SAN JOSE, Calif., March 10, 2026 /PRNewswire-PRWeb/ — VDURA, a modern data storage infrastructure software company for AI, GPU, and high-performance workloads, today announced a strategic collaboration to certify WD’s Ultrastar® Data60 and Ultrastar Data102 Hybrid Storage Platforms within the VDURA Data Platform architecture. The announcement brings hyperscale storage economics to enterprise AI infrastructure by combining the scalable capacity and economics of WD’s Ultrastar Hybrid Storage Platforms with VDURA’s high-performance parallel file system.

As organizations scale AI training and inference workloads, the costs of data storage and movement have become a critical bottleneck. Flash storage alone cannot economically address the petabyte-scale capacity requirements of modern AI pipelines. The VDURA and WD solution will address this challenge through a mixed fleet, hybrid-tiered architecture that pairs the throughput of VDURA’s all-NVMe™ V5000 flash tier with the density and economics of WD’s Ultrastar Hybrid Storage Platforms capacity tier, all managed within a single, unified namespace.

Designed for modern AI and HPC environments, WD’s Ultrastar Hybrid Storage Platforms deliver the dense, cost-efficient capacity required to store and manage the massive data pipelines that power AI training and inference. WD’s Ultrastar Platforms feature patented IsoVibe™ vibration isolation and ArcticFlow™ thermal zone cooling technologies and is available in two configurations to match diverse capacity and density requirements. The Ultrastar Data60 Hybrid Storage Platform includes 60 drive bays in a 4U form factor supporting up to 1.92 PB of raw capacity when populated with 32TB HDDs. The Ultrastar Data102 Hybrid Storage Platform scales to 102 drive bays in the same 4U footprint, delivering up to 3.26 PB of raw capacity using 32TB HDDs. Both platforms include 12x12Gb/s SAS-3 host ports and feature redundant, hot-swappable power supplies, fans and I/O modules for high availability and continuous operation. Each platform also supports up to 24 hybrid SSD slots for mixed-media configurations that balance cost and performance within a single enclosure and can be daisy-chained to scale capacity further. These platforms and future generations will benefit from WD’s recently announced roadmap of 40TB and beyond.

“AI infrastructure requires storage architectures that balance performance with economics at massive scale,” said Scott Hamilton, senior director of product management, marketing and customer experience at WD. “By qualifying our Ultrastar Data60 and Data102 Hybrid Storage Platforms within the VDURA architecture, we’re enabling customers to build AI data pipelines that combine the density and cost-efficiency of HDD with the intelligent data management needed to keep GPUs fully utilized.”

The VDURA Data Platform is a fully software-defined, microservices-based parallel file system that separates the control plane from the data plane. Its DirectFlow client provides POSIX-compliant, cache-coherent parallel I/O directly from compute servers to storage nodes, eliminating the bottlenecks of traditional NFS architectures. With intelligent tiering, data is automatically placed across flash and HDD tiers based on access patterns, ensuring GPU compute resources remain fully utilized while bulk data resides on cost-effective capacity storage.

Storage Economics & Flash Volatility Index

To help organizations evaluate the economics of mixed fleet architectures, VDURA recently introduced the Storage Economics & Flash Volatility Index, a free interactive tool that tracks quarterly flash pricing, models cost scenarios across storage architectures and visualizes performance-capacity-cost trade-offs. The calculator enables organizations to compare all-flash deployments with mixed fleet configurations at varying SSD percentages using real market data, helping infrastructure teams design cost-optimized AI storage environments.

The interactive calculator is available at:

https://www.vdura.com/flash-volatility-index-and-storage-economics-optimizer-tool/