherry Servers vs Vultr - GPU Provider Comparison (April 2026)
Side-by-side comparison of Vast.ai vs RunPod vs Massed Compute vs Cherry Servers vs Vultr. Quickly compare GPU models, hourly pricing, spot instances, billing granularity, VRAM, regions, developer tools, Kubernetes support, and compliance to narrow down your cloud GPU provider shortlist. Data updated April 2026.
|
Vast.ai
Instant GPUs. Transparent Pricing.
|
RunPod
The cloud built for AI — deploy and scale GPU workloads from serverless inference to instant multi-node clusters on demand.
|
Massed Compute
GPU cloud with direct engineer support
|
Cherry Servers
Bare metal GPU servers with 24 years of hosting experience and full hardware-level control.
|
Vultr
High-performance cloud GPU across 32 global regions
|
|
|---|---|---|---|---|---|
| Overview | |||||
| Trustpilot Rating | 4.4 | 3.7 | 0 | 4.6 | 1.8 |
| Headquarters | United States | United States | United States | Lithuania | United States |
| Provider Type | GPU Marketplace | GPU-Focused | GPU-Focused | N/A | Multi-Cloud |
| Best For | AI training inference fine-tuning Stable Diffusion batch processing research LLM serving generative AI | AI training inference fine-tuning Stable Diffusion batch processing rendering research LLM serving generative AI | AI training inference VFX rendering generative AI fine-tuning HPC Stable Diffusion research | AI training inference fine-tuning rendering research HPC generative AI deep learning | AI training inference video rendering HPC Stable Diffusion game development generative AI fine-tuning research |
| GPU Hardware | |||||
| GPU Models | B200 H200 H100 SXM H100 NVL A100 SXM A100 PCIe RTX 5090 RTX 5080 RTX 5070 Ti RTX 6000 Pro RTX 6000 Ada RTX 4500 Ada RTX A6000 RTX A5000 RTX A4000 L40S L40 A40 A10 RTX 4090 RTX 4080 RTX 4070 Ti RTX 4070 RTX 4060 Ti RTX 4060 RTX 3090 Ti RTX 3090 RTX 3080 Ti RTX 3080 RTX 3070 Ti RTX 3070 Tesla V100 Tesla T4 A2 GTX 1080 | B300 B200 H200 H100 SXM H100 PCIe H100 NVL MI300X A100 SXM A100 PCIe RTX 5090 RTX PRO 6000 L40S L40 RTX 6000 Ada RTX 5000 Ada RTX A6000 RTX A5000 RTX 4090 RTX 4080 SUPER RTX 4080 RTX 4070 Ti RTX 3090 Ti RTX 3090 RTX 3080 Ti RTX 3080 RTX 3070 A40 A30 A2 L4 | A30 RTX A5000 RTX A6000 L40S A100 SXM H100 PCIe H100 SXM H100 NVL RTX PRO 6000 H200 NVL | A100 A40 A16 A10 A2 Tesla P4 | A16 A40 L40S A100 PCIe GH200 A100 SXM H100 SXM B200 B300 MI300X MI325X MI355X |
| Max VRAM (GB) | 192 | 288 | 141 | 80 | 288 |
| Max GPUs/Instance | 8 | 8 | 8 | 2 | 16 |
| Interconnect | NVLink, InfiniBand | NVLink | NVLink | PCIe | NVLink |
| Pricing | |||||
| Starting Price ($/hr) | $0.06/hr | $0.06/hr | $0.35/hr | $0.16/hr | $0.47/hr |
| Billing Granularity | Per-second | Per-second | Per-minute | Per-hour | Per-hour |
| Spot/Preemptible | Yes | Yes | No | No | Yes |
| Reserved Discounts | Up to 50% (1-6 month reserved) | 15-29% (1-month to 1-year plans) | N/A | N/A | N/A |
| Free Credits | Small test credit on signup | $5-$500 bonus after first $10 spend | None | None | Up to $300 free credit for 30 days |
| Egress Fees | Varies by host ($/TB) | None (Free) | None | N/A | Standard (varies by plan) |
| Storage | Varies by host ($/GB/hr, charged while instance exists) | Container/Volume ($0.10/GB/mo), Idle Volume ($0.20/GB/mo), Network Storage ($0.07/GB/mo 1TB) | Local NVMe included with instances | NVMe SSD, Elastic Block Storage ($0.071/GB/mo) | 350 GB - 61 TB NVMe (included), Block Storage at $0.10/GB/mo, S3-compatible Object Storage |
| Infrastructure | |||||
| Regions | 500+ locations, 40+ data centers | 31 global regions | United States (Tier III data centers) | Lithuania, Netherlands, Germany, Sweden, US, Singapore (6 locations) | 32 regions across 6 continents (Americas, Europe, Asia, Australia, Africa) |
| Uptime SLA | No formal SLA (host reliability scores visible) | 99.99% | Tier III (99.98% design) | 99.97% | 100% |
| Developer Experience | |||||
| Frameworks | PyTorch TensorFlow CUDA vLLM ComfyUI | PyTorch TensorFlow JAX ONNX CUDA | PyTorch TensorFlow CUDA cuDNN ComfyUI pre-configured ML templates | PyTorch TensorFlow CUDA (bare metal — full stack control) | PyTorch TensorFlow CUDA cuDNN ROCm Hugging Face NVIDIA NGC |
| Docker Support | Yes | Yes | Yes | Yes | Yes |
| SSH Access | Yes | Yes | Yes | Yes | Yes |
| Jupyter Notebooks | Yes | Yes | No | No | Yes |
| API / CLI | Yes | Yes | Yes | Yes | Yes |
| Setup Time | Seconds | Instant | Minutes | Minutes | Minutes |
| Kubernetes Support | No | No | No | Yes | Yes |
| Business Terms | |||||
| Min Commitment | None | None | None | None | None |
| Compliance | SOC 2 Type 2 HIPAA GDPR CCPA | SOC 2 Type II | SOC 2 Type II HIPAA | ISO 27001 ISO 20000-1 GDPR PCI DSS | SOC 2+ (HIPAA) PCI ISO 27001 ISO 27017 ISO 27018 ISO 20000-1 CSA STAR Level 1 |
Vast.ai
RunPod
Massed Compute
Cherry Servers
Vultr