/dqc/media/media_files/2025/11/17/western-digital-unveils-ai-hpc-storage-innovations-at-sc25-2025-11-17-10-51-38.png)
Western Digital unveils AI & HPC storage innovations at SC25
Western Digital showcased its next-generation innovations for AI and high-performance computing (HPC) at Supercomputing 2025, unveiling new solutions, partnerships, and ecosystem expansions that redefine how performance, capacity, flexibility, and scalability are delivered across modern data environments.
This year’s showcase reinforces Western Digital’s commitment to helping organisations overcome storage bottlenecks that limit GPU utilisation, slow model development, and increase infrastructure costs, all amid unprecedented growth in AI workloads.
Powering AI Breakthroughs with Advanced NVMe-oF Architectures
At the event, Western Digital demonstrated how its platforms integrate with leading AI infrastructure providers to accelerate discovery.
In collaboration with PEAK: AIO, the company highlighted how modern NVMe-oF architectures eliminate traditional storage bottlenecks, enabling:
Higher throughput for AI training
Independent scaling of compute and storage
Greater GPU utilisation
Faster time-to-insight across demanding AI and HPC workloads
These real-world performance demonstrations showcased how storage innovation directly impacts scientific research, generative AI development, and enterprise-scale AI deployments.
Democratizing UltraSMR for AI and HPC Environments
Western Digital continues to expand access to its UltraSMR technology beyond major hyperscalers.
Through ecosystem collaborations with Leil Storage and Swiss Vault, the company demonstrated how specialised file system optimisations unlock SMR’s:
Superior capacity efficiency
Improved sequential write performance
Lower power consumption per terabyte
The company’s Ultrastar SMR-enabled JBOD platforms, including Data60 and Data102, now offer dramatically higher density.
With the new 32TB UltraSMR HDDs, a single Data102 can deliver up to 3.26PB in one enclosure, making:
Exabyte-scale AI pipelines
Research data lakes
HPC simulation repositories
more cost-effective and energy-efficient for research institutes, universities, labs, and mid-market enterprises.
Expanded Open Composable Compatibility Lab (OCCL) Ecosystem
Western Digital’s OCCL ecosystem continues to grow, now including new participants such as:
ASUS
Leil Storage
Open-E
Solidigm
Swiss Vault
Alongside existing collaborators DapuStor, Phison, and SanDisk, the expanded OCCL provides pre-validated, interoperable solutions that customers can deploy confidently.
This vendor-neutral proving ground helps organisations:
Reduce integration risk
Avoid vendor lock-in
Accelerate deployment
Scale compute and storage independently
Improve supply chain resilience
The updated validation lineup includes new SSDs, control planes, and disaggregated infrastructure building blocks.
A Platform Designed for the Future of AI at Scale
“With Supercomputing 2025, we’re demonstrating Western Digital’s leadership in enabling AI and HPC workloads to scale efficiently across every environment,” said Kurt Chan, Vice President and GM, Western Digital Platforms Business.
“Our platforms serve as a force multiplier, from democratizing SMR for exabyte-scale deployments to expanding the OCCL ecosystem, we’re architecting the foundation for tomorrow’s most ambitious computational workloads.”
Western Digital will showcase its technologies at Supercomputing 2025 from November 16–21, 2025, at the America’s Centre Convention Complex in St. Louis, Missouri, located at Booth #6529. Exhibit hours are scheduled for November 18 and 19 from 10 AM to 6 PM, and November 20 from 10 AM to 3 PM. Visitors will be able to experience firsthand the storage solutions built to support next-generation AI and HPC infrastructure.
Read More:
GPT-5.1, a new chapter in Developer AI with agentic capabilities
From Cloud to AI Agents: Innovacx CEO on the future of enterprise tech
/dqc/media/agency_attachments/3bO5lX4bneNNijz3HbB7.jpg)
Follow Us