Rent A40
Compare live on-demand and spot rental prices across 97+ cloud providers. 48GB GDDR6 at a lower price than A100. Good for workloads needing large VRAM without HBM cost.
Choosing the right billing model for A40
Provision and terminate at any time. Ideal for development, short experiments, and workloads with unpredictable duration.
Instances can be reclaimed when demand spikes. Best for fault-tolerant batch jobs, training with checkpointing, and preprocessing.
Lock in a rate for 1–3 months. Right for sustained production inference or long training runs where cost predictability matters.
Best use cases
- Visualization + compute
- Mid-size inference
- Virtual workstations
A40 — Specs & Benchmarks
Performance bars, compute tiers (FP32/FP16/BF16/FP8/INT8), memory specs, LLM model size guidance, and related GPU comparisons.
A40 Rental Guide
Use an A40 when a single 48GB data-center GPU is needed for visualization, virtual workstation workloads, or model serving without paying for A100 HBM capacity. Its cost case is strongest when the application benefits from both professional graphics support and ample VRAM.
Spot capacity works well for render farms, batch image generation, and offline inference that can be retried. On-demand is a better choice for interactive VDI or customer-facing services, where GPU replacement and availability need to be predictable.
Compare the A40 with L40S for newer Ada efficiency and stronger compute, and with A100 40GB/80GB for HBM bandwidth and multi-GPU topology. The A40’s lack of NVLink, FP8, and HBM means it should not be selected solely on the basis of its 48GB headline capacity.
Frequently Asked Questions
How much does it cost to rent a A40?
A40 on-demand rental prices vary by provider and region. On-demand rates typically range based on availability and provider margins — use the comparison table above to see current live rates across all providers. Spot instances are generally 40–70% cheaper than on-demand but can be interrupted. Monthly cost estimates (hourly rate × 730 hours) are shown in the table for sustained workloads.
Which cloud provider has the cheapest A40?
The cheapest A40 provider changes as providers update their pricing. The comparison table above shows live rates sorted by price, so the cheapest option is always at the top. Factors beyond headline price include region (latency to your users), availability (high/medium/low), and billing granularity (per-second vs per-hour minimums).
What can I run on a A40?
With 48GB of GDDR6, the A40 can run LLM models up to approximately 24B parameters at FP16, 48B at INT8, or 96B at INT4/GGUF quantization. Common workloads include: Visualization + compute, Mid-size inference, Virtual workstations. 48GB GDDR6 at a lower price than A100. Good for workloads needing large VRAM without HBM cost.
Should I use on-demand or spot pricing for A40?
Spot instances save 40–70% vs on-demand but can be interrupted when the provider needs capacity back. Use spot for: batch inference jobs, training runs with checkpointing, preprocessing pipelines, and any workload that can tolerate interruption and restart. Use on-demand for: production inference serving, interactive workloads, and jobs that cannot be interrupted. Most providers bill per second, so short on-demand jobs are not penalized by hourly minimums.
How does the A40 compare to the H100 for cloud rental?
The H100 80GB delivers 1,979 TFLOPS FP16 with 3,350 GB/s HBM3 bandwidth, compared to the A40's 74.8 TFLOPS FP16 and 696 GB/s bandwidth. The H100 is significantly more expensive — typically $2.50–$5.00/hr vs lower rates for the A40. For workloads that fit within 48GB and don't require FP8 precision, the A40 often delivers better cost-per-token than the H100.
What is the memory bandwidth of the A40 and why does it matter?
The A40 has 696 GB/s of memory bandwidth. For LLM inference, memory bandwidth is often more important than raw TFLOPS — each autoregressive token generation reads the full model weight matrix from VRAM, so bandwidth directly determines tokens-per-second throughput. Higher bandwidth means faster inference for the same model at the same batch size. For batch inference (processing many requests simultaneously), compute throughput becomes more important.
Can I use the A40 for Stable Diffusion or image generation?
Yes — the A40 is capable for Stable Diffusion and image generation workloads. Image generation is primarily FP32 and FP16 compute-bound, and the A40's 37.4 TFLOPS FP32 throughput determines images-per-second. The 48GB VRAM fits SDXL (requires ~6GB) and most ControlNet pipelines. For high-throughput image generation at scale, compare cost-per-image across providers using the GPU cost calculator.