The HDD Shortage Persists Because Nobody Is Curing It

The consensus on Seagate this year is straightforward: AI data centers are consuming every hard drive the industry can make, the biggest maker is sold out of nearline capacity, and the shortage is a demand problem that keeps compounding. Western Digital confirmed its entire 2026 capacity is effectively sold out, with firm orders stretching into 2028. Seagate reports every nearline exabyte it can produce is pre-sold through 2028. The stock has responded accordingly, roughly tripling over the past year.

That framing gets the direction right and the mechanism wrong. The shortage is not primarily a demand story. It is a supply decision. The three surviving hard-disk manufacturers — Seagate, Western Digital, and Toshiba, out of an industry that once held more than 200 names — are choosing not to build their way out of it, and Seagate's own capital spending shows the choice in dollars. That distinction determines whether the current economics are a cyclical windfall or a durable re-pricing of the entire storage industry.

The growth is in capacity per drive, not in drives shipped

The first thing to check is what is actually growing. The HDD market has split into two measures that point in opposite directions. Total industry unit shipments have been declining for years, as consumers and mainstream PCs migrate to solid-state drives. Data capacity shipped has gone the other way: industry exabyte shipments rose roughly 71% year over year in mid-2025, driven by hyperscalers loading high-capacity drives into AI and cold-storage tiers. Seagate shipped 218 exabytes in its fiscal fourth quarter, up 34% from a year earlier, with 89% going to data-center customers.

This is the ASP-versus-unit divergence in its cleanest form. Revenue and margin are growing not because Seagate is shipping more drives but because each drive carries more terabytes at a better price. That mix — fewer, denser, more expensive drives — is precisely the structure that lets a supplier raise price without losing the customer's scale of purchasing. It is the reason gross margin hit a record 52.7% in the July quarter, up about 1,480 basis points year over year and the thirteenth consecutive quarter of gross-margin gains.

The capex line shows why the shortage has legs

The load-bearing evidence that this is supply discipline rather than a passing inventory crunch sits in Seagate's cash-flow statement. In the twelve months through the latest reported period, Seagate generated roughly $3.1 billion of free cash flow and spent only about $570 million on capital expenditures — one dollar of capex for every five dollars of cash flow. A hardware maker whose product is sold out through 2028 and still constrained on output does not behave that way. It behaves that way when the bottleneck is not capacity but technology, and when management has decided that incremental dollars are better spent migrating to denser drives than adding headcount and plants to chase volume.

That is the capital-reallocation pattern in miniature. The industry consolidated to three players and learned the memory lesson: after a demand shock collapsed pricing, the survivors prioritized technology migration and restrained supply over volume chasing. Seagate is running the same play. Management targets mid-20% annual exabyte growth without added unit capacity, funding the ramp by pushing drives from 44-terabyte Mozaic 4-class HAMR to a 50TB Mozaic 5 platform, qualification slated for late 2027 and volume in 2028. HAMR already accounts for roughly 40% of the nearline exabyte run rate.

The consequence is pricing power that is contractually locked in, not ephemeral. Seagate's nearline capacity is allocated through calendar 2027 under build-to-order agreements that fix both configuration and price, and the shortage is now re-pricing the spot market — high-capacity 24TB drives were up as much as 60%, with average increases around 46%, in early 2026. Sold-out supply plus locked pricing plus record margins is why the shortage has legs: nobody in a three-player industry has the incentive to cure it.

What would break the shortage

The strongest challenge to this setup is not a fourth hard-drive maker. It is a different storage medium. The same AI workloads that are emptying hard-drive inventories also push data centers toward all-flash architecture: PCIe SSDs save power and rack space, and flash vendors are scaling hard, with QLC drives on public roadmaps toward 245 terabytes. There is a structural crossover in progress. For deep-cold, price-sensitive archiving, HDD still wins on raw cost per terabyte, and HAMR's denser platters are Seagate's answer for keeping that advantage. But if NAND costs keep falling and data centers keep prioritizing watts and density, some nearline demand that today has nowhere to go but hard drives will migrate. That is the constraint migration that would deflate the shortage rather than let it re-price.

The second caveat is valuation, and it is a big one. After the run, Seagate trades at roughly 59 times trailing earnings and around 43 times EV/EBITDA, against about 33 times for Western Digital — a wide premium for a company still tied to the same upgrade cycle. The market has already bought the idea that HDD pricing is durably resetting. The thesis no longer needs to be right so much as to keep coming in ahead of an aggressive bar.

The honest reading is that the shortage is real, structurally reinforced, and convertible into margin because Seagate and its two rivals are refusing to build supply into it. The excess return from here, though, depends on the thing that breaks it — whether flash economics keep the re-pricing alive, or whether substitution and a fully-priced stock turn a durable re-rating back into a cycle. The key issue is not whether AI demand lasts. It is whether the industry's supply restraint survives the moment when adding capacity starts to look attractive again.