Pure Storage gets new Cisco FlashStack for AI certification added to Nvidia BasePOD validation

Pure Storage says it has new CiscoFlashStack configs complementing its Nvidia’s BasePOD reference architecture certification, burnishing its AI storage street cred.

Cisco’s FlashStack for AI is a set of Cisco validated designs (CVDs) for AI systems using Cisco UCS server and networking components along with Nvidia GPUs in the servers and Pure FlashBlade storage arrays.

Nvidia’s DGX BasePOD is a reference architecture used to build A systems with Nvidia A100 and H100 GPUs, ConnectX network interface cards,  QM and SN switches, and third-party storage. BasePOD is smaller than Nvidia’s SuperPOD, and Pure’s AIRI AI system is built on the BasePOD reference architecture and uses the latest FlashBlade//S array.

A statement from Tony Paikeday, Senior Director, DGX platform at Nvidia, said:“Businesses across industries are leveraging generative AI to enrich customer experiences and transform operational efficiency. With NVIDIA DGX BasePOD certification, Pure Storage can help customers simplify and scale their initiatives, speeding their return on investment with AI-powered insights.”

Integration layers for DGX BasePOD

DGX BasePOD system Architecture with A100s

There is a rush by businesses to integrate AI across their organizations to drive real-time decision making, operational efficiency, and enhanced scalability. They will struggle, Pure says, unless they have the right storage infrastructure in place. Its storage infrastructure can support, it says, both enterprise development of large-scale AI training environments and the deployment of large language models for AI inference.

Pure says it supports over 100 customers across many AI use cases, including self-driving cars, financial services, genomics, gaming, manufacturing, and more. It wants us to know that it long anticipated the rising demand for AI, delivering an efficient, high performance, container-ready data storage platform to fuel the most advanced enterprise AI initiatives. 

We should not, Blocks & Files imagines, forget this when reading about VAST Data and CoreWeave and Lambda Labs, IBM’s AI-focused Storage Scale 6000, or DDN and its AI experiences.

Pure commissioned a report from Wakefield Research, “Drivers of Change: Meeting the Energy and Data Challenges of AI Adoption,” that says AI-using organizations are using more electrical energy for AI than anticipated and this could set back their ESG goals. Pure, of course, can help. Read the report here.