TensorDock

What is TensorDock?

TensorDock is a GPU cloud marketplace where you rent virtual machines by the hour for AI training, inference, rendering and cloud gaming. Independent hosts list their hardware on the platform, so the same GPU can carry different prices depending on location and redundancy. Servers run on KVM virtualization with dedicated GPUs and root access, Docker is included on VM templates, and Windows 10 is supported. The company also sells CPU-only virtual machines and a container platform for inference that is still in beta. You deposit funds, usage is deducted continuously, and servers are deleted automatically if the balance reaches zero. A REST API lets developers script deployments, which cloud gaming and rendering companies on the site use for Windows machines. TensorDock pricing starts around $0.10 per hour for smaller GPUs, with the H100 SXM5 80GB listed at $2.25 per hour, and reserved pricing is available by contacting sales.

Use cases

• Rent a GPU server for model training
• Run LLM inference on a budget
• Render animations on cloud GPUs
• Host Windows cloud gaming machines
• Deploy GPU servers through an API
• Run batch and transcoding jobs on CPU VMs

Who it's for

• ML engineers
• Indie developers
• University researchers
• Startups

Why use TensorDock

• Starts small: you can deposit as little as $5 and launch a GPU server in about 30 seconds.
• No quotas or long contracts, with reserved pricing available if you later need long-term capacity.
• A wide range of GPU models lets you match cost to workload, from consumer cards to H100s.
• Root access and dedicated GPUs mean you control the full software stack on each VM.
• Bandwidth is included, so there are no ingress or egress fees on the GPU cloud.

TensorDock features

• On-demand GPU servers from RTX A4000 to H100 SXM5
• KVM virtualization with dedicated GPUs and root access
• Docker included on all VM templates
• Windows 10 support for rendering and cloud gaming
• Per-second billing with prepaid balance
• REST API for deploying and managing servers
• CPU-only VMs on Intel Xeon and AMD EPYC
• Container platform for inference (beta)
• Hosts vetted to a 99.99% uptime standard

How to get started with TensorDock

1. Sign up for an account on tensordock.com.
2. Deposit funds, starting from as little as $5.
3. Pick a GPU model, location and resources on the dashboard or through the API.
4. Deploy the server from a template and connect with root access.
5. Run your workload, then delete the server to stop billing.

TensorDock limitations

• Prices come from independent hosts, so the same GPU can cost different amounts
• Servers are deleted automatically when the prepaid balance reaches $0
• You manage your own drivers, frameworks and operating system
• The pricing page was last updated July 24, 2024 and points to the dashboard for live rates

TensorDock pricing

TensorDock has no free plan. Plans start at $0.1 usage-based, according to the official pricing page on 24 Oct 2026.

TensorDock plans

• H100 SXM5 80GB

  $2.25/hour

  Typical price, varies by host

  • Training and inference
  • Billed per second

• A100 SXM4 80GB

  $1.80/hour

  Typical price, varies by host

  • AI inference
  • Billed per second

• RTX 4090 24GB

  $0.35/hour

  Typical price, varies by host

  • Image processing, gaming, rendering

• RTX A4000 16GB

  $0.10/hour

  Typical price, varies by host

  • Entry-level GPU workloads

• CPU Cloud

  From $0.012/hour

  • Xeon and EPYC CPUs
  • Transcoding and batch jobs

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

TensorDock alternatives

The main TensorDock alternatives listed on Debuly are Vast.ai, Runpod, Lambda, CoreWeave, Crusoe Cloud and Novita AI. The table compares pricing and platforms; order is A–Z.

1. 

   CoreWeave

   AI cloud that rents NVIDIA GPU capacity by the hour, with on-demand, spot and reserved plans for model training and inference.

   Developer platformsCoding
   PaidVisit CoreWeave (official site)
2. 

   Crusoe Cloud

   AI cloud with on-demand GPU instances, managed Kubernetes and Slurm, and managed inference and fine-tuning for open models.

   Developer platformsCoding
   PaidVisit Crusoe Cloud (official site)
3. 

   Lambda

   GPU cloud for AI with on-demand NVIDIA instances and reserved 1-Click Clusters for training, fine-tuning and inference.

   Developer platformsCoding
   PaidVisit Lambda (official site)
4. 

   Novita AI

   AI cloud with pay-per-token model APIs for 200+ models, dedicated GPU endpoints, agent sandboxes and GPU rentals.

   Developer platformsCodingAutomation and agents
   PaidVisit Novita AI (official site)
5. 

   Runpod

   Pay-per-second GPU cloud for AI: on-demand Pods, Serverless endpoints, multi-node Clusters and ready-made public model APIs.

   Developer platformsCoding
   PaidVisit Runpod (official site)
6. 

   Vast.ai

   GPU cloud marketplace for renting on-demand, interruptible or reserved GPUs with per-second billing.

   Developer platformsAutomation and agents
   PaidVisit Vast.ai (official site)

TensorDock FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from tensordock.com · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.