5 days Ago By Chris Mellor
Pure Storage has announced certification of its FlashBlade//S500 storage with Nvidia’s DGX SuperPOD along with a new GenAI Pod system for inferencing. The FlashBlade//S500 is a high-end, performance-optimized FlashBlade system, with the mid-range S200 and entry-level S100 completing the //S range. The GenAI Pod is a “full-stack solution providing turnkey designs” based on Pure Storage arrays.
SuperPOD is Nvidia’s largest GPU server and there were six : DDN and its A3I AI400X2T Lustre array, Dell with PowerScale using Ethernet storage, IBM with Storage Scale System 6000, NetApp with EF600 running BeeGFS, VAST Data with its Data Platform, and WEKA. Pure Storage’s FlashBlade//S500 with Ethernet is now the seventh SuperPOD storage supplier. Nvidia’s Tony Paikeday, Senior Director of AI Systems, stated: “The Pure Storage GenAI Pod with Nvidia AI Enterprise software and the certification of FlashBlade//S500 with Nvidia DGX SuperPOD can help organizations eliminate infrastructure complexity, speed deployments, and simplify operations.
” Pure’s GenAI Pod includes validated designs that enable turnkey solutions for GenAI use cases, such as drug discovery, trade research and investment analysis, and RAG with agentic frameworks for semantic search, knowledge management, and chatbots. It integrates Pure’s Portworx with its Kubernetes operations, container storage, and backup software facilities. GenAI Pod has services that provide automated deployments of Nvidia’s NeMo and NIM microservices and the Milvus vector database, with, Pure says, a one-click deployment and streamlined Day 2 operations for vector databases and foundation models.
The full-stack aspect means the GenAI Pod includes hardware, software, foundational models, and professional services from industry AI vendors: Arista, Cisco, KX, Meta, Nvidia, Red Hat, Supermicro, and WWT. Pure Pure Fusion for Files, Zero Move Tiering, and Real-time Enterprise File services with always-on multi-protocol, along with auditing, QoS, SMT for file, and an AI Copilot to FlashBlade in September. It also added 150 TB flash drive (Direct Flash Module) support then as well.
Pure’s AI infrastructure storage system, which uses FlashBlade//S storage, has Nvidia BasePOD certification. Up until now, Pure has not had SuperPOD certification, putting it at a disadvantage in bids where data is served to SuperPODs. Wedbush financial analyst Matt Bryson commented: “The GenAI Pod may have more significant longer-term ramifications for Pure Storage if we see increased enterprise adoption of AI hardware (perhaps as inference applications take off) and Pure’s solution enables easier/better enterprise AI implementations using Pure hardware.
However, in the near term, we believe finally achieving SuperPOD certification might be the more important accomplishment, given Pure can finally point to checking that box during competitive bake-offs.” Pure says the new GenAI validated designs, along with FlashBlade//S500 SuperPOD certification, expand upon its AI system portfolio, including with Nvidia DGX BasePOD, validated Nvidia OVX servers, and FlashStack for AI with Cisco. The Pure Storage GenAI Pod is expected to be generally available in the first half of 2025.
.
