tensordock.comReviews of tensordock.com

As so many others have noted, this is a truly terrible company to do business with. They claim that they have thousands of GPU instances around the world, but when you come to actually launch one of them, there isn't any inventory available. I tried on multiple occasions and was happy to accept any GPU in any location, but they could never provide one for me. I contacted them asking for a refund and, not surprisingly, they just ignored me.

Worst provider ever! I couldn't start a VM for 2 days. After finally getting a Windows VM up and running, I still couldn't connect to it via RDP. Customer support remains silent. Don't even bother adding funds.

Just Bad!!! they take your money and you are not even able to deploy a server and forget something called customer support it doesn't even exist you have more chances marrying Sydney Sweeney then to get either a decent experience or you money's worth

scam!!!!!!

No answer from support. total scam , they never refund your money. Stay away !! Lost a lot of work because of these scammers

TensorDock looked promising initially, but the overall experience turned into a frustrating waste of time. The VM behaved inconsistently despite showing healthy resources on the dashboard, and after spending hours configuring CUDA, PyTorch, ComfyUI, the instance suddenly lost access to the GPU entirely — `nvidia-smi` could no longer communicate with the NVIDIA driver even though the rental page still showed an active RTX A4000 attached. The environment felt unstable and poorly maintained for serious GPU workloads, with mismatched drivers, missing packages, inconsistent system behavior, and unexplained failures after rebooting or configuring services. What made the experience even worse was the complete lack of support — emails went unanswered, there was effectively no response from the team, and there was no meaningful help while the VM was failing. For anyone trying to run production or long-running AI workloads like ComfyUI/Flux remotely, the instability combined with nonexistent support makes this platform very difficult to trust.

I attempted to create a VM equipped with an RTX 5000 Ada GPU. However, the machine failed to boot properly. The status indicator remained red, and I was completely unable to access or connect to the instance. Assuming it was a temporary glitch, I waited for over an hour for the provisioning to complete, but the status did not change. I then deleted that VM and deployed a completely new one with the exact same RTX 5000 Ada configuration. Unfortunately, the new instance is stuck in the exact same red error state.

I rented a server with with A6000 GPU, 64 GB RAM and 500 GB disk. From the start on, the GPU was offline on Docker level. I sent many support queries, to no avail. Most of the time when I tried to start the server, it was not available. In times I managed to start it, the GPU was offline. So the credit I loaded slowly depleted itself as they charge for the disk space. Some guy once responded to the online chat saying he will personally check my case, but never came back. They basically drained all money I loaded without providing any service.