Specifications Compared
| Spec | H100 NVL | RTX 4090 |
|---|---|---|
| TDP | 700W | 450W |
| VRAM | 94 GB | 24 GB |
| CUDA Cores | 16,896 | 16,384 |
| FP8 (dense) | 1,979 TFLOPS | 330.3 TFLOPS |
| Memory Type | HBM3 | GDDR6X |
| Architecture | Hopper | Ada Lovelace |
| FP16 (dense) | 989 TFLOPS | 165.2 TFLOPS |
| Form Factors | NVL (PCIe card, NVLink bridged) | PCIe |
| INT8 (dense) | 1,979 TOPS | 660.6 TOPS |
| Interconnect | NVLink bridge, PCIe 5.0 | PCIe 4.0 |
| Tensor Cores | 528 | 512 |
| FP32 Performance | 67 TFLOPS | 82.6 TFLOPS |
| FP64 Performance | 34 TFLOPS | 1.3 TFLOPS |
| Memory Bandwidth | 3,350 GB/s | 1,008 GB/s |
| FP8 (with sparsity) | 3,958 TFLOPS | 660.6 TFLOPS |
| FP16 (with sparsity) | 1,979 TFLOPS | 330.4 TFLOPS |
| INT8 (with sparsity) | 3,958 TOPS | 1,321.2 TOPS |
Performance Analysis
The H100 NVL family delivers 989 TFLOPS FP16 dense throughput and 67 TFLOPS FP32 throughput while the RTX 4090 family delivers 165.2 TFLOPS FP16 dense throughput and 82.6 TFLOPS FP32 throughput. This FP16 dense gap means the H100 NVL processes larger matrix operations in training and inference at higher throughput rates than the RTX 4090. Memory bandwidth shared across each family stands at 3350 GB/s for the H100 NVL versus 1008 GB/s for the RTX 4090 so the H100 NVL sustains larger batch sizes during memory intensive operations.
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.
H100 NVL
| Provider | Region | GPUs | Per GPU / hr | Instance / hr | Deploy |
|---|---|---|---|---|---|
| Vast.ai | Czechia, CZ | 1 | $2.59 | — | Deploy |
| QuantaCloud | us-midwest-2 | 4 | $2.59 | $10.36 | Deploy |
| Massed Compute | us-central-3 | 1 | $2.73 | — | Deploy |
| Ori | dallas-2 | 2 | $2.90 | $5.80 | Deploy |
| RunPod | global | 1 | $3.99 | — | Deploy |
9 providers in stock, 24 offers (cheapest per provider shown). All H100 NVL offers, price history and alerts
Notify me when H100 NVL drops below a price
One email when the cheapest in-stock on-demand price per GPU-hour falls below your threshold. Cheapest right now: $2.59/GPU-hr.
QuantaCloud
Comparing H-series providers? We broker across all of them.
Hopper stock changes by the hour and prices differ by provider. If you need 16+ GPUs reserved or a cluster in the next 90 days, we quote H-series or B300 inventory at partner rates: one quote, 24h turnaround.
When to Choose the H100 NVL
The H100 NVL variant suits scenarios that require 94 GB HBM3 capacity and NVLink bridge connectivity for distributed multi GPU setups. Its 700 W TDP and 3350 GB/s bandwidth support sustained high throughput on FP16 dense workloads that exceed the limits of 24 GB configurations.
When to Choose the RTX 4090
The RTX 4090 variant fits single node tasks where 24 GB GDDR6X and 450 W TDP provide adequate capacity at lower power draw. Its 82.6 TFLOPS FP32 throughput exceeds the H100 NVL family figure of 67 TFLOPS FP32 for certain compute bound single precision applications.
Use Cases
The H100 NVL variant supplies 94 GB HBM3 and 989 TFLOPS FP16 dense throughput that accommodate large model training beyond 24 GB limits.
The H100 NVL variant provides 3350 GB/s bandwidth and 1979 TFLOPS FP8 dense throughput that enable higher concurrency than the RTX 4090 family figures.
The H100 NVL variant offers 94 GB capacity and NVLink bridge support that handle parameter updates on models exceeding 24 GB VRAM.
The RTX 4090 variant delivers 82.6 TFLOPS FP32 throughput and 450 W TDP that suffice for image generation workloads within 24 GB.
The H100 NVL variant supplies higher bandwidth while the RTX 4090 variant supplies higher FP32 throughput so selection depends on exact precision needs.
Frequently Asked Questions
What memory capacities distinguish the H100 NVL variant from the RTX 4090 variant?▾
The H100 NVL variant contains 94 GB HBM3 memory. The RTX 4090 variant contains 24 GB GDDR6X memory.
How do the FP16 dense figures compare between these two GPU families?▾
The H100 NVL family reports 989 TFLOPS FP16 dense. The RTX 4090 family reports 165.2 TFLOPS FP16 dense.
Which interconnect options appear on the H100 NVL variant versus the RTX 4090 variant?▾
The H100 NVL variant uses NVLink bridge together with PCIe 5.0. The RTX 4090 variant uses PCIe 4.0 only.
What TDP values apply to each GPU family?▾
The H100 NVL family operates at 700 W TDP. The RTX 4090 family operates at 450 W TDP.
Do the two variants share the same memory bandwidth?▾
No. The H100 NVL family bandwidth measures 3350 GB/s. The RTX 4090 family bandwidth measures 1008 GB/s.
Which architecture powers each specific variant?▾
The H100 NVL variant uses Hopper architecture. The RTX 4090 variant uses Ada Lovelace architecture.
Which is cheaper to rent, the H100 NVL or the RTX 4090?▾
Cloud rental prices for both the H100 NVL and RTX 4090 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 H100 NVL have compared to the RTX 4090?▾
The H100 NVL has 94 GB of HBM3 memory. The RTX 4090 has 24 GB of GDDR6X memory.
Can I find H100 NVL and RTX 4090 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 H100 NVL and the RTX 4090?▾
The H100 NVL uses the Hopper architecture (2022) while the RTX 4090 uses Ada Lovelace (2022). The H100 NVL delivers 6.0x the dense FP16 throughput (989 vs 165.2 TFLOPS, both without sparsity) and 3.3x the memory bandwidth of the RTX 4090.
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 H100 NVL and the RTX 4090. 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 H100Every current offer by provider, daily price history and a price alert.
- Rent RTX 4090Every 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.