Data orchestrator and GPU data feeder Hammerspace is providing a data foundation for AI agents by adopting Nvidia’s AI Data Platform reference design.
Hammerspace is effectively bolting this, its Data Platform for AI Anywhere, on its Global Data Environment (GDE) offering which virtualizes a customer’s file and object data estate across multiple vendor’s storage silos. It has a single global namespace which helps enables activity level-based data placement decisions, including tier zero, the locally attached NVMe flash drives on GPU servers, and fast data supply to GPU servers with GPU Direct. The company says it has automated data objectives, integrated Milvus vector database, MCP support, and tight integration with AI agents. It software ensures, it says, that training and inference workloads always have immediate access to the data they need, without manual data movement or complex integration layers.
Jeff Echols, VP for Strategic Partnerships at Hammerspace, said: “The Hammerspace Data Platform eliminates the chaos of legacy enterprise data silos, allowing organizations to instantly make data available to AI agents anywhere, while maintaining full control and governance.”
Hammerspace’s AI Data Platform is based on Nvidia’s AI Enterprise software, and integrated with Nvidia GPUs such as the RTX PRP6000 Blackwell products, and Spectrum-X Ethernet networking. It supports Nvidia’s NI and NeMo Retriever microservices. View it as a Hammerspace reference design for Nvidia’s AI Data Platform, which itself is a reference architecture.
It’s delivered as a validated software system, and only available from strategic Hammerspace channel partners. As we understand it, the Hammerspace channel program does not have specific named tiers like gold and silver. Instead there are ordinary or general channel partners and strategic ones who who go beyond transactional relationships, ie. merely reselling, to co-develop solutions, expand market reach, and prompt things like AI infrastructure development.
Hammerspace says customers can start with its AI Data Platform at small scale, using pilot projects to validate AI initiatives for example. They can then, assuming success, scale linearly to multi-site or global deployments.
Jack Hogan, VP of Strategy and Solutions at SHI Internationa, a strategic partner we understand, said: “We’re excited to showcase this solution in the SHI AI & Cyber Labs, where customers can experience firsthand how it simplifies data orchestration and accelerates AI adoption.”
The Hammerspace reference design for the Nvidia AI Data Platform will be showcased at NVIDIA GTC in Washington, D.C. and will be available through authorized Hammerspace partners in late 2025.
Learn more at Hammerspace’s AI Data Platform website.
