← All rental offers

Your rental shortlist

Compare the whole instance

Host resources, billing rules and workload costs for up to three offers. Missing facts stay unknown.

Resources and prices are for one whole instance. Host RAM is separate from GPU VRAM. GB and GiB retain the source's units.
What is includedTensorDock ↗Browse TensorDock ↗H100 SXM 80 GBH100 SXM5 marketplace starting rateRemove TensorDock H100 SXM5 marketplace starting rate
GPU capacityInstance GPU count unknown · 80 GB per GPU
Compute priceWhole-instance price unknownUSD 2.25 / hr per GPU · Starting price
Rental / commitmentHourly · On-demandNo monthly commitmentPay-as-you-go marketplace rate with no commitment.
CountryUnspecified / multiple countries
Provider zoneUnknown
Source location / coverageUnknown
Cloud tierUnknown
Host CPUUnknown
Host RAMUnknown
Included local storageUnknown
Boot storageUnknown
Billing minimum / roundingUnknown
Additional persistent storageUnknown
Outbound transferUnknown
Setup feeUnknown
Public IP / other required extrasUnknown
TaxesUnknown; excluded from estimate
Stock / regionStock not verified · Region unknown
EvidenceCompute pricing ↗Price checked Oct 24, 2026, 09:02 EDT · StaleDetails checked Unknown · Recheck required

Plan your workload

Estimate the known costs

One instance, one continuous allocation. Enter the time it remains billable, including idle time. Additional persistent storage is separate from included local disks. Storage retention can outlast compute; monthly storage uses full billing months, without guessed prorating.