Dedicated accelerator compute
Rent server capacity for model training and inference, with published GPU configurations, system memory, local storage, and networking specifications.
Cirrascale
Paid
Dedicated AI cloud compute, storage, networking, and deployment services for training and inference.
O que é Cirrascale?
Cirrascale provides dedicated cloud compute, storage, and networking for AI training and inference. Teams can select accelerator servers, bring their applications and data, and arrange the configuration or managed support their workload needs. The cloud service supports hardware options from NVIDIA, AMD, and Qualcomm, with published server configurations and term-based prices.
Teams can also arrange the separately scoped Cirrascale Inference Platform for model routing, endpoint management, private RAG chat, and consumption controls.
O que Cirrascale pode fazer?
Storage and networking
Use multi-tiered storage and high-bandwidth networking for AI workloads. Cirrascale does not charge ingress or egress data-transfer fees.
Scoped deployment support
Arrange professional or managed services around your configuration. Support beyond your service plan can carry additional fees.
Inference Platform option
Use the separately scoped platform for model provisioning, routing, RAG-assisted chat, token controls, and central endpoint administration.
Casos de uso
“Dedicated AI workloads”
Run training or inference applications on reserved accelerator servers, selecting a configuration for your workload and project term.
“Capacity and cost planning”
Compare per-server monthly prices, commitment periods, and storage tiers before reserving capacity.
“Enterprise model serving”
Evaluate the Inference Platform when your deployment needs model endpoints, private knowledge chat, or team-level AI consumption controls.
Perguntas frequentes
Can I rent servers by the hour?
No. Cirrascale does not provide hourly server rentals. GPU-hour equivalents are comparison figures, while server prices are quoted per server per month with the selected commitment period.
How should I read the published server prices?
For example, an eight-GPU AMD MI300X server is listed at $22,499 per month on a monthly term, or $17,999 per month with an annual term. The annual-term figure is not the total yearly bill. Three- and six-month terms are also listed for this configuration.
Are transfer fees and all other costs included?
There are no ingress or egress data-transfer fees. Storage, configuration add-ons, out-of-plan professional services, and applicable taxes can be separate; card or PayPal payments may carry an additional transaction fee under the terms.
Who manages applications and backups?
For dedicated servers, you receive administrator access and take responsibility for applications, data security, and independent backups. Managed services should be agreed within your service scope; standard hosting is not a promise of fully managed applications.
What payment terms apply?
The service terms state that payments are non-refundable. Recurring plans can have an upfront charge followed by agreed recurring charges, so confirm the commitment and cancellation notice when ordering.
Informações rápidas
- Product type
- Dedicated AI cloud compute and related services
- Core workloads
- Training and inference
- Server billing
- Per server per month; commitment-dependent pricing
- Hourly rentals
- Not offered
Análise de tráfego de Cirrascale
Os dados de tráfego são apenas para referência.
- Classificação do domínio
- 59
Alternativas ao Cirrascale
Scaleway GPU Instances
www.scaleway.com
Scaleway GPU Instances provide on-demand GPU compute for AI training, fine-tuning, inference, video, and rendering workloads. Teams can select among GPU families and provision instances through the console, CLI, or Terraform.
NVIDIA DGX Cloud Lepton
lepton.ai
NVIDIA DGX Cloud Lepton connects developers to a global network of GPU compute through a cloud-based offering. The available source positions it for developers who need access to accelerated computing resources.