Cirrascale Cloud Services
US cloud provider of multi-accelerator AI training and inference infrastructure, offering NVIDIA, AMD Instinct, Qualcomm and Tenstorrent systems plus a managed inference platform. Sells hourly and in 3-, 6- and 12-month blocks.
Visit Cirrascale Cloud Services Compare with others Request a quote
Facts
Source per value- Type What kind of provider it is: hyperscaler, managed HPC service, GPU cloud, sovereign or EU-hosted cloud, publicly funded supercomputer, or simulation-software cloud.
- gpu-cloud 1
- Fabric Primary node-to-node network the provider states for its HPC or GPU clusters. RDMA means the provider names RDMA cluster networking without naming the transport.
- Not published
- Interconnect detail Speeds and instance families as the provider states them.
- Up to 3,200 Gb per server (fabric type not named) 5
- Schedulers Job schedulers and orchestrators the provider offers or documents.
- Not published
- Slurm Offers or documents Slurm (managed, or as a supported deployment).
- Not published
- Kubernetes Offers managed Kubernetes for its clusters.
- Not published
- CPUs CPU models the provider names for its HPC or GPU hosts.
- Not published
- GPUs / accelerators Accelerators the provider lists, as listed on its own pages.
- GB300, B300, B200, RTX PRO 6000, MI355X 5
- H100 Lists NVIDIA H100.
- No 1
- Blackwell Lists at least one NVIDIA Blackwell data-centre GPU (B200, B300, GB200, GB300).
- Yes 1
- AMD Instinct Lists at least one AMD Instinct GPU.
- Yes 1
- Bare metal Sells non-virtualised bare-metal nodes.
- Not published
- H100 $/GPU-hr $/GPU-hr Published per-GPU-hour price of an H100. Basis (on-demand, capacity block, derived from an 8-GPU price) is in the next column.
- Not published
- H100 price basis What the H100 price is: purchase mode, instance size and any derivation.
- Not published
- Price checked Date the price was read from the provider page.
- 2026-11-09 1
- Pricing model How the provider charges, as it states it.
- hourly + term blocks 1
- Regions Locations the provider names on its own pages.
- US (US-West, US-East, US-South per company about page) 3
- EU-hosted Names data-centre capacity in an EU member state.
- Not published
- Sovereignty claim The provider's own sovereignty or compliance statement, quoted or closely paraphrased.
- Not published
- Software catalogue Pre-configured application catalogue a platform provider advertises.
- Not published
- Runs on Clouds and hardware partners a platform provider lists.
- Not published
- Pricing model
- hourly + term blocks
- From
- $4.79 GPU-hour
- Note
- NVIDIA HGX B200 'as low as' $4.79/GPU-hr on cirrascale.com/pricing (2026-11-09); page says prices are examples and term blocks save up to 20%, so not confirmed as on-demand list price (list_price_b200_hr left null). No H100 listed. Others: B300 $5.69, GB300 $8.09, MI355X $4.29, MI350 $3.89. Founded = cloud division became a separate company in early 2017 (about page). ClusterMAX 3.0 lists it as 'Unavailable' (not rated).
Sources
- Cirrascale pricing accessed 2026-11-09
- Cirrascale homepage accessed 2026-11-09
- Cirrascale about accessed 2026-11-09
- SemiAnalysis ClusterMAX 3.0 (Sep 2026) rankings image accessed 2026-11-09
- Cirrascale pricing accessed 2026-11-11
Related providers and lists
Get a quote from Cirrascale Cloud Services and similar providers
Tell us what you need once. We send it to matching providers and they reply to you directly.