"Weka and Stability AI Aim to Maximize Cloud Advantage for AI"

WEKA and Stability AI recently announced a partnership to create a new AI solution aimed at training multiple AI models in the cloud.

WEKA's data platform integration cloud solution aims to enhance Stability AI's ability to train multiple AI models, including its popular Stable Diffusion model, and extend the benefits of efficiency, cost, and sustainability to its customers. Both companies are committed to maximizing the cloud advantages of AI model training.

Both companies have always combined AI with the background of enterprises and data, and this collaboration will only strengthen their customers' application of AI.

Supporting AI models and improving use cases

In particular, WEKA has revealed many infrastructure challenges that affect AI applications, with a survey showing that 32% of respondents consider data management as a technical factor inhibiting AI and ML deployments.

WEKA has been working hard to develop a unique way to significantly improve the experience of its cloud-first generative AI customers. The integration cloud solution of the WEKA data platform is the first extended storage for running cloud-based instances with deep learning capabilities. It leverages the transient local storage and memory of cloud-based AI instances to reduce costs and provide performance improvements for large-scale generative AI resources compared to traditional data architectures.

Stability AI hopes to make more efficient use of existing cloud resources. The company has partnered with WEKA to develop a "fusion cloud" approach and has recently successfully trialed the solution. This approach aims to provide a set of high-performance resources for both applications and data platforms.

As a result, Stability AI's models can be trained more optimally and are more useful for its research team.

"Running AI model training and inference in the public cloud simply and affordably"

"As the only independent, open, and multimodal generative AI company, we understand the limitations of running fusion workloads on-premises," said Tom Mason, CTO of Stability AI.

"WEKA enables us to offer our customers more choices by enabling fusion mode training in the public cloud. We hope this collaboration will increase our hardware utilization and scale it to thousands of instances, enabling us to accomplish more work more efficiently while being more sustainable."

Once deployed, Stability AI plans to achieve performance improvements and significant cost savings in the cloud from WEKA's data platform. The company expects to leverage WEKA's solution to support its upcoming sustainability plans as it deploys and scales.

"The WEKA data platform can improve GPU stack storage efficiency by 10-50 times, bringing significant cost advantages to generative AI companies and helping them maximize their paid resources," said Liran Zvibel, Co-founder and CEO of WEKA.

"At the same time, WEKA can easily and affordably run AI model training and inference in the public cloud, helping to reduce data infrastructure as well as energy and carbon emissions, providing dual bottom-line benefits to customers looking to leverage AI without compromising their enterprise's sustainability goals."