We've got your use case covered
AI Exploration
Efficiently and easily explore paths to incorporate AI into your business
H100 SXM, H200 SXM & A100 SXM instances with InfiniBand networks speed up distributed training on clusters ranging from 8 to 8k+ GPU's
Model Training
Optimize the development or refinement of the largest AI models
H100 SXM & A100 SXM instances with InfiniBand networks speed up distributed training
Scalable Inference
Integrate AI models into real-time applications with high reliability at any level of demand
H100, A100 & A40 GPUs enable 20B+ parameter model inference on a single card
GPU Compute
| GPU Compute | On-Demand | 6-month reserved | 1-year reserved | 3-year reserved |
|---|---|---|---|---|
| H200 141GB SXM | $ 4.29 /hr | $ 3.43 /hr | $ 3.22 /hr | $ 2.79 /hr |
| H100 80GB SXM | $ 3.90 /hr | $ 3.12 /hr | $ 2.93 /hr | $ 2.54 /hr |
| MI300x 192GB OAM | Contact Us | |||
| A100 80GB SXM | $ 1.95 /hr | $ 1.56 /hr | $ 1.46 /hr | $ 1.27 /hr |
| A100 80GB PCIe | $ 1.65 /hr | $ 1.32 /hr | $ 1.24 /hr | $ 1.07 /hr |
| A100 40GB PCIe | $ 1.45 /hr | $ 1.16 /hr | $ 1.09 /hr | $ 0.94 /hr |
| L40S 80GB PCIe | $ 1.00 /hr | $ 0.80 /hr | $ 0.75 /hr | $ 0.65 /hr |
| A40 48GB PCIe | $ 1.10 /hr | $ 0.88 /hr | $ 0.83 /hr | $ 0.72 /hr |
NVIDIA H200 Tensor Core GPU
On-Demand
$ 4.29/hr
6-month reserved
$ 3.43/hr
1-year reserved
$ 3.22/hr
3-year reserved
$ 2.79/hr
NVIDIA H100 Tensor Core GPU
On-Demand
$ 3.90/hr
6-month reserved
$ 3.12/hr
1-year reserved
$ 2.93/hr
3-year reserved
$ 2.54/hr
MI300x
On-Demand
Contact Us
6-month reserved
Contact Us
1-year reserved
Contact Us
3-year reserved
Contact Us
NVIDIA A100 Tensor Core GPU
On-Demand
$ 1.95/hr
6-month reserved
$ 1.56/hr
1-year reserved
$ 1.46/hr
3-year reserved
$ 1.27/hr
NVIDIA A100 Tensor Core GPU
On-Demand
$ 1.65/hr
6-month reserved
$ 1.32/hr
1-year reserved
$ 1.24/hr
3-year reserved
$ 1.07/hr
NVIDIA A100 Tensor Core GPU
On-Demand
$ 1.45/hr
6-month reserved
$ 1.16/hr
1-year reserved
$ 1.09/hr
3-year reserved
$ 0.94/hr
NVIDIA L40S Tensor Core GPU
On-Demand
$ 1.00/hr
6-month reserved
$ 0.80/hr
1-year reserved
$ 0.75/hr
3-year reserved
$ 0.65/hr
NVIDIA A40 GPU
On-Demand
$ 1.10/hr
6-month reserved
$ 0.88/hr
1-year reserved
$ 0.83/hr
3-year reserved
$ 0.72/hr
CPU Compute & Storage
| CPU Compute & Storage | On-Demand | 6-month reserved | 1-year reserved | 3-year reserved |
|---|---|---|---|---|
| General-Purpose CPU per vCPU/hour | $ 0.04 | $ 0.03 | $ 0.03 | $ 0.03 |
| Storage-Optimized CPU per vCPU/hour | $ 0.09 | $ 0.07 | $ 0.07 | $ 0.06 |
| Block Storage per GiB/month | $ 0.08 | $ 0.08 | $ 0.08 | $ 0.08 |
| Shared Disks per GiB/month | $ 0.07 | $ 0.07 | $ 0.07 | $ 0.07 |
General-Purpose CPU
per vCPU/hourOn-Demand
$ 0.04
6-month reserved
$ 0.03
1-year reserved
$ 0.03
3-year reserved
$ 0.03
Storage-Optimized CPU
per vCPU/hourOn-Demand
$ 0.09
6-month reserved
$ 0.07
1-year reserved
$ 0.07
3-year reserved
$ 0.06
Block Storage
per GiB/monthOn-Demand
$ 0.08
6-month reserved
$ 0.08
1-year reserved
$ 0.08
3-year reserved
$ 0.08
Shared Disks
per GiB/monthOn-Demand
$ 0.07
6-month reserved
$ 0.07
1-year reserved
$ 0.07
3-year reserved
$ 0.07
Networking
InfiniBand Networks
Highest throughput & lowest latency GPU RDMA
VPC Networks
Fully non-blocking Ethernet fabric in each zone
VPC Firewall Rules
Seamlessly manage network access to your VMs
DNS
Easy access to instances with internal & external addresses
Developers Love Crusoe Cloud
Crusoe has not only provided us with a robust infrastructure platform for our machine learning training and inference workloads but a management console that is intuitive and easy to use. With Crusoe Cloud we're able to get jobs up and running within a couple of hours.
Aleks Kamko
AI Researcher
Codeium is proud to run distributed training on Crusoe Cloud using a cluster of their newly deployed NVIDIA HGX H100s with 3,200Gbps InfiniBand.
Varun Mohan
Co-Founder
We chose Crusoe Cloud because of its simple user interface, reliability, and positive environmental impact. Crusoe makes it extremely easy to set up an account, spin-up a virtual machine (VM), and start training ML models right away. We were able to select and spin-up an 8 x NVIDIA A40 48GB node in under three minutes.
SangEon Park
MIT Researcher
Recent News
October 15, 2024
Crusoe, Blue Owl Capital and Primary Digital Infrastructure Enter $3.4 billion Joint Venture for AI Data Center Development
August 16, 2024
100 Jobs Coming to Tulsa through AI Tech Firm's Expansion Plan
July 18, 2024
Crusoe to Build Initial 200 MW AI Data Center With Plans to Expand at 1.2 GW Lancium Clean Campus
June 28, 2024
Crusoe & Lowercarbon to Host AI Hackathon to Accelerate Clean Energy Development, OpenAI Offering Credits and Mentoring Hackers, Department of Energy to Speak at Public Workshop