Boost AI training capacity and price performance with one click on the Metacloud

Habana All in on efficiency

Habana partners with

the industry’s most flexible end-to-end machine learning operating system, to enable developers to easily deploy and manage Habana Gaudi AI training workloads, so they can reduce the cost of training models, and innovate faster by managing and optimizing training capacity.

Developers can leverage the efficiency of Gaudi processors on the platform to build frictionless machine learning pipelines in just a few clicks.

End-to-end development: Developers can instantly build production-ready pipelines with
custom or pre-built ML components.

Portable: cnvrg’s MLOPs and container-based infrastructure simplifies engineering of heavy
and complex tasks.

Optimization: Enables developers to orchestrate any ML task to run on any AI stack to maximize
performance and increase efficiency.


Optimize AI workflows with Gaudi’s price performance and Metacloud’s flexibility.
With’s Metacloud AI developers have the flexibility to run end-to-end AI flows on the
mix of cloud or on-premises compute resources or storage they choose.

On-premises operators have the benefit of dynamically expanding their AI processing capability on
an as-needed basis with Metacloud’s cloud-bursting onto Amazon EC2 DL1 instances powered by
Gaudi accelerators. Amazon instances based on Habana Gaudi accelerators deliver up to 40% better
price performance than existing GPU-based EC2 instances, enabling customers to train more, while
spending less. With the combination of Gaudi on Metacloud, AI workloads are:

Flexible and seamless – enabling developers to choose the mix infrastructure and hardware desired
to optimize capacity or cost—even within the same AI/ML workflow or pipeline

Optimized – developers can choose best of breed compute and cloud solutions for each

Easily deployed and managed – Metacloud provides a developer-friendly portal to set-up and launch
AI/ML workflows using the developer’s choice of hardware

Dynamic and fast – developers can manage data and develop, train and deploy models on a mix of
infrastructure instantly

For more information on how developers can leverage the price/performance of Gaudi with the ease, flexibility and expandability of Metacloud, read our blog >

For a deep-dive, how to implement Habana Gaudi AI training on the platform, watch here >

To get more information on how to easily and cost-efficiently boost AI capacity and performance, please contact Habana.