Analyst(s): Brad Shimmin
Publication Date: April 9, 2025
At its annual user conference, NVIDIA introduced the NVIDIA AI Data Platform, a partner-led reference design centered on optimizing AI infrastructure for AI inference and agentic AI workloads. Several technology partners, including VAST Data, Dell, HPE, Pure Storage, and IBM, made supportive product announcements, highlighting the pivotal role of data in accelerating AI reasoning through efficient data access.
Key Points:
- NVIDIA has launched a new reference platform, NVIDIA AI Data Platform, designed to put the company’s AI acceleration hardware to good use in minimizing data bottlenecks and maximizing AI inference token throughput, particularly for agentic systems.
- Efficient data handling is paramount for advanced AI, driving the need for integrated solutions capable of accelerating data access, processing, and retrieval techniques, such as retrieval-augmented generation (RAG) techniques.
- Building a large and diverse partner ecosystem across infrastructure and software players is critical to NVIDIA’s long-term success in promoting the idea of AI factories where model input and output tokens serve as the principal capital under management.
Overview:
Held in late March, NVIDIA’s GPU Technology Conference (GTC) 2025 signaled a clear evolution in the broader AI landscape, one that emphasizes operational efficiency and the foundational role of data infrastructure, especially for building sophisticated agentic AI solutions. NVIDIA drove this theme with the introduction of the NVIDIA AI Data Platform. This initiative provides a customizable reference architecture, rather than a single product, enabling NVIDIA and its partners to construct AI infrastructure that is finely tuned to the intense data demands of AI inference and agentic AI processes. Primary research from Futurum corroborates NVIDIA’s viewpoint on this shift toward optimization, noting that more than one-third of current enterprise GenAI practitioners are beginning to optimize their approach to generative AI (see Figure 1).
Figure 1: Decision Maker: Current Gen AI Usage, 2024
The NVIDIA conference underscored that contemporary AI success increasingly depends on creating efficient “AI factories,” with NVIDIA Founder and CEO Jensen Huang stating that in the future, the value of AI will be constrained by the amount of energy available to its consumers. In other words, AI performance is measured by the speed and cost-effectiveness involved in processing AI tokens, the fundamental units consumed and produced by large language models (LLMs).
At GTC 2025, NVIDIA announced NVIDIA AI Data Platform, directly targeting this opportunity by integrating high-performance elements such as NVIDIA Blackwell GPUs, BlueField DPUs, Spectrum-X networking, and essential software such as the Dynamo inference library and NIM microservices. The primary objective is to optimize the entire data lifecycle, spanning data ingestion, processing, and retrieval. In doing so, companies can reduce the performance bottlenecks and cost constraints increasingly common in complex AI systems, particularly agentic applications that heavily rely on exchanging contextual information through methods such as Retrieval-Augmented Generation (RAG).
This platform’s realization fundamentally depends on collaboration within and support among NVIDIA’s extensive partner ecosystem, with storage vendors playing a crucial role in providing the enterprise-grade data foundation. Numerous partners unveiled supporting solutions at GTC 2025, all geared toward positioning data more effectively for AI inferencing and token processing workloads to enhance speed and efficiency.
Dell Technologies, for instance, launched the Dell AI Factory with NVIDIA. This end-to-end integrated system, based on Dell PowerEdge servers and Dell Pro Max storage, utilizes the AI Data Platform architecture to deliver optimized access to high-quality data specifically for agentic AI. Similarly, Hewlett Packard Enterprise (HPE) demonstrated its commitment by announcing comprehensive support for the NVIDIA AI Data Platform across its major portfolios, including HPE Private Cloud AI, Alletra Storage, and HPE GreenLake solutions. HPE’s integration leverages NVIDIA technologies such as AI-Q Blueprints and NIM microservices to offer infrastructure optimized for demanding AI data workflows.
The full report is available via subscription to the Data Management & Analytics IQ service from Futurum Intelligence—click here for inquiry and access.
For more information, please see the full press release on NVIDIA’s AI Data Platform: NVIDIA and Storage Industry Leaders Unveil New Class of Enterprise Infrastructure for the Age of AI.
Futurum clients can read more in the Data Management & Analytics Platform, and nonclients can learn more here: Data Management & Analytics Practice.
About the Futurum Data Management & Analytics
The Futurum Data Management & Analytics Practice provides actionable, objective insights for market leaders and their teams so they can respond to emerging opportunities and innovate. Public access to our coverage can be seen here. Follow news and updates from the Futurum Practice on LinkedIn and X. Visit the Futurum Newsroom for more information and insights.
Author Information
Brad Shimmin is Vice President and Practice Lead, Data and Analytics at Futurum. He provides strategic direction and market analysis to help organizations maximize their investments in data and analytics. Currently, Brad is focused on helping companies establish an AI-first data strategy.
With over 30 years of experience in enterprise IT and emerging technologies, Brad is a distinguished thought leader specializing in data, analytics, artificial intelligence, and enterprise software development. Consulting with Fortune 100 vendors, Brad specializes in industry thought leadership, worldwide market analysis, client development, and strategic advisory services.
Brad earned his Bachelor of Arts from Utah State University, where he graduated Magna Cum Laude. Brad lives in Longmeadow, MA, with his beautiful wife and far too many LEGO sets.
