VDURA Summer 2025 Update
Summer has been anything but slow for VDURA. We’ve continued pushing the boundaries of what performance, scalability, and efficiency mean for AI and HPC infrastructure.
Thor Sets the Bar (and Breaks a Record)
At the 2025 World Deadlift Championships, our longtime collaborator Hafþór Björnsson, also known as The Mountain, set a new world record with a 505kg deadlift, wearing VDURA. A powerful reminder that performance is more than speed; it’s consistency under pressure.
On the Road: Summer Events Recap
It’s not just our infrastructure that scales. Our team has been actively engaging with the AI, HPC, and research communities across the globe this summer:
- PEARC 2025 (Columbus, OH): We joined thought leaders in high-performance computing to discuss data bottlenecks in AI and how VDURA addresses storage complexity in scientific research workflows.
- ISC 2025 (Hamburg, Germany): Our return to ISC featured the launch of the AI Data Challenge, a packed happy hour with partners Starfish Storage and Cornelis Networks, and a show-stopping appearance by Hafþór Björnsson.
Product Update: All-Flash VDURA V5000
In June, we launched the VDURA V5000 All-NVMe Flash System, purpose-built for today’s high-performance AI and data-intensive workloads.
- Effective capacity starting at 286TB
- 7.2M IOPS and 210 GB/s throughput
- Powered by VeLO™, our flash-optimized metadata engine
- Ultra-low latency, intelligent automation, and minimal overhead
- Starting at $699K, including hardware, software, and support
The V5000 isn’t a one-size-fits-all box. It’s a modular platform that scales with you. Whether you’re starting with a pilot deployment or expanding to multi-petabyte GPU clusters, VDURA grows as your workloads evolve. We offer flexible system builds optimized for training, inference, mixed I/O, and value-tier capacity, so you can align infrastructure directly to your performance and budget goals.
Looking Ahead to SC25
We’re gearing up for Supercomputing 2025 with several exciting activations and technical showcases on the horizon. If you’re attending, keep an eye out, VDURA will be there in force.
Whether it’s transforming federal-scale AI clusters, supporting cutting-edge research, or redefining what’s possible in the data center, VDURA is helping organizations move faster, scale smarter, and perform with purpose.
Powering Research at Goethe University
Goethe University has selected VDURA to support advanced computational research initiatives. With a fully parallel, high-throughput architecture, the deployment delivers scalable performance that meets the demands of modern AI pipelines, without adding operational complexity.
Chosen for One of the World’s Largest Defense GPU Clusters
A leading U.S. federal systems integrator has selected VDURA to power a multi-year, mission-critical AI and CFD program.
- Phase 1 (2025): 20PB total capacity, 100 GB/s sustained performance
- Phase 2 (2026): Scaling to ~200PB useable, 2.5 TB/s sustained
- Key success factors: sub-millisecond latency, automated tiering across NVMe and HDD, 50% lower cost per TB than flash-only, and 44% greater energy efficiency
Learn more at vdura.com