The ultimate in cloud-based machine learning and high performance computing
From IoT services and virtual workstations to AI image generation and HPC computations, our next level GPU-based solutions deliver scalable and cost-effective performance from the cloud.
Supercharge Your AI and Machine Learning Projects
Accelerate your AI and machine learning projects with state-of-the-art GPUs. Designed to handle complex computations at lightning speeds, enabling faster model training and more efficient algorithm development.
Accelerate Your Innovation with Cloud GPUs
Whether you’re training complex machine learning models, rendering high-definition graphics, or analysing large datasets, our GPUs attached to virtual machines offer the scalability and flexibility to meet your evolving needs.
Unleash the Power of GPUs in the Cloud
Experience unparalleled speed and efficiency with cloud-based GPUs. Ideal for everything from AI and machine learning to 3D rendering and scientific simulations, providing the computational power you need to accelerate your most demanding tasks.
Fully Customisable for Maximum Flexibility
Our range of on-demand GPU options can be attached to any Rock or Boulder virtual machine for any duration. Simply launch your chosen system size, attach your chosen GPU, and unlock the full power of NVIDIA Tensor, CUDA, and Ray Tracing cores.
Looking for high-CPU virtual machines?
Pair your GPU with any of our high-frequency CPU instances to power the most demanding workloads.
GPU Pricing
Only pay for the time a GPU is attached to a virtual machine. GPUs can be detached from virtual machines at any time.
| GPU | Performance | Tensor Cores | CUDA Cores | Memory | Bandwidth | Price |
|---|---|---|---|---|---|---|
| Nvidia T4 | 8.1 TFLOPS (FP32) | 320 | 2,560 | 16GB | 320GB/s | £201.60/mo£0.28/hr |
| Nvidia L4 | 30.3 TFLOPS (FP32) | 300 | 7,424 | 24GB | 300GB/s | £268.80/mo£0.37/hr |
| Nvidia A100 | 312 TFLOPS (TF32) | 432 | 6,912 | 80GB | 1,935GB/s | Coming soon |
| Nvidia H100 | 989 TFLOPS (TF32) | 528 | 16,896 | 80GB | 3.35TB/s | Coming soon |
| Nvidia H200 | 989 TFLOPS (TF32) | 528 | 16,896 | 141GB | 3.8TB/s | Coming soon |
Start building today with £100 free credit
You’ll be up and running in less than a minute.
The Krystal Difference
Hardware
Industry-leading CPUs
Capable of the most strenuous workloads thanks to high-frequency AMD EPYC chips. High-memory solutions are also available.
Clustered Storage
We utilize a distributed storage architecture with no single points of failure and store three replicas of system disks at no extra cost.
High Availability
Engineered for resilience, node failover ensures vital compute resources stay online.
Network
Pooled Outbound Transfer
Each virtual machine provides free transfer dependent on its size, pooled across all VMs in your organization for maximum flexibility.
Inclusive DDoS Protection
Real-time, fully automated DDoS protection is included with every virtual machine.
Free Inbound/Internal Transfer
Transfer to our data centres and internal transfer between storage locations or VMs comes at no cost.