H100 vs H20
The H100 offers significantly higher compute and 900 GB/s NVLink, while the H20 has 96GB memory at lower compute. Choose H100 for performance, H20 for export-compliant memory-bound inference.
Overview
Nvidia H100
|
Nvidia H20
|
|
|---|---|---|
| Summary | The standard data center GPU for large-scale AI training and inference. | Export-compliant Hopper inference GPU with lower compute than H100. |
| Architecture | Hopper | Hopper |
| Memory | 80GB HBM3 | 96GB HBM3 |
| Memory Bandwidth | 3,350GB/s | 4,000 GB/s |
| Performance | 1,979 TFLOPS (FP16) | 148 TFLOPS (FP16) |
| Release Date | Q3 2022 | Early 2024 |
| Performance Segment | High Performance | High Performance |
| Datasheet | Link | -- |
Price comparison
H100 Pricing
We're tracking 45 cloud providers for the H100, and the pricing spread is significant. While the average sits at $3.16/hr, the lowest price is currently $0.36/hr per GPU (spot instance).
How we sort results
Default ranking
Our algorithm weighs five factors to find the most relevant matches for you:
- Price: We blend total hourly price with price-per-GPU to balance affordability and value.
- Specs: We favor offers with higher CPU, RAM, and GPU Memory.
- Billing: We favor on-demand billing for simplicity and flexibility over spot instances, reservations, and custom quotes.
- Location: We blend both datacenter proximity and provider HQ location. Datacenter location matters for latency, while HQ location matters for compliance and support.
- Provider diversity: Each time a provider appears in the list, their subsequent offerings are ranked a little lower, so one provider's offerings don't crowd out the top positions.
Sorting and filtering
Click "View all offerings" to access the full list with sortable columns and filters for billing type, GPU count, vCPUs, and RAM.
Transparency and funding
Ads and sponsors: Any paid placements are fixed at the top and clearly labeled as sponsored content.
Affiliates: Any affiliate links will be indicated to you as well. We may earn a commission if you click them, but this never influences the ranking order.
| Provider | GPUs | Billing | Total/h | |
|---|---|---|---|---|
Thunder Compute
Our sponsor
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $1.38 | View |
Runpod
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $1.99 | View |
Theta EdgeCloud
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $2.29 | View |
Novita
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $2.59 | View |
Lambda Labs
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $3.29 | View |
DigitalOcean
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $3.39 | View |
Together
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $3.36 | View |
Paperspace
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $5.99 | View |
CoreWeave
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $6.16 | View |
Azure
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $6.98 | View |
AWS
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $6.88 | View |
Vultr
|
8x | On-Demand Pay-as-you-go pricing. No term commitments. | $23.92 | View |
Runpod
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $2.59 | View |
Hyperstack
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $1.90 | View |
UpCloud
|
1x | On-Demand Pay-as-you-go pricing. No term commitments. | $2.05 | View |
| No offerings matching your filters. | ||||
H20 Pricing
The H20 is not currently offered by any cloud providers we track.
Heads up: We do our best to keep these specs & prices accurate. However, cloud costs may fluctuate based on region, usage, and other factors not listed here. These are estimates based on common setups and are for informational purposes only. Always verify current rates & exact specs with the provider before provisioning.
Alternative GPUs
Here's how these GPUs compare against alternatives:
Our data for Nvidia H100 was last updated on March 25, 2026, and for Nvidia H20 on March 25, 2026.