DISCOVER THE FASTEST FULLY PARALLEL AND DISTRIBUTED FILE SYSTEM PERFECT FOR AI

Cirrascale has announced its partnership with Weka.io to deliver a distrbuted file system that's even faster than local NVMe and perfect for handling small files efficiently. Learn more about our partnership today.

TEST DRIVE H2O DRIVERLESS AI ON IBM POWER SYSTEMS

Cirrascale has teamed up with H2O and IBM to deliver H2O Driverless AI on IBM Power Systems to deliver automated machine learning solutions. Qualifying companies can receive a one week trial.

MULTI-GPU, DEDICATED DEEP LEARNING CLOUD SERVERS

Our service provides highly available, dedicated, multi-GPU servers with state-of-the-art configurations. In fact, there's no one in the industry that offers the level of performance, density and high-end technology that we do.

NVIDIA® TESLA® V100 GPU ACCELERATORS ARE NOW AVAILABLE

It's our promise to you to always have the latest technology available for your deep learning development and training. The NVIDIA® Volta® architecture has made its way to our cloud service. Discover the latest GPU architecture developed by the engineering geniuses at NVIDIA.

POWER9 CLOUD SERVICES NOW AVAILABLE

With the largest variety of architecture available in the cloud, you'll want to try our POWER9 cloud services. With eight threads per processor core, POWER9 is able to deliver parallel processing of data queries like never before, resolving queries faster than other processor architectures.

Dedicated Servers

There's no virtualization here. Within our private cloud, you get the entire system, its resources, and all its GPUs (up to 16 of them). Now that's some serious dedication.

Largest Variety of GPUs

Other providers give you two generation old GPUS. Not us, we use the latest NVIDIA┬« Tesla┬« GPU accelerators, powered by NVIDIA Volta™ and Pascal™ architecture.

Perfect for Deep Learning Frameworks

Building the ultimate deep learning application takes different tools along the way. Our variety of configuration offerings are set up to allow you to start small and scale up.

Up to 16 GPUs on a Single Root Hub

All of our x86 dedicated cloud servers utilize PCIe switch riser technology enabling up to 16 GPUs to communicate directly with each other on the same PCI bus.

Customers select us over AWS, Azure, and GCP because we have no extra fees, and "no surprises" billing.

GET STARTED

We've teamed up with some great companies.