TensorDock
Marketplace for on-demand GPU servers and cloud instances
TensorDock operates a global marketplace connecting users with independent hosts offering GPU servers. It provides access to over 30,000 accelerators across more than 100 locations, aiming to lower costs for AI training and inference compared to traditional hyperscale clouds.
- Access to up to 30,000 GPUs via partner hosts
- Servers deployed in seconds with no long-term contracts
- Supports Nvidia H100, A100, and consumer RTX cards
- Hosts vetted for 99.99% uptime and security standards
- Pay-as-you-go pricing with no hidden fees or quotas
What it is
TensorDock is a cloud computing marketplace that aggregates GPU resources from independent hosts. It offers virtual machines and bare-metal servers for artificial intelligence workloads, gaming, and rendering. The platform claims to provide access to up to 30,000 graphics processing units distributed across more than 100 locations in over 20 countries. Users can deploy servers quickly without long-term commitments or complex quotas, paying only for the time they use.
The compute
The platform supports a wide range of Nvidia hardware, including consumer-grade RTX 4090 cards and enterprise HGX H100 SXM5 modules. It offers both Linux and Windows virtual machines with root access, allowing users to manage their own drivers and software environments. Docker is included in all templates to simplify deployment. The system uses a multithreaded stack designed for rapid virtual machine provisioning, often completing launches in under a minute.
Where the GPUs are
Compute resources are hosted in Tier 3 and Tier 4 data centres globally. TensorDock vets each host for hardware quality and technical capability, enforcing a 99.99 per cent uptime standard. Hosts must schedule maintenance at least two weeks in advance. The company states that most hardware now resides in certified facilities, with security measures including revoked SSH access for hosts to prevent unauthorised data viewing.
Customers and deals
Clients include AI startups, academic researchers, and creative agencies. Florida State University professor Skyler Liang uses the service for generative adversarial network research due to its cost efficiency. AirGPU leverages the platform to handle peak demand for cloud gaming services. Creavite utilises CPU-only instances and Windows virtual machines for rendering workflows and logo animations.
Money and ownership
TensorDock operates on a pay-as-you-go model where users deposit funds and are charged continuously. Prices vary because independent hosts set their own rates based on location and redundancy. The company claims its rates are up to 80 per cent lower than major cloud providers. H100 instances start at $2.25 per hour, while consumer GPUs like the RTX 4090 are available from $0.12 per hour.
Its place in AI
The service positions itself as an alternative to hyperscale clouds by offering greater flexibility and lower prices. It serves niches such as batch processing, transcoding, and test workloads. By aggregating supply from converted mining rigs and professional data centres, it aims to balance price and performance for builders who need shell access and dedicated hardware without large upfront investments.
Questions people ask about TensorDock
How does TensorDock pricing work?
Prices are set by independent hosts who compete on the marketplace. This structure allows users to find the best rates for their specific needs. Costs vary by location, hardware type, and redundancy levels. Users pay hourly from a deposited balance, with servers deleting automatically when funds reach zero.
What types of GPUs are available on the platform?
The marketplace offers over 45 GPU models, ranging from consumer cards like the RTX 3090 and 4090 to enterprise hardware such as the Nvidia A100 and H100. This variety supports diverse workloads including gaming, image rendering, and large-scale AI model training or inference tasks.
Is my data secure on TensorDock servers?
Most hardware is hosted in certified Tier 3 or 4 data centres. TensorDock revokes SSH access for hosts to prevent them from viewing customer data. An agent monitors host nodes for suspicious login activity. The company also requires hosts to meet strict uptime and maintenance scheduling standards.
Good reading and sources
There is no English Wikipedia article on TensorDock yet, so the facts on this page come from the sources listed above: tensordock.com/. Some are the subject's own pages, so its claims about itself are reported as its claims. EveryNeocloud wrote the text in its own words, with AI assistance, and compares it with the sources to make sure no passage is copied.
It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.