Utilizing 100% renewable energy
View more
The ML Cloud
Premium dedicated GPU servers and clusters
Model inference services
Easy sign up · Pay as you go
![](/_image?href=%2Fassets%2Fhero.C2uJVEAp.png&f=webp)
Partners who trust our services:
![Product](/_image?href=%2Fassets%2Finstance.BeFejk4H.png&h=176&f=webp)
GPU Instances
On-demand and long-term instances on DataCrunch infrastructure
![Product](/_image?href=%2Fassets%2Fcluster.D17-all1.png&h=176&f=webp)
GPU Clusters
Custom built GPU clusters, managed by DataCrunch with custom software stack
![Product](/_image?href=%2Fassets%2Finference.BAT2W_zh.png&h=176&f=webp)
Serverless inference
Autoscaling and low-cost deployment for your models with zero setup
![Product](/_image?href=%2Fassets%2Fh100-cluster.BQSHrRO0.png&h=288&f=webp)
H100 clusters
At the forefront of digital intelligence Our clusters exclusively use the NVIDIA H100 SXM5 80GB NVLINK modules. Via NVLINK, the H100 achieves a chip-to-chip interconnect bandwidth of 900GB/s and leverages a 3200gbit/s InfiniBand™ interconnect. Contact us for more information.
Fast Dedicated hardware for max speed and security
Productive Start, stop, hibernate instantly via the dashboard or API
Reliable A historical uptime of over 99.9%
Protected DataCrunch is ISO27001 certified
Price Calculator
Number of GPUs
1x H100 SXM5 80GB
- 30 CPUs
- 120 GB RAM
- 80 GB GPU RAM
This setup is best for:
Gargantuan ML modelsMulti-GPU trainingFP64 HPCNVLINK
- $3.17/h
- $3.07/h
- $3.04/h
- $2.92/h
- $2.38/h
Price on demand
All infrastructure powered by renewable energy
High-end GPU instances
Order and access your GPU in just minutes via our intuitive dashboard or API.
Your machine is automatically setup according to your preferences and you receive access in a few minutes.