At its Innovation Day 2026 occasion, Western Digital (now rebranded as WD) unveiled a customer-centric storage roadmap to reinvent laborious drives for AI-era infrastructure, spanning 100TB+ HDDs, new performance- and power-optimized architectures, and an clever platform designed to enhance storage economics and time-to-value.
WD confirmed its 40TB UltraSMR ePMR drive is presently in buyer qualification, with quantity manufacturing deliberate for the second half of 2026. HAMR-based drives are additionally transferring via qualification, with ramp manufacturing anticipated in 2027.
WD will prolong ePMR to 60TB by leveraging HAMR improvements with out rising energy consumption, whereas HAMR is projected to scale to 100TB by 2029. As a result of each applied sciences share a typical structure, WD says clients profit from smoother transitions, improved manufacturing effectivity, and predictable capability planning with out compelled platform modifications.
Boosting HDD Efficiency for AI Workloads
To deal with rising efficiency calls for, WD launched two new drive applied sciences:
- Excessive-bandwidth drive know-how, which permits a number of heads to learn and write concurrently, delivers as much as twice the bandwidth of at the moment’s HDDs and provides a longer-term path to a lot bigger beneficial properties. Prospects are already validating this know-how.
- Twin Pivot know-how provides a second independently working actuator inside a typical 3.5-inch drive. In contrast to earlier dual-actuator designs, this strategy preserves capability and avoids software program modifications whereas delivering as much as double the sequential I/O. Functions see one storage machine to handle, considerably enhancing compatibility throughout the board. Tightening disk spacing additionally permits extra platters per drive. HDDs with Twin Pivot are anticipated to be accessible in 2028.
WD positions these improvements as basically resetting HDD efficiency, enabling workloads beforehand thought of flash-only to run on laborious drives at far decrease value. The corporate says the mixed impression of Excessive Bandwidth Drive and Twin Pivot applied sciences can in the end ship as much as 4x sequential I/O, whereas sustaining relative I/O per TB as capacities climb.
Decrease-Energy Drives for Chilly AI Knowledge
AI workloads generate massive volumes of knowledge which are accessed occasionally however should stay accessible rapidly. WD says its upcoming power-optimized HDDs goal this hole between heat and chilly storage.
WD says these power-optimized HDDs will devour roughly 20% much less energy, enabling clients to decrease working prices, construct lower-cost storage tiers, and enhance sustainability, whereas sustaining sub-second entry for chilly AI information that’s too energetic for tape however too costly for conventional capability drives.
Bringing Hyperscale-Type Storage to Extra Prospects
WD additionally introduced an enlargement of its Platforms enterprise, together with an clever software program layer delivered via an open API, focused for launch in 2027. Geared toward organizations managing 200 petabytes or extra, the platform is designed to simplify deployment throughout WD’s HDD and flash merchandise, shorten time-to-production, and cut back qualification danger.
Relatively than requiring architectural modifications, the software program is supposed to sit down on high of current infrastructure, serving to mid-scale clients obtain storage effectivity nearer to hyperscale environments.
AI and Cloud Centered Enterprise
WD famous that roughly 90% of its income now comes from AI and cloud clients. The corporate says operational enhancements greater than doubled gross revenue yr over yr, serving to drive its inclusion within the Nasdaq 100 and a rating among the many high S&P 500 performers in 2025. WD additionally shared a brand new monetary mannequin outlining expectations for the subsequent three to 5 years.
Market Affect
WD’s Innovation Day paints an image of an organization intent on retaining HDDs related properly into the AI period. However the success of that technique in the end hinges on how these applied sciences translate into usable efficiency, economics, and scale. StorageReview President Brian Beeler weighs in:
“Knowledge development isn’t slowing, however information availability is more and more changing into the larger problem as organizations work to extract worth from their AI investments. Entry to extra information, at scale and on the proper value, is rapidly changing into an attention-grabbing AI workflow.
“HDDs face an uphill battle, although, as QLC SSDs proceed to push into larger capacities and broader deployment, providing compelling density and latency benefits regardless of their larger value. That stated, WD’s roadmap for HDD capability enlargement and performance-focused architectures is essentially the most bold effort we’ve seen to reposition laborious drives for AI-era workloads.
“If WD can ship on its claims of materially larger bandwidth, improved sequential efficiency, and sustained I/O as capacities climb, HDDs may stay a significant a part of future AI information tiers fairly than being relegated purely to deep archive. The performance-oriented applied sciences particularly stand out, and we’d welcome the chance to get hands-on with these designs at scale as soon as WD strikes additional down the qualification and manufacturing curve.”
