WD SVP Tim Rausch Details AI Storage Demand and HDD Roadmap at AI Infra Summit 2026
In a keynote at the AI Infra Summit 2026, Western Digital's SVP of Product Engineering Tim Rausch argues that AI is not just a compute problem but a data problem. Citing IDC projections of 274 zettabytes of data created in 2026 alone, he demonstrates how AI-generated video significantly amplifies backend storage needs. The presentation also outlines WD's HDD roadmap, including HAMR technology for higher capacities and High Bandwidth Drives for improved throughput.
ImportanceLocalEvidenceE2 unreplicatedWrite-upQuick
In a video published on September 22, 2026, Tim Rausch, Senior Vice President of Product Engineering at Western Digital (WD), delivered a keynote at the AI Infra Summit 2026. This resource offers a specific vendor perspective for readers tracking physical infrastructure bottlenecks in AI.
According to the official episode description, Rausch's central argument is that AI isn't just a compute problem—it's a data problem. He cites IDC data projecting that 274 zettabytes of data will be created in 2026 alone, growing to 718 zettabytes by 2030. To illustrate this trend, he presents a live demonstration showing that a single 8-second AI-generated video stores 130 times more data on the backend than what a user actually downloads.
Furthermore, Rausch covers WD's HDD roadmap built for the AI era. This includes scaling capacity from today's 30–40TB drives to 60TB by 2028 and 100TB by 2029–2030 using HAMR technology, as well as shipping High Bandwidth Drives with 2X throughput and a path to 8X concurrent read/write heads by the early 2030s. These technical details aim to address the dual demands of capacity and bandwidth for AI workloads.