Rent V100 32GB
Compare live on-demand and spot rental prices across 97+ cloud providers. Previous-gen workhorse. 32GB HBM2 at low cost. Still widely available. No FP8 or BF16 hardware support.
Choosing the right billing model for V100 32GB
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
- Legacy ML training
- HPC workloads
- Budget multi-GPU setups
V100 32GB — Specs & Benchmarks
Performance bars, compute tiers (FP32/FP16/BF16/FP8/INT8), memory specs, LLM model size guidance, and related GPU comparisons.
V100 32GB Rental Guide
The V100 32GB is the most affordable option for workloads requiring 32GB of HBM memory. At $0.50–$1.50/hr on-demand — significantly cheaper than A100 — it is a viable choice for legacy ML training, HPC workloads, and budget-constrained teams. It is widely available on AWS p3.2xlarge and p3.8xlarge instances, making it accessible without specialized provider relationships.
Spot instances for V100 are available at 40–70% below on-demand rates on AWS and other providers. For batch training jobs with checkpointing, V100 spot instances can deliver competitive cost-per-FLOP for workloads that do not require BF16 or FP8. On-demand is appropriate for interactive workloads where uptime matters.
The main limitations to plan around are the lack of BF16 and FP8 hardware support, and the older PCIe 3.0 host interface. Modern transformer training frameworks are optimized for BF16 — running in FP16 on V100 requires careful loss scaling to avoid numerical instability. For new workloads, the A100 40GB is often available at comparable cost with significantly better performance and BF16 support. Use the GPU cost calculator to compare V100 vs A100 total cost for your specific training duration.
Frequently Asked Questions
How much does it cost to rent a V100 32GB?
V100 32GB 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 V100 32GB?
The cheapest V100 32GB 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 V100 32GB?
With 32GB of HBM2, the V100 32GB can run LLM models up to approximately 16B parameters at FP16, 32B at INT8, or 64B at INT4/GGUF quantization. Common workloads include: Legacy ML training, HPC workloads, Budget multi-GPU setups. Previous-gen workhorse. 32GB HBM2 at low cost. Still widely available. No FP8 or BF16 hardware support.
Should I use on-demand or spot pricing for V100 32GB?
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 V100 32GB 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 V100 32GB's 112 TFLOPS FP16 and 900 GB/s bandwidth. The H100 is significantly more expensive — typically $2.50–$5.00/hr vs lower rates for the V100 32GB. For workloads that fit within 32GB and don't require FP8 precision, the V100 32GB often delivers better cost-per-token than the H100.
What is the memory bandwidth of the V100 32GB and why does it matter?
The V100 32GB has 900 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 V100 32GB for Stable Diffusion or image generation?
Yes — the V100 32GB is capable for Stable Diffusion and image generation workloads. Image generation is primarily FP32 and FP16 compute-bound, and the V100 32GB's 14 TFLOPS FP32 throughput determines images-per-second. The 32GB 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.