Specifications Compared
| Spec | A40 | RTX 5090 |
|---|---|---|
| TDP | 300W | 575W |
| VRAM | 48 GB | 32 GB |
| CUDA Cores | 10,752 | 21,760 |
| FP4 (dense) | Not published | 1,676 TFLOPS |
| FP8 (dense) | Not published | 419 TFLOPS |
| Memory Type | GDDR6 | GDDR7 |
| Architecture | Ampere | Blackwell |
| FP16 (dense) | 149.7 TFLOPS | 209.5 TFLOPS |
| Form Factors | PCIe | PCIe |
| INT8 (dense) | 299.3 TOPS | 838 TOPS |
| Interconnect | NVLink, PCIe 4.0 | PCIe 5.0 |
| Tensor Cores | 336 | 680 |
| FP32 Performance | 37.4 TFLOPS | 104.8 TFLOPS |
| FP64 Performance | Not published | 1.6 TFLOPS |
| Memory Bandwidth | 696 GB/s | 1,792 GB/s |
| FP4 (with sparsity) | Not published | 3,352 TFLOPS |
| FP8 (with sparsity) | Not published | 838 TFLOPS |
| FP16 (with sparsity) | 299.4 TFLOPS | 419 TFLOPS |
| INT8 (with sparsity) | 598.6 TOPS | 1,676 TOPS |
Performance Analysis
The RTX 5090 delivers higher dense FP16 performance at 209.5 TFLOPS compared to the A40 dense FP16 performance at 149.7 TFLOPS. This ratio favors the RTX 5090 by a factor of 1.4 times for dense FP16 operations in both training and inference. The A40 sparse FP16 performance reaches 299.4 TFLOPS while the RTX 5090 sparse FP16 performance reaches 419 TFLOPS. Higher memory bandwidth on the RTX 5090 at 1792 GB/s versus 696 GB/s on the A40 supports larger batch sizes during model execution. The FP32 delta shows the RTX 5090 at 104.8 TFLOPS against the A40 at 37.4 TFLOPS which affects precision sensitive scientific workloads.
Current On-Demand Offers
Cheapest secure on-demand offer per provider that is reported in stock right now, per GPU per hour; multi-GPU instances show the whole-instance price alongside. Deploy opens the DeployGPU console for providers on the platform, otherwise the provider's own site. Stock is rechecked every minute.
A40
| Provider | Region | GPUs | Per GPU / hr | Instance / hr | Deploy |
|---|---|---|---|---|---|
| LeaderGPU | The Netherlands | 8 | $0.52 | $4.13 | Deploy |
| RunPod | global | 1 | $0.59 | — | Deploy |
2 providers in stock, 2 offers. All A40 offers, price history and alerts
RTX 5090
| Provider | Region | GPUs | Per GPU / hr | Instance / hr | Deploy |
|---|---|---|---|---|---|
| Vast.ai | Alberta, CA | 8 | $0.87 | $6.93 | Deploy |
| LeaderGPU | The Netherlands | 1 | $2.29 | — | Deploy |
2 providers in stock, 19 offers (cheapest per provider shown). All RTX 5090 offers, price history and alerts
Notify me when A40 drops below a price
One email when the cheapest in-stock on-demand price per GPU-hour falls below your threshold. Cheapest right now: $0.52/GPU-hr.
QuantaCloud
Comparing providers? We broker across all of them.
Stop tab-switching between pricing pages. Tell us what you need, 16+ GPUs reserved or cluster capacity, and we return one quote at partner rates within 24 hours.
When to Choose the A40
The A40 suits scenarios that require 48 GB of memory capacity for large models that exceed 32 GB limits. Its 300W TDP enables denser deployments in power constrained environments. NVLink interconnect supports multi GPU scaling where PCIe 5.0 alone proves insufficient for data exchange rates.
When to Choose the RTX 5090
Its 104.8 TFLOPS FP32 performance accelerates compute heavy tasks compared to 37.4 TFLOPS on the A40. Newer architecture supports FP8 dense at 419 TFLOPS where the A40 provides no published figure.
Use Cases
The RTX 5090 provides 209.5 TFLOPS dense FP16 compared to 149.7 TFLOPS on the A40 which accelerates training iterations.
Higher dense FP16 at 209.5 TFLOPS and bandwidth at 1792 GB/s on the RTX 5090 support larger batch sizes than the A40.
The A40 48 GB memory capacity accommodates larger fine tuning datasets that exceed the 32 GB limit on the RTX 5090.
The RTX 5090 delivers 419 TFLOPS sparse FP16 versus 299.4 TFLOPS on the A40 for faster image generation passes.
FP32 performance at 104.8 TFLOPS on the RTX 5090 surpasses 37.4 TFLOPS on the A40 for precision calculations.
Frequently Asked Questions
How does FP16 performance compare between A40 and RTX 5090?▾
The A40 provides 149.7 TFLOPS dense FP16 and 299.4 TFLOPS with sparsity. The RTX 5090 provides 209.5 TFLOPS dense FP16 and 419 TFLOPS with sparsity.
Which GPU has more memory for large models?▾
The A40 contains 48 GB GDDR6 while the RTX 5090 contains 32 GB GDDR7. This difference affects maximum model size without offloading.
What are the power requirements for each GPU?▾
The A40 TDP is 300W and the RTX 5090 TDP is 575W. Power delivery infrastructure must match these ratings for stable operation.
Does the A40 support NVLink for multi GPU setups?▾
The A40 includes NVLink interconnect alongside PCIe 4.0. The RTX 5090 relies on PCIe 5.0 without NVLink support.
How does memory bandwidth differ in practice?▾
The RTX 5090 reaches 1792 GB/s while the A40 reaches 696 GB/s. This gap influences data movement speed during inference and training.
Which is cheaper to rent, the A40 or the RTX 5090?▾
Cloud rental prices for both the A40 and RTX 5090 vary by provider, configuration, and availability. This page shows live pricing from 25+ providers updated every 60 seconds. Scroll to the Live Cloud Pricing section to compare current rates.
How much VRAM does the A40 have compared to the RTX 5090?▾
The A40 has 48 GB of GDDR6 memory. The RTX 5090 has 32 GB of GDDR7 memory.
Can I find A40 and RTX 5090 GPUs available to rent right now?▾
Yes. This page shows real-time availability across 25+ cloud GPU providers. The Live Cloud Pricing section displays only in-stock offers with current pricing.
What is the main difference between the A40 and the RTX 5090?▾
The A40 uses the Ampere architecture (2020) while the RTX 5090 uses Blackwell (2025). The RTX 5090 delivers 1.4x the dense FP16 throughput (209.5 vs 149.7 TFLOPS, both without sparsity) and 2.6x the memory bandwidth of the A40.
Rent these GPUs
Each GPU page lists every current on-demand offer by provider, per GPU-hour, updated every minute.
Related comparisons
How this page is made
- Specifications come from the NVIDIA, AMD and Intel datasheets for the A40 and the RTX 5090. Dense and sparse throughput are listed separately.
- Prices and availability are live: every offer shown is an in-stock, on-demand listing from a provider we track, rechecked every minute.
- The written comparison (overview, performance notes, when to choose each card, verdict, use cases and FAQ) was drafted with an AI model from the specification table above and passed an automated check that rejects any figure not in that table. It was last generated on . It contains no prices; those are always read live.
- Read how we collect the data, or report an error on this page.
Next steps
- Rent A40Every current offer by provider, daily price history and a price alert.
- Rent RTX 5090Every current offer by provider, daily price history and a price alert.
- GPU price indexHow on-demand prices for the major GPUs have moved, updated daily.
- All GPUsEvery GPU we track, with current lows and provider counts.