TensorPool is a GPU infrastructure tool that allows users to deploy GPU clusters on demand. It provides simple, fast, and affordable GPU infrastructure for machine learning training and inference.
TensorPool offers on-demand GPU clusters for machine learning training and inference, enabling users to deploy multi-node GPU clusters in seconds. It provides high-performance shared storage and colocated object storage, along with high-speed interconnects for distributed training. The platform supports various GPU types, including B200 SXM, H200 SXM, B300 SXM, H100 SXM, and L40S, with competitive hourly pricing. Users can access up to 128 GPUs on the standard plan and scale to thousands of GPUs with enterprise solutions. TensorPool aims to simplify the process of acquiring and managing GPU resources for developers and data scientists.
Best used for
Ideal for developers and data scientists who need to quickly provision GPU resources for machine learning training, deploy multi-node GPU clusters for distributed workloads, and manage high-performance storage solutions. Especially valuable for those seeking affordable and scalable GPU infrastructure without the complexities of traditional cloud providers.
Common actions
deploy GPU clusters
train machine learning models
manage GPU infrastructure
run distributed training
AI developmentAI infrastructureMLOpsresource managementDeep learningdistributed computingmachine learningDeveloper Tools
Capabilities
Key features
On-demand GPU clusters
High-performance shared storage
High-speed interconnects
CLI & API access
Distributed training support
Scalable GPU infrastructure
Target Audience
developerdata scientist
Integrations
Not yet documented
Pricing & Plans
Paid · Usage-based · Enterprise
Not Disclosed
FAQs
What types of GPUs are available on TensorPool and what are their hourly costs?
TensorPool offers a range of GPUs including B200 SXM ($4.99/hr), H200 SXM ($2.99/hr), B300 SXM ($5.49/hr), H100 SXM ($1.99/hr), and L40S ($1.49/hr). CPU usage is also available at $0.015/hr, providing flexible options for various ML workloads.
Does TensorPool offer shared storage and what are its costs?
Yes, TensorPool provides high-performance shared storage at $100 per TB per month, with speeds up to 300 GB/s read and 150 GB/s write. Object storage is also available at $50 per TB per month, with no egress or ingress charges.
What are the key differences between the Standard Plan and Enterprise Plan?
The Standard Plan offers on-demand access to up to 128 GPUs, 24/7 support, and shared storage. The Enterprise Plan provides custom solutions for thousands of GPUs, dedicated support, custom storage, analytics, monitoring, and spend controls, with a 99.9% Uptime SLA.
Shypd.aiThe largest open index of AI tools
