Nvidia remains a dominant force in the artificial intelligence (AI) hardware sector despite a recent decline in its stock price. During the GTC 2025 event held in San Jose, California, the company introduced its AI Data Platform while emphasizing concepts such as “agentic AI” and “reasoning.” This initiative is expected to influence the storage supply industry significantly.
Nvidia’s Blackwell Ultra GPU
A key highlight of the event was the unveiling of the Blackwell Ultra graphics processing unit (GPU), engineered specifically for AI datacenter applications. This GPU is tailored for reasoning models like DeepSeek R1, enhancing both memory and inference performance. Alongside this advancement, Nvidia announced a new range of rack-scale products within its GB/NVL line, which will integrate the Blackwell Ultra GPU. This includes updated DGX family SuperPod clusters, workstations, network interface cards (NICs), and laptop GPUs.
Storage Partners and Product Announcements
Recognizing the critical role of data processing in AI, Nvidia’s announcements at GTC 2025 resonated across major storage vendors. Notable storage solutions that align with Nvidia’s AI Data Platform reference architecture were confirmed, allowing third-party suppliers to develop products that meet Nvidia’s specifications for workloads involving agentic and reasoning methodologies. Among the partners listed were DDN, Dell, HPE, Hitachi Vantara, IBM, NetApp, Pure Storage, Vast Data, and Weka.
Key Developments from Storage Suppliers
Storage suppliers made several key announcements during the event:
- DDN: Introduced the Inferno fast object appliance, which integrates Nvidia’s Spectrum-X switch with DDN Infinia storage, currently limited to S3 object storage.
- Dell: Announced powerful 20-petaflop-scale PCs for AI, alongside validation of its PowerScale scale-out file system for enterprise AI factory deployments.
- HPE: Highlighted its new “unified data layer” that accommodates both structured and unstructured data, along with upgrades to its MP B10000 array for unified block and file access.
- Hitachi Vantara: Launched the Hitachi iQ M Series, which integrates its Virtual Storage Platform One with Nvidia AI Enterprise software, supporting agentic AI initiatives.
- IBM: Announced new collaborative efforts with Nvidia, including the launch of a content-aware storage capability for IBM Fusion and expanded Watsonx integrations.
- NetApp: Achieved validation for its AFF A90 product with DGX SuperPOD and received the new Nvidia-Certified Storage designation for its AIPod product.
- Pure Storage: Revealed compatibility with the Nvidia AI Data Platform following its FlashBlade//Exa announcement.
- Vast Data: Launched an enterprise-ready AI Stack that combines Vast’s InsightEngine with Nvidia DGX products and advanced networking solutions.
- Weka: Announced data store certification for Nvidia GB200 deployments, including WEKApod Nitro Data Platform Appliances certified for Nvidia Cloud Partner deployments.
This collaboration among storage suppliers and Nvidia underscores the increasing demand for high-speed data processing capabilities necessary for effective AI training and inference.