Searching the live catalogue1,024offerings42providers81GPU models59regions
56 matching configurations
42 providers evaluated
Interpreted requirements
Accelerator: H100
Quantity: 8 GPUs
Price ceiling: $2.50 / GPU-hour
Start window: Within 7 days
Ranked using workload compatibility, placement, observed pricing and source evidence. Unknown requirements remain open for provider confirmation; results are not reservations.
Placement & cost estimate
8 of 8 GPUs in the planning allocation · $13.90/hour.
Recorded capacity covers 0 GPUs; 8 use configuration limits, not confirmed stock. Remaining shortfall: 0.
8 allocated GPUs rely on configuration evidence rather than confirmed aggregate live capacity.
Best match
GMI Cloud · H100 SXM
Global / region-aware
Available
$2.00 USD / GPU-hourCompute estimate $11,680.00 · Grade A
1 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling. Tracked availability indicates immediate access; provider reconfirmation is required.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$1.70 USD / GPU-hourCompute estimate $9,928.00 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. A single tracked configuration accommodates the requested 8 GPUs. Comparable price is within the $2.50 / GPU-hour ceiling.
Live aggregate available GPU count is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$2.44 USD / GPU-hourCompute estimate $14,242.30 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. A single tracked configuration accommodates the requested 8 GPUs. Comparable price is within the $2.50 / GPU-hour ceiling.
Live aggregate available GPU count is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked.
$2.45 USD / GPU-hourCompute estimate $14,308.00 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. A single tracked configuration accommodates the requested 8 GPUs. Comparable price is within the $2.50 / GPU-hour ceiling.
Live aggregate available GPU count is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked.
$2.46 USD / GPU-hourCompute estimate $14,388.30 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. A single tracked configuration accommodates the requested 8 GPUs. Comparable price is within the $2.50 / GPU-hour ceiling.
Live aggregate available GPU count is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked.
$2.49 USD / GPU-hourCompute estimate $14,541.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. A single tracked configuration accommodates the requested 8 GPUs. Comparable price is within the $2.50 / GPU-hour ceiling.
Live aggregate available GPU count is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked.
From $0.79 USD / GPU-hourCompute estimate $4,613.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Quantity compatibility cannot be confirmed from current capacity evidence. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
Latitude.sh · H100 80GB — provider does not specify PCIe/SXM
Unspecified
Starting-price plan; location availability varies
From $1.68 USD / GPU-hourCompute estimate $9,811.20 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Quantity compatibility cannot be confirmed from current capacity evidence. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked.
$1.74 USD / GPU-hourCompute estimate $10,161.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$1.74 USD / GPU-hourCompute estimate $10,161.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
Hyperstack · H100 80GB — provider does not specify PCIe/SXM
Unspecified
Availability unconfirmed
From $1.75 USD / GPU-hourCompute estimate $10,220.00 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Quantity compatibility cannot be confirmed from current capacity evidence. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
From $1.82 USD / GPU-hourCompute estimate $10,628.80 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Quantity compatibility cannot be confirmed from current capacity evidence. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$1.92 USD / GPU-hourCompute estimate $11,212.80 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$1.99 USD / GPU-hourCompute estimate $11,621.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$1.99 USD / GPU-hourCompute estimate $11,621.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$1.99 USD / GPU-hourCompute estimate $11,621.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$2.00 USD / GPU-hourCompute estimate $11,680.00 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Quantity compatibility cannot be confirmed from current capacity evidence. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$2.25 USD / GPU-hourCompute estimate $13,140.00 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
$2.49 USD / GPU-hourCompute estimate $14,541.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked.
$2.49 USD / GPU-hourCompute estimate $14,541.60 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Request exceeds the per-configuration maximum of 1 GPUs; multiple configurations would be required and aggregate capacity is not confirmed. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
Hyperstack · H100 80GB — provider does not specify PCIe/SXM
Unspecified
Unconfirmed
$2.50 USD / GPU-hourCompute estimate $14,600.00 · Grade A
2 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. Comparable price is within the $2.50 / GPU-hour ceiling.
Quantity compatibility cannot be confirmed from current capacity evidence. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
Quote requiredCompute estimate Not priced · Grade A
3 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. A single tracked configuration accommodates the requested 8 GPUs.
Live aggregate available GPU count is not confirmed. Comparable USD / GPU-hour price is not available. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Comparable USD / GPU-hour price is not tracked. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked.
Quote requiredCompute estimate Not priced · Grade A
3 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. A single tracked configuration accommodates the requested 8 GPUs.
Live aggregate available GPU count is not confirmed. Comparable USD / GPU-hour price is not available. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Comparable USD / GPU-hour price is not tracked. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
Quote requiredCompute estimate Not priced · Grade A
3 requirements need confirmation. Compute-only estimate; additional charges and final terms may apply.
Why this match · economics & evidence
GPU requirement matches the tracked accelerator. A single tracked configuration accommodates the requested 8 GPUs.
Live aggregate available GPU count is not confirmed. Comparable USD / GPU-hour price is not available. Start within 7 days is not confirmed by tracked availability evidence.
Billed planning month is 730 hours unless the workload uses an explicit run duration. Utilization assumption is 100.0%. No egress volume was requested, so egress cost is excluded from the scenario. This scenario models compute, requested egress, setup and entered overhead only; taxes, storage and other unmodeled charges are not a full provider invoice. Comparable USD / GPU-hour price is not tracked. Setup fee is undisclosed or invalid; it is not assumed to be zero. No applicable public comparable performance benchmark or explicit throughput assumption is available. Network bandwidth is not included in the cost model because a comparable cost/performance basis is not tracked. Storage pricing is not tracked in this economics model.
