The 100TB HDD roadmap is no longer theoretical—it is now the central battleground for storage vendors racing to meet AI and high-performance computing demand. Seagate and Western Digital have outlined fundamentally different paths to reach 100TB capacity within the next five years, each betting on distinct technologies and timelines.
Key Takeaways
- Seagate ships 44TB drives now, plans 100TB HDDs by 2030 using HAMR technology exclusively.
- WD targets 100TB by 2029 with overlapping ePMR and HAMR architectures for smoother scaling.
- WD’s 40TB UltraSMR ePMR drive enters volume production H2 2026; HAMR ramp begins 2027.
- AI and HPC markets are driving zettabyte-scale storage demand, making HDDs cheaper than SSD alternatives.
- WD introduces Dual Pivot Technology (second independent actuator) launching 2028 for 4x sequential I/O without capacity loss.
Seagate’s HAMR-First Strategy: Scaling to 100TB by 2030
Seagate is pursuing a straightforward path: HAMR (Heat-Assisted Magnetic Recording) as the sole technology for scaling beyond 44TB. The company is shipping 44TB Mozaic drives now and plans to reach 100TB within five years, according to Seagate Chief Commercial Officer BS Teh. “There is a plan and a clear road map for the company to go from 40TB to 100TB per drive,” Teh told CNBC. Seagate’s roadmap includes 36TB, 40TB (10-platter), and 50TB variants before the final push to 100TB, all leveraging HAMR’s ability to write data at higher densities using heat.
The company frames Mozaic as foundational for AI workloads. Seagate’s 40TB Mozaic drives replace 42,000 conventional 24TB units with just 25,000 advanced units in data centers, cutting both physical space and power consumption significantly. Teh emphasized the strategic necessity: “This is a key enabler for the industry to be able to deliver the storage capacity that the market needs. There’s no other technology that’s able to produce this capacity of storage technology to meet the growth that the market needs”. This confidence in HAMR’s scaling potential sets Seagate apart as the vendor most committed to a single architectural path.
WD’s Dual-Architecture Approach: ePMR First, HAMR Second
Western Digital is taking a more cautious, overlapping strategy. Instead of jumping straight to HAMR, WD is scaling ePMR (Energy-Assisted Perpendicular Magnetic Recording) first, then transitioning to HAMR as reliability and manufacturing maturity improve. The company unveiled this 100TB HDD roadmap at Innovation Day 2026, revealing a staggered rollout designed to minimize risk and manufacturing disruption.
WD’s 40TB UltraSMR ePMR drive is currently in customer qualification, with volume production targeted for the second half of 2026. The company plans to push ePMR to 60TB by 2028 while simultaneously qualifying HAMR-based drives for production ramp in 2027. HAMR drives will then scale from 40TB and 44TB in 2027 to 100TB by 2029, extending beyond 2030 if needed. WD Chief Product Officer Ahmed Shihab described the approach: “We’re reinventing the hard drive that the cloud and AI depend on so that we can deliver more capacity sooner while doubling both bandwidth and I/O”.
Performance and Power: WD’s Hidden Advantage
Beyond raw capacity, WD is addressing two pain points data centers actually care about: speed and power consumption. WD’s High Bandwidth Drive technology, already in customer hands, delivers sequential I/O performance around 500-510 MB/s—saturating the SATA interface’s 530 MB/s ceiling without requiring new infrastructure. Shihab noted: “The performance is about 500 Megabytes per second and change. So we saturate the SATA link that’s already in existence today”.
More intriguing is Dual Pivot Technology, which WD is moving from lab to production in 2028. Unlike earlier dual-actuator designs, Dual Pivot adds a second independent actuator without sacrificing capacity or requiring software changes—a critical advantage for hyperscalers managing thousands of drives. The technology enables up to 4x sequential I/O improvement, addressing a key bottleneck in large-scale deployments. WD is also shipping power-optimized drives with 20 percent less power draw and 10 percent more capacity, entering customer qualification in 2027.
Why This Matters: AI and the HDD Comeback
The surge toward 100TB HDDs reflects a fundamental shift in data center economics. AI training and inference models generate massive datasets that would be prohibitively expensive to store on SSDs alone. HDDs, despite slower random access, offer dramatically lower cost per terabyte and consume less power at scale. Zettabyte-scale storage—the industry’s current target—would be unaffordable without advanced HDD technology. Seagate and WD’s competing roadmaps are not academic exercises; they are existential bets on which vendor can deliver capacity faster while keeping power and cooling costs manageable.
The architectural divergence matters. Seagate’s single-technology approach simplifies manufacturing and messaging but puts all chips on HAMR’s reliability at extreme densities. WD’s overlapping strategy hedges risk: if ePMR hits unforeseen scaling limits, HAMR is already ramping. If HAMR stumbles, ePMR can carry the load longer. For hyperscalers, this matters—one vendor’s manufacturing crisis could constrain supply for years.
What About Toshiba?
The industry’s third major HDD vendor, Toshiba, has outlined a distinct strategy but specific details remain sparse. While Toshiba is part of the 100TB race, the company has not disclosed concrete timelines or technological approaches matching the transparency of Seagate and WD.
Is the 100TB HDD roadmap realistic?
Both Seagate and WD have shipped drives at or near their stated capacities on schedule in the past. Seagate’s 44TB drives are shipping now; WD’s 40TB ePMR drives are in customer qualification. The five-year window to 100TB is aggressive but not unprecedented for the industry. HAMR technology has been in development for over a decade, and both vendors have invested heavily in manufacturing tooling.
Will 100TB HDDs replace SSDs in data centers?
No. 100TB HDDs will coexist with SSDs, not replace them. HDDs excel at sequential, large-capacity workloads (backups, archives, AI training data). SSDs remain essential for random-access, latency-sensitive tasks (databases, caching, real-time inference). The real win is cost efficiency—fewer 100TB HDDs mean lower power, cooling, and space costs compared to smaller drives, reducing the SSD burden on hyperscale budgets.
The 100TB HDD roadmap represents the storage industry’s answer to AI’s insatiable appetite for data. Seagate is betting speed wins; WD is betting risk mitigation and performance features win. Whichever vendor reaches 100TB first will own a massive competitive advantage, but the real winner is the data center operator who can finally afford to store the internet’s intelligence.
Edited by the All Things Geek team.
Source: Tom's Hardware
