Western Digital's Sold-Out Hard Drives: The AI 'Demand' Story Is Really a Supply Story

By late summer, Western Digital had a simple message for anyone watching the AI trade: demand for hard disk drives will outrun what the industry can build, through 2027. The company is sold out for calendar 2026, has firm purchase orders from its top seven customers, and has negotiated long-term agreements that stretch into 2027 and 2028. On its fiscal fourth-quarter call, management framed the outlook as one of "durability of demand."

That framing is worth testing, because the numbers behind it do not behave like a demand story. They behave like a supply story. Western Digital is one of three suppliers — with Seagate and Toshiba — that controls essentially the entire market for the high-capacity drives the big cloud companies now hoard. Having consolidated and then disciplined its capacity after the 2022 downturn, that oligopoly never rebuilt the surplus that used to make hard drives a commodity. Demand surged and simply ran into a wall that had already been built. The result is not more unit sales so much as pricing power, and that distinction decides how durable this trend is.

The price did the work

Western Digital's fiscal fourth quarter, ended July 3, 2026, makes the mechanism legible. Revenue rose 44% year over year to $3.75 billion. But the volume behind it grew much less: shipments rose 22% to 231 exabytes. Pricing, in other words, supplied a large part of the growth on top of real unit increases, and that pricing shows up in the margin line. Non-GAAP gross margin expanded 13.1 percentage points to 54.4%, and adjusted earnings per share grew 109% to $3.56.

That is the signature of a supply-constrained market. When a shortage is driven purely by unit demand, revenue and volume rise together and margins stay roughly flat. Here volume grew moderately while the price and margin components moved far more — the tell of a market where sellers, not buyers, control the terms. The pattern is visible at the shelf level too: 24 TB drives saw price increases of up to 60% in early 2026, with average increases around 46%.

The industry backed into this position deliberately. Western Digital's capital spending remained modest relative to a company pulling in $12.9 billion of fiscal-2026 revenue; the goal of the past several years was technology migration, not new factories. Discipline, not demand, is what emptied the 2026 and 2027 order books.

Two markets, not one

The shortages are not evenly distributed, and that split matters as much as the total. Roughly 89% of Western Digital's quarterly revenue came from cloud customers — the hyperscalers that sign long-term agreements for nearline capacity at negotiated prices. The consumer market made up a sliver, and it is being starved as the manufacturers allocate nearly all of what they can make to the highest-paying enterprise buyers. That is why an ordinary buyer now pays record retail prices or simply cannot find a large drive in stock.

In a supply-constrained market, the value accrues to whoever controls the constraint — here Western Digital extracting record margins from hyperscalers — while the residual customer (the consumer, buying at spot) is a price-taker with no negotiating position. The market has bifurcated into a structural winner and a structural loser by customer class.

What breaks the shortage

The honest question for a stock already up roughly 160% for the year is not whether 2026 is tight — it clearly is — but whether the supply constraint holds through 2027, and the risk registers are all supply-side, not demand-side.

The first is new product density. Western Digital is on track to launch a 44 TB hard drive in the first half of calendar 2027 using heat-assisted magnetic recording, a technology that packs more bits onto each platter. HAMR does not require a new factory — it raises the capacity of existing capacity, which is the classic way this industry relieves a shortage without breaking pricing. UltraSMR, a stacking technology, is expected to reach about 60% of nearline exabyte shipments by the end of fiscal 2027. Both are forms of supply growth arriving precisely when the "outpacing supply" thesis is being tested.

The second is substitution. When hard drives get expensive enough and scarce enough, cloud builders shift more cold-storage workloads to cheaper flash, and the pricing lever that produced a 54% gross margin starts to bend. The constraint that is working for Western Digital today is the same constraint that, at a high enough price, hands share to an alternative.

This is what makes the stock's round trip instructive. Western Digital kept beating earnings — including in August, when it topped estimates on revenue, margin, and profit — yet the shares fell sharply and are down roughly 45% from the June all-time high of $799.87. The beat didn't disappoint; the price had already run far enough ahead of even 44% growth that guidance could not satisfy it. Even after the drawdown, the stock carries a rich multiple that embeds the assumption that pricing power compounds rather than fades.

The fundamental story — a disciplined oligopoly with record margins and strong free cash flow — is real. The mistake would be to read "demand outpacing supply" and conclude the earnings picture explains the price. The stock has already priced a great deal of the shortage. The forward question is entirely supply-side: whether HAMR's density gains and customer digestion let the constraint hold, or quietly dissolve the pricing that made Western Digital the AI trade. That, not unit demand, is what determines whether the margin stays at 54% or reverts toward the 30s it spent most of its history in.