İçeriğe geç
Tres Teknoloji
Services / Compute / GPU Server Hourly GPU, live in a minute

Hourly GPU servers for training and inference

Virtual machines powered by NVIDIA L4, L40S and H100 accelerators. CUDA, cuDNN and drivers come pre-installed; power it off when you are done and billing stops.

On quote
GPU server
H100 · training/inference
live
GPU
98%
epoch
H100 · 80 GBNVLinkCUDA ready
Deep-learning model training
Pricing

A quote for your needs

The exact price is set by your needs and scale. Get a quote or schedule a meeting with your account manager.

Get a quote for a custom configuration
Plan GPU VRAM vCPU RAM Price
g1.l4 1× L4 24 GB 8 32 GB On quote
g1.l40s 1× L40S 48 GB 16 64 GB On quote
g1.h100 1× H100 80 GB 24 128 GB On quote
g2.h100x4 4× H100 320 GB 96 512 GB On quote
Prices exclude VAT and are quoted based on your needs.
Features

From model training to inference, your GPU workflow is ready

Preloaded CUDA stack

Start in minutes with images that already have CUDA, cuDNN and the NVIDIA drivers installed.

Single-tenant GPU

The GPU is not split or shared; the VRAM and cores are fully allocated to you.

NVLink multi-GPU

On multi-GPU nodes the accelerators talk over NVLink at high bandwidth.

Local NVMe dataset

Put the training dataset on the local NVMe disk so disk I/O does not limit your training speed.

Containers and Jupyter

The NVIDIA Container Toolkit is ready; run your own image or a ready-made Jupyter environment.

Connect to object storage

Write checkpoints and model weights directly to S3-compatible object storage.

S.S.S.

Frequently asked questions about GPU Server

We offer the NVIDIA L4, L40S and H100 accelerators. Multi-GPU (e.g. 4× H100) configurations are provided on request with a quote.

Yes. Our ready-made images ship with the current NVIDIA driver, CUDA and cuDNN; you can also open PyTorch and TensorFlow environments in one click.

No. The GPU you reserve is single-tenant; the VRAM and cores are not shared with another customer and you experience no noisy-neighbor effects.

Stopping (powering off) the server does not stop billing; the GPU and disk resources stay reserved and billed until you delete it. When training ends, move your checkpoint to object storage and delete the server — that stops the GPU charge entirely.

Yes. We offer configurations with multiple H100s in a single node connected over NVLink; these multi-GPU options are provided on request with a quote.

For continuously running workloads we offer reserved GPUs with a monthly or yearly commitment; this provides a significant discount over the hourly price. For short jobs the hourly model is ideal.

Datasets and model weights are kept on the local NVMe disk or in S3-compatible object storage, on the infrastructure we operate in Tier III+ data centers, in Turkey.

Customer reviews

What do teams using GPU Server say?

“We run our model training on hourly H100 servers; when the job ends we power off and stop the cost. Because CUDA came installed, we spent no time on setup.”
AY Arda Yılmaz ML Engineer · Anatolia Analytics
“On L40S servers our inference latency dropped noticeably. Thanks to the single-tenant GPU we experience no performance fluctuation, predictable all day.”
BA Buse Aydın AI Lead · Meridyen E-ticaret
“We connected our dataset directly from object storage; we write checkpoints there too. Keeping the data in Turkey was essential for us.”
Emre Çelik Data Science Manager · Pusula Ödeme
“Getting GPU Server (NVIDIA-accelerated) up and running took minutes; the documentation is clear and the Turkish-language support made it easy.”
FY Furkan Yavuz Backend Team Lead · Loop Yazılım
“After switching to GPU Server (NVIDIA-accelerated) we measured the performance difference clearly in our application response times.”
DY Deniz Yılmaz DevOps Engineer · Trendhane
“We can reach the support team even in the middle of the night; response times are genuinely as fast as promised.”
Ece Şahin Software Team Lead · Bulut Market
“They gave us free consulting before the migration; together we worked out how much of each resource we needed.”
EK Elif Korkmaz IT Coordinator · Akademi Online
“The contract and SLA terms are clear; we built a transparent relationship from the start, not a negotiation.”
FY Furkan Yavuz Backend Team Lead · Loop Yazılım
“The price/performance balance came out better than overseas providers, and we are invoiced in TRY.”
DY Deniz Yılmaz DevOps Engineer · Trendhane
Related services
Cloud Server Live in seconds $0.0123/hr
Bare Metal Unshared physical hardware On quote
Kubernetes Managed Kubernetes $9.00/mo
Serverless Scales to zero On quote