Specifications Compared
| Spec | H200 | L40 |
|---|---|---|
| TDP | 700W | 300W |
| VRAM | 141 GB | 48 GB |
| CUDA Cores | 16,896 | 18,176 |
| FP8 (dense) | 1,979 TFLOPS | 362 TFLOPS |
| Memory Type | HBM3e | GDDR6 |
| Architecture | Hopper | Ada Lovelace |
| FP16 (dense) | 989 TFLOPS | 181.05 TFLOPS |
| Form Factors | SXM, NVL | PCIe |
| INT8 (dense) | 1,979 TOPS | 362 TOPS |
| Interconnect | NVLink, PCIe 5.0, InfiniBand | PCIe 4.0 |
| Tensor Cores | 528 | 568 |
| FP32 Performance | 67 TFLOPS | 90.5 TFLOPS |
| FP64 Performance | 34 TFLOPS | Not published |
| Memory Bandwidth | 4,800 GB/s | 864 GB/s |
| FP8 (with sparsity) | 3,958 TFLOPS | 724 TFLOPS |
| FP16 (with sparsity) | 1,979 TFLOPS | 362.1 TFLOPS |
| INT8 (with sparsity) | 3,958 TOPS | 724 TOPS |
Performance Analysis
The H200 records 989 TFLOPS FP16 dense performance against the L40 recording 181.05 TFLOPS FP16 dense performance. This difference produces more than five times the throughput for dense FP16 operations on the H200. The same pattern holds for sparse figures where the H200 reaches 1979 TFLOPS FP16 with sparsity against 362.1 TFLOPS FP16 with sparsity on the L40. Memory bandwidth of 4800 GB/s on the H200 versus 864 GB/s on the L40 permits substantially larger batch sizes during training and inference passes. The L40 instead supplies 90.5 TFLOPS FP32 performance compared with 67 TFLOPS FP32 on the H200 which favors precision sensitive non tensor tasks.
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.
H200
| Provider | Region | GPUs | Per GPU / hr | Instance / hr | Deploy |
|---|---|---|---|---|---|
| RunPod | global | 1 | $5.29 | — | Deploy |
| Lyceum | Europe | 8 | $6.19 | $49.52 | Deploy |
| LeaderGPU | The Netherlands | 1 | $6.22 | — | Deploy |
3 providers in stock, 3 offers. All H200 offers, price history and alerts
L40
| Provider | Region | GPUs | Per GPU / hr | Instance / hr | Deploy |
|---|---|---|---|---|---|
| ThunderCompute | USA | 1 | $0.79 | — | Deploy |
| Massed Compute | us-central-3 | 1 | $0.86 | — | Deploy |
| QuantaCloud | us-midwest-2 | 4 | $0.94 | $3.75 | Deploy |
| Hyperstack | CANADA-1 | 1 | $1.00 | — | Deploy |
4 providers in stock, 12 offers (cheapest per provider shown). All L40 offers, price history and alerts
Notify me when H200 drops below a price
One email when the cheapest in-stock on-demand price per GPU-hour falls below your threshold. Cheapest right now: $5.29/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 H200
The H200 suits deployments that require 141 GB VRAM capacity together with 4800 GB/s bandwidth for large scale tensor workloads. Its 989 TFLOPS FP16 dense rating and 1979 TFLOPS FP8 dense rating accelerate both training and inference phases beyond the capabilities of the L40. Teams select the H200 when model sizes exceed the 48 GB limit of the L40.
When to Choose the L40
The L40 fits environments constrained by 300W TDP limits where 90.5 TFLOPS FP32 performance provides an advantage over the 67 TFLOPS FP32 of the H200. Its PCIe form factor and 362 TFLOPS FP8 dense rating remain adequate for lighter inference or graphics adjacent tasks that do not demand the 141 GB capacity of the H200.
Use Cases
The H200 supplies 989 TFLOPS FP16 dense performance and 141 GB VRAM which accommodate larger models than the 48 GB limit of the L40.
The H200 achieves 1979 TFLOPS FP8 dense performance which supports higher throughput than the 362 TFLOPS FP8 dense performance of the L40.
The H200 provides 4800 GB/s bandwidth and 141 GB memory which enable larger batch sizes than the 864 GB/s and 48 GB of the L40.
The L40 delivers 90.5 TFLOPS FP32 performance which exceeds the 67 TFLOPS FP32 of the H200 for graphics oriented precision workloads.
The L40 operates at 300W TDP and supplies 90.5 TFLOPS FP32 performance which suits power limited FP32 calculations better than the 700W TDP of the H200.
Frequently Asked Questions
How much more memory does the H200 have than the L40?▾
The H200 contains 141 GB HBM3e memory while the L40 contains 48 GB GDDR6 memory. This difference equals nearly three times the capacity on the H200.
What is the FP16 dense performance gap between the two GPUs?▾
The H200 reaches 989 TFLOPS FP16 dense while the L40 reaches 181.05 TFLOPS FP16 dense. The H200 therefore supplies more than five times the dense FP16 throughput.
Which GPU offers higher FP32 performance?▾
The L40 provides 90.5 TFLOPS FP32 performance compared with 67 TFLOPS FP32 on the H200. The L40 holds the advantage in this metric.
How does memory bandwidth compare between the H200 and L40?▾
The H200 features 4800 GB/s bandwidth while the L40 features 864 GB/s bandwidth. The H200 bandwidth exceeds the L40 bandwidth by a factor of more than five.
What TDP values distinguish the H200 from the L40?▾
The H200 lists a 700W TDP and the L40 lists a 300W TDP. The L40 consumes less than half the power of the H200.
Which interconnect options appear on each card?▾
The H200 supports NVLink together with PCIe 5.0 and InfiniBand while the L40 supports only PCIe 4.0.
Which is cheaper to rent, the H200 or the L40?▾
Cloud rental prices for both the H200 and L40 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 H200 have compared to the L40?▾
The H200 has 141 GB of HBM3e memory. The L40 has 48 GB of GDDR6 memory.
Can I find H200 and L40 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 H200 and the L40?▾
The H200 uses the Hopper architecture (2024) while the L40 uses Ada Lovelace (2023). The H200 delivers 5.5x the dense FP16 throughput (989 vs 181.05 TFLOPS, both without sparsity) and 5.6x the memory bandwidth of the L40.
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 H200 and the L40. 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 H200Every current offer by provider, daily price history and a price alert.
- Rent L40Every 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.