Pure Storage, Nvidia partner to democratize AI with new infrastructure solutions

Pure Storage, a pioneer in delivering advanced data storage technology and services, unveiled a suite of new AI infrastructure solutions and validated reference architectures developed in partnership with Nvidia at last week’s GTC conference. The announcement aimed to provide enterprises with proven frameworks to manage the high-performance data and compute requirements necessary for successful AI deployments. 

Miroslav Klivansky, the Global Practice Leader for AI and Analytics at Pure Storage, recently discussed with VentureBeat the company’s strengthened partnership with Nvidia and its potential impact on the adoption of artificial intelligence by enterprises.

The partnership between Pure Storage and Nvidia comes at a critical juncture, as enterprises across various industries are increasingly adopting AI to drive innovation, optimize operations, and gain a competitive advantage; however, the majority of AI deployments are currently scattered across fragmented data environments, which can hinder the full realization of AI’s potential. Pure Storage and Nvidia’s collaboration seeks to address this challenge head-on.

“Pure Storage has had a long-standing partnership with Nvidia,” Klivansky told VentureBeat. “In 2018, in collaboration with Nvidia, we launched AIRI, the first AI-ready infrastructure reference architecture, purpose-built to enable organizations to achieve better utilization and uptime and ultimately scale their AI investments without complexity. Since then, Pure Storage was among the first enterprise data storage vendors to become Nvidia DGX BasePOD certified.”

Simplifying the AI puzzle with validated reference architectures

Klivansky emphasized the importance of their collaboration with Nvidia, stating, “Today’s announcement is a testament to this ongoing, collaborative relationship with Nvidia to support enterprise AI deployments. As a leader in AI, Pure Storage, in collaboration with Nvidia, is arming global customers with proven solutions to address the high-performance data and compute requirements they need to drive successful AI deployments with faster model training and inference, with greater operational efficiency, plus lower total cost and energy requirements.”

One of the key solutions unveiled at GTC was a Retrieval Augmented Generation (RAG) Pipeline for AI Inference, which combines Nvidia GPUs for compute and Pure Storage’s all-flash enterprise storage. “Using Pure Storage’s data storage infrastructure for retrieval-augmented generation, customers can enhance standard large language models (LLMs) with custom corporate data to provide greater relevance, accuracy, and currency to chatbots and avoid the massive investment of building their own custom large language models,” explained Klivansky.

Expanding choice with Nvidia OVX server storage certification

Another significant development announced at GTC was Pure Storage’s achievement of Nvidia OVX Server Storage certification. This certification provides customers with flexible storage reference architectures validated against key benchmarks, ensuring a robust infrastructure foundation for cost and performance-optimized AI hardware and software solutions.

“With the latest Nvidia OVX Server Storage validation, we can now provide enterprise customers and channel partners with flexible storage reference architectures, validated against key benchmarks to provide a strong infrastructure foundation for cost and performance-optimized AI hardware and software solutions,” said Klivansky. “OVX computing platforms are powered by Nvidia L40s GPUs, which means our customers now have their choice of validated reference architectures from Pure Storage with greater compute GPU platform options and availability.”

Klivansky highlighted the unique advantages of Pure Storage’s solution, stating, “What differentiates Pure’s solution from competitive solutions is Pure’s data storage platform, optimized for AI workloads. Pure’s FlashBlade//S provides a combination of leading performance, operational efficiency, energy and cost savings, and a reliable, future-proof storage foundation that grows with changing AI data requirements.”

Vertical-specific AI solutions starting with finance

In addition to the broader AI infrastructure solutions, Pure Storage and Nvidia are also focusing on developing vertical-specific AI applications. The first of these is a Financial RAG using FinGPT, designed specifically for the finance industry.

Klivansky highlighted the prevalence of finance industry-specific AI use cases in large enterprises, stating, “Generative AI using retrieval augmented generation (RAG) and pre-trained LLMs can be used to create financial analyses that are more accurate, real-time/faster, and more cost-efficient than those created by humans since they are able to crunch a lot more information at a much higher rate.”

The development of vertical-specific AI solutions demonstrates Pure Storage and Nvidia’s commitment to addressing the unique needs of different industries. By providing tailored AI applications, the companies aim to accelerate the adoption of AI in sectors such as finance, healthcare, and the public sector.

Expanding the AI partner ecosystem

Pure Storage’s dedication to meeting the evolving data storage needs of AI is further evidenced by its expanding partner ecosystem. The company has forged partnerships with leading AI platform providers like Run.AI and Weights & Biases, as well as channel partners such as WWT, ePlus, CDW, and Insight.

“We’re excited to expand investment into our AI partner ecosystem with partners like Run.AI and Weights & Biases,” said Klivansky. “Run.AI complements Pure’s storage performance and energy efficiency value for AI infrastructure by optimizing the use of GPU-powered compute resources to accelerate development. Weights & Biases provides an easy-to-use framework for model building and RAG pipelines — which augments Pure’s data storage platform for AI, further enriching operational efficiency and accelerating end-to-end ML workflows.”

The collaboration with these partners enables Pure Storage to offer a comprehensive and integrated AI solution stack, simplifying the deployment and management of AI infrastructure for enterprises. By leveraging the expertise of these partners, Pure Storage can deliver a seamless and optimized AI experience, from data storage to model development and deployment.

Enabling sustainable AI with energy-efficient storage

As the adoption of AI continues to grow, the environmental impact of data centers has become a significant concern. Pure Storage recognizes this challenge and has designed its storage platform to help customers reduce their energy consumption and carbon footprint.

“Pure has designed and built its platform to allow customers to dramatically decrease their energy and carbon footprints,” explained Klivansky. “Pure Storage’s flash-optimized systems generally use between 2x and 5x less power than competitive SSD-based systems and between 5x and 10x less power than the hard disk systems we replace.”

By providing energy-efficient storage solutions, Pure Storage enables enterprises to build more sustainable AI infrastructure. This not only helps organizations meet their environmental goals but also results in significant cost savings in terms of power, cooling, and data center space.

Paving the way for the future of enterprise AI

As Klivansky stated, “Pure Storage anticipated the needs and potential of AI at the inception of this latest wave, delivering the high-performance, efficient, and container-ready storage platform the industry needed to capitalize on and truly derive value from the massive amounts of data used to fuel this technology.”

With its proven track record in delivering advanced data storage solutions and its commitment to innovation, Pure Storage is well-positioned to play a pivotal role in the ongoing AI revolution. The company’s dedication to meeting the unique data storage requirements of AI workloads is evident in its continuous efforts to develop cutting-edge solutions and forge strategic partnerships.

“Over the past years, Pure Storage has been at the forefront of innovation to meet the escalating data storage needs of AI deployments,” said Klivansky. “Recognizing the challenge posed by fragmented data environments, Pure Storage pioneered an enterprise data storage platform to meet the unique needs of AI. By offering a simple, reliable, and efficient storage platform, Pure Storage enables enterprises to fully harness the potential of AI while mitigating risks, reducing costs, and minimizing energy consumption.”

The collaboration between Pure Storage and Nvidia is not only a testament to their shared vision for the future of AI but also a reflection of their commitment to enabling enterprises to unlock the full potential of this transformative technology. By providing a comprehensive and validated AI infrastructure solution, the companies are helping organizations across various industries to accelerate their AI adoption journey and gain a competitive edge in an increasingly data-driven world.