A New Era of AI-Powered Infrastructure
NVIDIA has recently introduced the NVIDIA AI Data Platform, a customizable reference design poised to revolutionize AI infrastructure. This groundbreaking platform is being adopted by leading technology providers to construct a new generation of infrastructure specifically engineered for the demanding requirements of AI inference workloads. At its core, this innovative design centers around enterprise storage platforms enhanced with AI query agents, all powered by NVIDIA’s cutting-edge accelerated computing, networking, and software technologies. This platform represents a significant leap forward, enabling businesses to harness the full potential of their data in the age of AI.
Empowering Businesses with Real-Time Insights
The NVIDIA AI Data Platform enables NVIDIA-Certified Storage providers to build infrastructure that significantly accelerates AI reasoning workloads. This is achieved through the integration of specialized AI query agents. These intelligent agents empower businesses to extract valuable insights from their data in near real-time. This capability is fueled by NVIDIA AI Enterprise software, a comprehensive suite that includes NVIDIA NIM™ microservices designed for the new NVIDIA Llama Nemotron models. These models boast advanced reasoning capabilities, further enhancing the platform’s ability to deliver actionable intelligence.
The NVIDIA Llama Nemotron models are a key component of this platform, providing the advanced reasoning capabilities necessary for complex data analysis. These models are designed to handle a wide range of data types and sources, enabling businesses to gain a deeper understanding of their operations and make more informed decisions. The integration of these models with the NVIDIA NIM microservices further enhances the platform’s efficiency and scalability.
Complementing this is the new NVIDIA AI-Q Blueprint, a framework that streamlines the development of agentic systems. The NVIDIA AI-Q Blueprint provides a standardized approach to building AI query agents, making it easier for developers to create and deploy these powerful tools. This blueprint ensures that the agents are optimized for performance and can seamlessly integrate with the NVIDIA AI Data Platform.
Optimizing Infrastructure for AI Query Agents
Storage providers can leverage the NVIDIA AI Data Platform to optimize their infrastructure, ensuring it can effectively power these sophisticated AI query agents. This optimization involves incorporating key NVIDIA technologies, including NVIDIA Blackwell GPUs, NVIDIA BlueField® DPUs, NVIDIA Spectrum-X™ networking, and the NVIDIA Dynamo open-source inference library. These components work synergistically to create a high-performance environment tailored for AI-driven data processing.
NVIDIA Blackwell GPUs provide the computational power necessary for handling the complex calculations involved in AI inference. These GPUs are designed to deliver exceptional performance and efficiency, making them ideal for demanding AI workloads.
NVIDIA BlueField DPUs offload and accelerate networking and storage tasks, freeing up the GPUs to focus on AI processing. This improves overall system performance and reduces latency, ensuring that data can be accessed and processed quickly.
NVIDIA Spectrum-X networking provides high-bandwidth, low-latency connectivity between the different components of the platform. This ensures that data can be transferred quickly and efficiently, minimizing bottlenecks and maximizing performance.
The NVIDIA Dynamo open-source inference library provides a set of optimized tools and libraries for deploying and managing AI models. This library simplifies the process of integrating AI into existing infrastructure and makes it easier to scale AI deployments.
A Collaborative Effort Across the Industry
A remarkable collaboration is underway, with leading data platform and storage providers joining forces with NVIDIA. Industry giants such as DDN, Dell Technologies, Hewlett Packard Enterprise, Hitachi Vantara, IBM, NetApp, Nutanix, Pure Storage, VAST Data, and WEKA are actively working to create customized AI data platforms. These platforms are designed to harness the vast potential of enterprise data, enabling businesses to reason and respond to complex queries with unprecedented speed and accuracy. This collaborative effort demonstrates the widespread recognition of the importance of AI in the modern business landscape and the commitment of these companies to providing their customers with the best possible solutions.
Jensen Huang’s Vision: Data as the Fuel for the AI Age
Jensen Huang, the founder and CEO of NVIDIA, aptly describes data as “the raw material powering industries in the age of AI.” He emphasizes that this collaboration with the world’s storage leaders is forging a “new class of enterprise infrastructure.” This next-generation infrastructure is essential for companies to deploy and scale agentic AI across their hybrid data centers, unlocking new levels of efficiency and innovation. Huang’s vision highlights the transformative potential of AI and the critical role that data plays in realizing this potential. The NVIDIA AI Data Platform is a key enabler of this vision, providing the foundation for a new era of AI-powered businesses.
Bringing Accelerated Computing to Enterprise Storage
The NVIDIA AI Data Platform marks a significant advancement by bringing the power of accelerated computing and AI to the vast number of businesses that rely on enterprise storage. These businesses utilize storage solutions to manage the critical data that drives their operations. The integration of NVIDIA’s technologies into this ecosystem opens up a new realm of possibilities. By leveraging the power of accelerated computing, businesses can now process and analyze their data faster and more efficiently than ever before. This enables them to gain deeper insights, make better decisions, and ultimately achieve greater success.
The Power of NVIDIA Technologies: Blackwell, BlueField, and Spectrum-X
The platform leverages the combined capabilities of NVIDIA Blackwell GPUs, BlueField DPUs, and Spectrum-X networking. These components form an accelerated engine that dramatically speeds up AI query agent access to data residing on enterprise storage systems. This synergistic combination of technologies creates a powerful and efficient platform for AI-driven data processing.
BlueField DPUs play a crucial role in enhancing performance and efficiency. They deliver up to 1.6 times higher performance compared to CPU-based storage solutions. Simultaneously, they reduce power consumption by up to 50%. This translates to a remarkable improvement in performance per watt, exceeding traditional solutions by more than threefold. These improvements are achieved through the offloading of networking and storage tasks from the CPU, allowing the CPU to focus on other critical tasks.
Spectrum-X takes on the task of accelerating AI storage traffic. It achieves up to a 48% increase in speed compared to traditional Ethernet. This is accomplished through the implementation of adaptive routing and congestion control mechanisms, optimizing data flow and minimizing bottlenecks. Adaptive routing allows the network to dynamically adjust to changing traffic patterns, ensuring that data is always routed along the most efficient path. Congestion control mechanisms prevent the network from becoming overloaded, ensuring that data can be transferred reliably and with minimal delay.
The NVIDIA AI-Q Blueprint: A Framework for Agentic Systems
The NVIDIA AI Data Platform’s storage infrastructure utilizes the NVIDIA AI-Q Blueprint. This blueprint serves as a guide for developing agentic systems capable of reasoning and connecting to enterprise data. AI-Q leverages NVIDIA NeMo Retriever™ microservices, which play a pivotal role in accelerating data extraction and retrieval. These microservices can boost data access speeds by up to 15 times on NVIDIA GPUs, significantly enhancing the overall efficiency of the platform. The NVIDIA AI-Q Blueprint provides a standardized and optimized approach to building agentic systems, ensuring that they are both powerful and efficient.
NVIDIA NeMo Retriever microservices are designed to quickly and efficiently retrieve data from a variety of sources. These microservices are optimized for use with NVIDIA GPUs, allowing them to take full advantage of the parallel processing capabilities of these devices. This results in significantly faster data access speeds, enabling AI query agents to respond to queries more quickly and accurately.
Enhancing Responses with Contextual Awareness
AI query agents built using the AI-Q Blueprint are designed to connect to data during the inference process. This crucial connection enables them to provide responses that are more accurate and contextually aware. These agents possess the ability to quickly access large-scale data and process various data types. This includes structured, semi-structured, and unstructured data originating from multiple sources, such as text documents, PDFs, images, and even video content. This ability to process a wide range of data types and sources allows the AI query agents to provide more comprehensive and insightful responses.
The contextual awareness of the AI query agents is a key differentiator. By connecting to data during the inference process, the agents can take into account the specific context of the query and provide responses that are tailored to the user’s needs. This results in a more personalized and relevant experience for the user.
Partner Initiatives: Customizing the AI Data Platform
NVIDIA-Certified Storage partners are actively collaborating with NVIDIA to develop custom AI data platforms tailored to their specific offerings. This collaboration ensures that the NVIDIA AI Data Platform can be seamlessly integrated into a wide range of existing storage solutions. Here’s a glimpse into some of these initiatives:
DDN: DDN is integrating AI Data Platform capabilities into its DDN Infinia AI platform, enhancing its performance and functionality. The DDN Infinia AI platform is a high-performance storage solution designed for AI and machine learning workloads. The integration of the NVIDIA AI Data Platform will further enhance its capabilities, making it an even more powerful solution for these demanding applications.
Dell Technologies: Dell is creating AI data platforms for its range of Dell PowerScale and Project Lightning solutions, expanding their capabilities in the AI domain. Dell PowerScale is a scale-out NAS solution that provides high performance and scalability for a variety of workloads. Project Lightning is a next-generation storage platform designed for AI and machine learning. The integration of the NVIDIA AI Data Platform will enable these solutions to better support AI-driven data processing.
Hewlett Packard Enterprise: HPE is incorporating AI Data Platform capabilities into several of its offerings, including HPE Private Cloud for AI, HPE Data Fabric, HPE Alletra Storage MP, and HPE GreenLake for File Storage. These offerings provide a comprehensive suite of solutions for AI and data management. The integration of the NVIDIA AI Data Platform will enhance their capabilities and provide customers with a more powerful and efficient platform for AI-driven innovation.
Hitachi Vantara: Hitachi Vantara is integrating the AI Data Platform into its Hitachi IQ ecosystem. This integration aims to empower customers with innovative storage systems and data offerings that deliver tangible AI-driven outcomes. The Hitachi IQ ecosystem provides a unified platform for data management and analytics. The integration of the NVIDIA AI Data Platform will enable customers to leverage the power of AI to gain deeper insights from their data.
IBM: IBM is incorporating the AI Data Platform as part of its content-aware storage capability. This integration involves IBM Fusion and IBM Storage Scale technology and is designed to accelerate retrieval-augmented generation applications. IBM Fusion is a hybrid cloud data platform that provides a unified view of data across multiple environments. IBM Storage Scale is a high-performance storage solution designed for demanding workloads. The integration of the NVIDIA AI Data Platform will enable these solutions to better support retrieval-augmented generation applications, which require fast and efficient access to large datasets.
NetApp: NetApp is advancing enterprise storage for agentic AI with its NetApp AIPod solution, which is built upon the foundation of the AI Data Platform. The NetApp AIPod solution is a converged infrastructure solution designed for AI and machine learning workloads. The integration of the NVIDIA AI Data Platform will enhance its performance and scalability, making it an even more powerful solution for these demanding applications.
Nutanix: The Nutanix Cloud Platform with Nutanix Unified Storage will integrate with the NVIDIA AI Data Platform. This integration will enable inferencing and agentic workflows to be deployed across various environments, including the edge, data centers, and public clouds. The Nutanix Cloud Platform provides a unified platform for managing applications and data across multiple environments. The integration of the NVIDIA AI Data Platform will enable customers to deploy AI workloads seamlessly across their entire infrastructure.
Pure Storage: Pure Storage will deliver AI Data Platform capabilities through its Pure Storage FlashBlade solution, enhancing its performance and AI readiness. Pure Storage FlashBlade is a high-performance all-flash storage solution designed for unstructured data. The integration of the NVIDIA AI Data Platform will enhance its performance and make it an ideal solution for AI and machine learning workloads.
VAST Data: VAST Data is collaborating with the AI Data Platform to curate real-time insights using its VAST InsightEngine, unlocking new possibilities for data analysis. The VAST InsightEngine is a data analytics platform that provides real-time insights from large datasets. The integration of the NVIDIA AI Data Platform will enhance its performance and enable it to handle even larger and more complex datasets.
WEKA: The WEKA Data Platform software is integrating with NVIDIA GPUs, DPUs, and networking technologies. This integration optimizes data access for agentic AI reasoning and insights, providing a high-performance storage foundation that accelerates AI inference and token processing workloads. The WEKA Data Platform software is a high-performance file system designed for AI and machine learning workloads. The integration of NVIDIA technologies will further enhance its performance and make it an even more powerful solution for these demanding applications.
Availability and Future Outlook
NVIDIA-Certified Storageproviders are planning to begin offering solutions created with the NVIDIA AI Data Platform starting this month. This marks a significant step towards widespread adoption of this transformative technology. The future of enterprise infrastructure is undoubtedly being shaped by AI, and the NVIDIA AI Data Platform is at the forefront of this revolution. As more and more businesses adopt AI, the demand for high-performance, AI-optimized infrastructure will continue to grow. The NVIDIA AI Data Platform is well-positioned to meet this demand and help businesses unlock the full potential of their data in the age of AI. The platform’s flexibility, scalability, and performance make it an ideal solution for a wide range of AI workloads, from inference to training. As the technology continues to evolve, we can expect to see even more innovative applications of the NVIDIA AI Data Platform in the years to come.