Delivers GPU-native direct memory access on optimized infrastructure powered by AMD EPYC Turin processors and NVIDIA ConnectX-7 networking and previews intelligent context-aware data placement coming later this year.

VDURA today announced three major advances at NVIDIA GTC 2026: availability of Remote Direct Memory Access (RDMA) capability, the upcoming first phase of its Context-Aware Tiering technology planned for later this year, and optimized infrastructure configurations for the VDURA Data Platform built on AMD EPYC Turin processors and NVIDIA ConnectX-7 high-speed networking adapters. Together, these advances deliver on what AI operators need most: more performance from their GPU infrastructure and the storage efficiency to scale without runaway costs.

RDMA: GPU-Native Data Access, Now Available

VDURA's RDMA capability delivers GPU-to-storage data transfers that bypass the CPU entirely, enabling direct memory access between GPU server nodes and the VDURA Data Platform. All GPU server data transfers now leverage RDMA over the network tier, eliminating traditional CPU bottlenecks and delivering the low-latency, high-throughput data paths that modern AI training and inference workloads demand. With RDMA, VDURA DirectFlow enables AI clusters to sustain peak throughput without CPU involvement, freeing compute resources for model execution and reducing end-to-end latency across the data pipeline.

Context-Aware Tiering: Intelligent Data Placement Across the Storage Hierarchy

Building on RDMA, the first phase of VDURA Context-Aware Tiering extends the intelligence of the VDURA Data Platform to dynamically manage data placement across multiple storage tiers based on workload characteristics and access patterns. Planned for general availability later this year, the initial phase of Context-Aware Tiering introduces:

• Extended DirectFlow Buffer to Local SSD: Extends the DirectFlow buffer layer to local NVMe SSD, reducing dependency on network storage for hot data and minimizing latency for active AI workloads.
• KVCache Writeback for Persistence SLA: Intelligent writeback of KVCache data ensures only persistence-critical data is written back to durable storage, minimizing unnecessary I/O while maintaining SLA compliance for AI inference pipelines.
• Context Cache Tiering: A unified Context Cache Tiering framework enables seamless, high-speed read and write access across local SSD and DRAM tiers at LMCache speed, supporting AI inference use cases including long-context language model serving and retrieval-augmented generation.

These features represent the initial release of VDURA Context-Aware Tiering. VDURA has a robust roadmap of additional Context-Aware Tiering capabilities planned through 2027, encompassing deeper application-directed data placement, expanded cross-node cache coherence, and broader hardware support for NVIDIA BlueField-4 DPUs as AI data storage infrastructure continues to evolve.

Full-Stack AI Storage Performance

The combination of RDMA and Context-Aware Tiering positions VDURA as the storage platform of choice for production AI environments. RDMA delivers GPU-native data access that eliminates CPU bottlenecks, while Context-Aware Tiering ensures data is automatically placed in the optimal tier — from memory to long-term retention — based on workload demands. Together, they give organizations the performance, efficiency, and operational reliability to scale AI infrastructure without compromise.

"Today's announcements at GTC 2026 reflect our commitment to delivering the AI storage platform that spans the full data hierarchy — from memory to long-term retention — with no compromises on performance,” said Ken Claffey, CEO of VDURA. “RDMA gives AI teams direct, zero-CPU-overhead access to their data. Context-Aware Tiering brings intelligence to every tier of the extended storage hierarchy, so data is always in the right place at the right time. Together, these capabilities enable organizations to run larger models, serve more inference requests, and efficiently scale AI infrastructure with the operational reliability that production AI demands.”

Availability

RDMA capability is available now for all V5000 and V7000-class systems running the VDURA Data Platform. Context-Aware Tiering Phase 1 is planned for general availability later this year. Customers interested in early access to Context-Aware Tiering are encouraged to visit VDURA’s booth at GTC or visit vdura.com to reach our sales team.

About VDURA

VDURA builds the world's most powerful data platform for AI and high-performance computing, bringing hyperscale-class storage to the rest of the world, powered by HYDRA, the only high-performance distributed architecture purpose-built to unify memory, flash and disk in a single software-defined platform that keeps GPU clusters saturated while delivering hyperscale-class durability and economics. Visit vdura.com for more information.

View source version on businesswire.com: https://www.businesswire.com/news/home/20260316003962/en/

"Today's announcements at GTC 2026 reflect our commitment to delivering the AI storage platform that spans the full data hierarchy — from memory to long-term retention — with no compromises on performance,’’ said Ken Claffey, CEO of VDURA.