PODs are easy to create with the provide API. Calculations are stable. However, streamlined process to create a volume on a data center which has available resources matching the calculation needs is... See more
While we don't verify specific claims because reviewers' opinions are their own, we may label reviews as "Verified" when we can confirm a business interaction took place. Read more
To protect platform integrity, every review on our platform—verified or not—is screened by our 24/7 automated software. This technology is designed to identify and remove content that breaches our guidelines, including reviews that are not based on a genuine experience. We recognise we may not catch everything, and you can flag anything you think we may have missed. Read more
Review summary
Created with AI, based on recent reviews
What people talk about most
Reviews shaping this summary
Company details
- #22 of 26 best companies in Cloud Computing Service
- #10 of 11 best companies in Computer Service
- #281 of 347 best companies in Software Company
Written by the company
Runpod is the AI Developer Cloud for teams building, training, and scaling AI applications. Developers use Runpod to move from experiment to production in one platform, with Pods for building, Serverless for shipping inference workloads, and Clusters for scaling larger deployments. Runpod gives AI and ML builders access to GPU and CPU compute without forcing them into a heavyweight cloud workflow. Teams can launch GPU environments, bring their own containers and frameworks, deploy serverless GPU endpoints, manage storage, monitor workloads, and run closer to users across global regions. Runpod supports a broad range of GPU options, from workstation-class cards to high-memory accelerators, so developers can choose the right compute for prototyping, fine-tuning, inference, rendering, research, and production workloads. Runpod is built for developers who want infrastructure that stays out of the way. The platform supports common AI workflows such as model training, fine-tuning, image and video generation, LLM inference, agent systems, and compute-heavy applications. Startups, AI-native teams, researchers, and infrastructure teams use Runpod when they need flexible cloud compute, straightforward APIs, container control, and a path from early experimentation to production deployment. For product help, billing questions, platform status, and account support, visit runpod.io or contact support@runpod.io. Customers can also find platform updates and status information from the Runpod website.
Contact info
1181 Nixon Dr. #1158, 08057, Moorestown, United States
- support@runpod.io
- runpod.io
Asks customers to review
This company invites their customers to review, whether positive or negative
Replied to 57% of negative reviews
Typically replies within 1 week
How this company uses Trustpilot
See how their reviews and ratings are sourced, scored, and moderated.
They take themselves seriously
They trust their own product enough to hand me a coupon when availability was limited. Of course, they wouldn't be handing out coupons if they expected their product to not hold up under almost all circumstances.

Reply from Runpod
Fast dedicated and solved the isse by…
Fast dedicated and solved the isse by going the extra mile

Reply from Runpod
Well run, runs well
Great prices. Responsive customer service. Had some annoying problems when instances switched to networked storage (really bad slow throughput) which took a little while to fix but mostly seems to be fixed now. Still, nothing's perfect, easily the best cloud compute service I've ever used. Easy to use and works well.

Reply from Runpod
Laughable
Gave it 2 chances, both using featured 'ready-to-use' images they offer as templates.
Try 1: 40 minutes just trying to install the venv. Once I checked the logs to see what on earth was going on, the download speeds were around 12 KB/s. Tried it locally to see if it was an issue with the pull, but it worked normally on my PC. Gave up and wasted 1 hour of my time.
Try 2: Chose a different template, and this time it took about 15 minutes to install. Then, when trying to upload my own data to the pod, it was simply a nightmare. Abysmal upload speeds, a broken file manager, and a remote console that likes to reset during the upload process, making the same file be uploaded twice, overwriting itself and duplicating bandwidth.
When it FINALLY was uploaded, the server kept detecting the file as temporary and it couldn't be used, and when I chose to relaunch the UI (a normal action), the whole server crashed in cascade (Cloudflare error included!), forcing me to stop the pod and lose all the data, time, and money invested in absolutely no effective work.
Result: Wasted $10.
One of the worst services I've ever used.

Reply from Runpod
Hours lost fighting storage & uploads — gave up
I wanted Runpod to be a simple “spin up GPU and go” service. Instead, I spent ~5 hours just trying to move files.
Network Volume (EU-RO-1): mounting via S3 was a mess. I generated S3 API keys multiple times, used the documented endpoint (s3api-eu-ro-1.runpod.io), Signature V4, path-style addressing, and set the region to eu-ro-1. Tried several clients (Cyberduck, RaiDrive, S3 Browser). I kept getting auth/region errors (first “region us-east-1 is wrong; expecting eu-ro-1”, then “Invalid object path”) despite correct Access Key ID (user_…) and Secret.
Cloud Sync: not visible on the volume page; only appears once a pod is running. Confusing UX if you just want to load data into storage.
Uploads: Jupyter upload crawled around ~22 Mb/s even though my upstream tests at ~800 Mb/s. No straightforward “Files” manager for the volume itself.
Costs: paid for a network volume (~$7/mo) and ended with unused credit (~$17) with nothing to show.
I eventually gave up. My suggestions: fix S3 in EU-RO-1 (or clearly document the exact, working settings), provide a native file manager for volumes, surface Cloud Sync consistently, and return clearer error messages when signatures/regions don’t match.
Bottom line: I wanted to run a project, not become an S3 detective. I can’t recommend Runpod based on this experience.

Reply from Runpod
Here a Runpod user from South Korea!
Here a Runpod user from South Korea!
I've been frustrated by lack of my PC's performance, especially lack of memory. This situation pushes me to cliff, which means I have to buy a high-performance PC, but I'm absolute broke.
Then I found this Runpod service. I literally bet whether it makes or not. And I win the bet.
Google Cloud or AWS is too heavy, it fits more a company, not personal.
Runpot has great capability, easy to understand whole pipeline to build my own AI training environment.
I will highly recommend it to whole people around of me.
Thank you for this god-like services.

Reply from Runpod
Great service.
The comfy ui template are not all…
The comfy ui template are not all updated and the price for storage i believe is too high. Hope you could make the price a bit cheaper for storage

Reply from Runpod
Nice experience.
How could you rate it anything else
How could you rate it anything else? Immediate spin up, i was connected via SSH and literally training within 30 minutes. Ive topped my account up with credit I WILL be back, this is game changing.
The service is simple and easy to use
The service is simple and easy to use. I'm able to spin up a POD in seconds to do work on my network drive then spin it down so I'm not paying for it. I'm able to use and pay only for the GPU I'm using.

Reply from Runpod
Stellar performance.
Stellar performance.
Highly recommend runpod. Had a little problem with network volume the other day but the support team resolved it swiftly. very good support!

Reply from Runpod
Outdated Documentation
The documentation for a lot of stuff seems not very up to date/is worded terribly/not detailed enough. Never seen a website that doesn't have up to date documentation. There's issues I have run into where I have fed page by page every single relevant piece of documentation to gemini 2.5 + Gpt 5 and ill spend half an hour troubleshooting and end up with just 10 more questions on how they thought running their website the way they do was a great idea.
To clarify, I am pretty new to using cloud GPU rentals, but I am not new to using docker or comfyUI, and this has still been a major headache.
Despite all of that though, when I do get stuff to run it works great, so if you can get it to work definitely a quality service.

Reply from Runpod
gpu taken away on shut down pods
you rent a pod and shut it down, you keep paying idle costs which is fine, but then when you start it again your gpu is taken and you can only start without gpu. Horrible waste of money.

Reply from Runpod
Best Compute
Great ui , instant gpus. For a medium tier this cloud hosting best place. Templates are great but some have issues which stops container .

Reply from Runpod
Too cumbersome to setup serverless and…
Too cumbersome to setup serverless and still find out the model isn't support and the service is crashing now and then

Reply from Runpod
Amazing...
I've been using RunPod for a few days now, mainly utilising RTX 5090 GPU's for some model training, I have to say, it is truly excellent. I only wish I knew about this service sooner. I am most definitely impressed and the pricing is fantastic.
All very easy to use, I have been using vs code with remote ssh, all works great, no issues; just like working locally, but on a seriously powerful machine :)

Reply from Runpod
Fast support, great infra
Their support is fast. The infra is really good and the UI interface and API endpoints are really easy to use.
At my startup we are using it for model training and we have absolutely no complaints.
I would love for them to have their own private Image Registry on which to upload docker images, but other than that they are really good.

Reply from Runpod
The Trustpilot Experience
Anyone can write a Trustpilot review. People who write reviews have ownership to edit or delete them at any time, and they’ll be displayed as long as an account is active.
We use dedicated people and clever technology to safeguard our platform. Find out how we combat fake reviews.
Learn about Trustpilot’s review process.
Here are 8 tips for writing great reviews.
Verification can help ensure real people are writing the reviews you read on Trustpilot.
Offering incentives for reviews or asking for them selectively can bias the TrustScore, which goes against our guidelines.