TensorDock
A distributed cloud marketplace for renting GPU, CPU, and Windows virtual machines for compute-heavy workloads.
About TensorDock
TensorDock is a cloud infrastructure marketplace for renting GPU, CPU, and Windows virtual machines across a distributed network of independent hosts. It is designed for machine-learning training, animation rendering, cloud gaming, generative-art workloads, and other compute-heavy jobs that benefit from access to GPU capacity. Hosts set their own prices, so otherwise similar hardware can carry different rates depending on location, redundancy, and infrastructure type. Customers fund an account and are billed continuously after a server is deployed; a server is automatically deleted when the balance reaches zero. The service supports pay-as-you-go use, while customers planning longer rentals can contact TensorDock about reserved pricing. Its marketplace spans both higher-security data-center hardware and lower-cost converted mining rigs, giving buyers a tradeoff between infrastructure assurances and price-to-performance. TensorDock also highlights compatibility with TensorFlow and PyTorch, and its virtual-machine model can support Windows-based creative or gaming workflows. The practical fit is strongest for teams that want flexible compute without hyperscaler quotas and can compare host, region, redundancy, and price before deployment.
Best for
Best use cases
Deploy a GPU server for a training run and pay from a deposited balance while it remains active.
Render animation work on a Windows virtual machine while keeping related CPU workloads on the same cloud.
Add cloud-gaming capacity during a demand peak without depending on a fixed hardware allocation.
Compare before choosing
Marketplace pricing creates more host and location choices than a single-provider inventory, but rates vary.
Data-center hardware emphasizes security and reliability, while converted mining rigs emphasize price-to-performance.
Pay-as-you-go access supports short workloads, while long-term rentals require a reserved-pricing conversation.
Features
Machine-learning training
Deploy GPU compute for training machine-learning models with TensorFlow or PyTorch workflows.
Animation rendering
Run animation and creative rendering workloads on GPU, CPU, or Windows virtual machines.
Cloud gaming capacity
Deploy Windows virtual machines for cloud gaming and expand capacity when demand increases.
Pros
- Marketplace inventory offers multiple locations and hardware options for compute-heavy workloads.
- Pay-as-you-go billing lets teams fund short deployments without committing to a fixed term.
- GPU, CPU, and Windows virtual machines support training, rendering, and cloud-gaming workflows.
Cons
- Independent hosts set prices, so similar GPU or CPU configurations can carry different rates.
- A server is automatically deleted when the deposited account balance reaches zero.
- Infrastructure ranges from certified data centers to converted mining rigs, requiring a deliberate tradeoff between assurances and cost.
Pricing Plans
Pay as you go
- Deposit funds before deployment
- Balance is deducted continuously
- Servers are deleted when the balance reaches zero
- Host prices vary by location and redundancy
* Pricing information is for reference only. Please refer to the latest announcement on TensorDock's official website.