GPU Cloud
(Early Access)
Rent the GPU you need by the hour to train or run AI models, or to render. You are invoiced from Turkey and work with an engineer who speaks your language. Leave your request for early access and we will reserve capacity for you.
Key Features
- A wide choice from RTX 4090 to H100, H200 and B200
- Hourly usage; start and stop as needed
- SSH and Jupyter access, ready PyTorch / CUDA templates
- Invoiced in Turkey, no foreign card payments
- Engineering support for setup and environments
- Monthly reserved capacity for long-running work
How It Works
Your GPU server is ready in 4 steps.
Leave Your Request
Tell us the GPU model, quantity, duration and use case.
Quote and Planning
We check capacity and send a written price and start date.
Environment Setup
Your server starts with your chosen template; we send SSH or Jupyter details.
Use and Scale
Stop when done or add more GPUs; usage is billed monthly.
Use Cases
GPU servers are most often used for these workloads.
LLM Fine-Tuning
Train open models (Llama, Mistral, Qwen and others) on your own data.
AI Inference
Serve your models as an API for chatbots and assistants.
Image and Video Generation
Produce with Stable Diffusion, ComfyUI and video models.
3D Rendering
Cut render times with Blender and similar tools.
Academic Research
Access powerful GPUs for research without buying hardware.
Data Science
Speed up model experiments on large datasets.
Technical Specifications
$0.74 / hour
$1.11 / hour
$0.74 / hour
$1.36 / hour
$3.11 / hour
$2.24 / hour
$4.99 / hour
$6.61 / hour
$9.99 / hour
Per GPU, starting prices, excl. VAT; current price and capacity confirmed in the quote
Frequently Asked Questions
Everything you need to know about GPU cloud rental.
How are prices set?
Prices are per GPU per hour, excluding VAT: for example RTX 4090 from $1.11, A100 80 GB $2.24, H100 80 GB $4.99. Because GPU capacity and prices change, we confirm the exact price in a quote. Monthly reserved usage is discounted.
Where is data processed?
GPU servers run in our partner's data centers abroad; billing and support are provided by CloudRUX from Turkey. If you process personal data, this may count as a cross-border transfer under KVKK, so we recommend it for workloads without personal data or with anonymised data, and we help with standard contract processes where needed.
What does early access mean?
We bring the GPU service online based on demand. When you fill in the form we record your needs and let you know capacity and a start date. Applying creates no payment obligation.
How do I access the server?
Via SSH and Jupyter in the browser. PyTorch and CUDA templates are ready, or you can use your own container image.
Is my data persistent?
Temporary disks are deleted when the server stops. For models and datasets you want to keep, we attach a persistent volume or you can back up to our S3 storage.
How am I invoiced?
Usage is calculated monthly and CloudRUX issues an e-invoice from Turkey. Prices are in USD and converted to TRY at the invoice-date rate.
Get Started Today
Tell us which GPU you need and for how long; we'll check capacity and reply within one business day.
Contact Us