WD to showcase AI storage platforms at Computex 2026

WD will showcase its AI storage products and platforms at Computex 2026, with a focus on the role of data storage in AI infrastructure.

At the centre of the presentation is a keynote from Ahmed Shihab, Chief Product Officer, who is expected to argue that AI systems should be designed around persistent data growth, not compute alone. WD will also use the exhibition to advance its view that training, inference and other AI workloads generate data that compounds over time, increasing pressure on infrastructure economics as well as performance.

The display will include WD's Ultrastar HDD range, including products based on UltraSMR, ePMR and HAMR. It will also feature High Bandwidth Drive Technology and Dual Pivot Drive Technology, which WD said are projected to increase HDD throughput fourfold while maintaining the relative I/O per terabyte rate users currently expect.

That claim is central to WD's argument that larger AI data sets will require more efficient storage scaling. Higher hard-drive throughput, it said, could reduce the need for additional SSD deployment or broader changes to service architecture as data volumes rise.

Platform push

Alongside its drive products, WD plans to highlight several platform offerings aimed at large-scale data environments. These include Ultrastar Data Series JBODs, OpenFlex EBOF and the RapidFlex NVMe-oF controller, which WD positions for cloud, neocloud and high-performance computing users managing growing AI workloads.

Another product in focus is the new Ultrastar Data 3000 series JBOD, which WD said is designed to provide capacity for AI-related data growth and includes ArcticFlow multizone cooling and IsoVibe vibration isolation.

WD said those design features reduce drive return rates by up to 62% compared with a previous generation without them. It also said the system provides 12 ports of 24 Gb/s SAS-4 host connectivity, intended to address higher bandwidth demands as storage density and workload intensity increase.

Tiered storage architectures will also form part of the message. WD said it will show systems built on Ceph, as well as designs developed with IBM Storage Scale and XTAO, to match cost and performance to different stages of the AI data lifecycle.

That reflects a broader debate in the infrastructure market over how to contain the cost of AI deployments. Much of the early commercial discussion around AI infrastructure centred on graphics processors and networking, but storage suppliers increasingly argue that the persistence and growth of data are becoming limiting factors in system design.

John Chen, Vice President, TRENDFOCUS, linked WD's product direction to those wider market pressures. "The shift toward AI-driven workloads is placing new demands on storage systems, requiring higher throughput, greater capacity, and improved reliability across multiple tiers," Chen said. "WD's advancements in HDD and platform technologies reflect the kinds of innovations needed to support increasingly data-intensive AI environments at scale."

Data growth

WD's approach frames AI infrastructure as a data management problem as much as a compute one. That position supports continued demand for high-capacity hard disk drives, even as solid-state storage remains critical for lower-latency tasks and hotter data tiers.

For storage suppliers, the issue is not only how fast data can be written and retrieved, but how economically very large volumes can be retained over long periods. AI training sets, model outputs, logs, video, sensor information and archived data all contribute to a storage footprint that can expand long after initial model development.

By emphasising hard drives, JBOD systems and tiered storage designs, WD is making the case that AI spending will not be captured solely by chipmakers and server vendors. The strategy also reflects a push to show that established storage technologies remain relevant to AI infrastructure if they can meet throughput and reliability demands at a lower cost per terabyte.

In addition to AI-focused products, WD will display its WD Gold, WD Red and WD Purple lines, along with the G-DRIVE family of external storage products aimed at content creators. Those brands sit outside the company's data centre pitch, but their inclusion underlines the breadth of WD's storage portfolio as it seeks to connect AI infrastructure spending with broader data storage demand.

Stefan Mandl, Vice President of Worldwide Sales and Marketing, WD, said the company sees the market shifting quickly as organisations confront the consequences of rapid data expansion. "Computex is a great opportunity to connect with AI leaders and builders from around the world at a pivotal moment for the industry," Mandl said. "AI is driving data growth at a scale and pace the industry has never seen before, and the organisations that recognise AI as a data system are the ones that will build AI that scales economically and efficiently over the long term. WD provides the durable, scalable and economically efficient data infrastructure foundation required to support AI systems over the long term."