Seagate and Western Digital Drop 10.2% on Toshiba’s…

The selloff that hit Seagate Technology and Western Digital on Friday was not a demand scare. Nothing in the session said artificial-intelligence data centers had stopped ordering hard-disk drives, or that exabytes already spoken for had been torn up. The news attacked the scarcity premium, not the unit demand that created it. Nikkei Asia reported on Oct. 2, 2026, that Toshiba plans to double hard-disk capacity for AI data centers within fiscal 2027, from the 2025 level, with roughly 60 billion yen ($380 million) of investment in the Philippines, its first major hard-disk investment in about five years. Nasdaq’s official closes, retrieved Oct. 3, put Seagate Technology Holdings plc (NASDAQ: STX) at $848.99, down 10.21%, and Western Digital Corporation (NASDAQ: WDC) at $415.29, down 10.22%. The market marked down how long two suppliers can charge for being scarce, not whether the drives are still needed.

That distinction is the whole story. A scarcity premium is a price for time. Seagate and Western Digital spent 2026 getting paid because nearline capacity was tight, gross margins jumped, and customers were signing for supply years out. Toshiba adding a plant does not subtract a drive from this quarter’s shipments. It tells a buyer negotiating a 2029 or 2031 contract that a third factory might exist by then. Analysts who called Friday overdone were making a supply-timing argument, about heads, media, exabytes, and the calendar through 2028. They were not arguing that AI demand had been misread. The selloff repriced the duration of the premium.

The capacity plan, and what it does not say

Nikkei Asia’s dispatch, filed from Manila by Keigo Yoshida at 06:59 Japan time on Oct. 2, is a factory story with a date on it. Toshiba intends to double hard-disk capacity for AI data centers within fiscal 2027, from the 2025 level, spending roughly 60 billion yen, about $380 million, in the Philippines. New lines come with drives that raise capacity per unit by as much as 40%, and with further spending toward 65-terabyte-class drives in 2030 and, later, 100-terabyte-class drives. Automation of inspection and clean rooms is supposed to cut the extra staff this build would normally need by about 40%. Nikkei put Toshiba’s share by storage capacity at just over 10%, with a medium-term aim of 30%. The build is the company’s first major hard-disk investment in about five years.

“Double” and “30%” are not the same claim. Using 10% only because Nikkei said “just over 10%,” a doubling of units plus 40% more capacity on each drive takes Toshiba to about 2.8 times its starting exabytes, roughly 28% of the old industry pool. If the other 90% does not shrink, the new share is 28 divided by 118, about 24%, not 30%. If Seagate and Western Digital also add terabytes per drive, the gain is smaller. The 30% figure is an ambition. The near-term fact is a plant and a denser drive.

“Within fiscal 2027” is not this quarter. Nikkei does not name the month that year ends. A same-day AskTraders brief glossed it as the year through March 2027. That date is the brief’s, not a line in the Nikkei text reviewed here. Either reading is a multi-year build. In the same piece Nikkei said solid-state drives cost around 20 times more than hard disks, and, citing IDC, that annual data generation is projected to reach 718 zettabytes in 2030, about four times the 2024 level, with around 60% of it on hard disks. Demand language, inside a supply headline.

Disks still clear an AI budget because memory is scarce. FinanceFeeds has framed Micron’s shortage as running toward 2028. Kioxia’s cap on NAND pricing after a 70% quarter is the other half: Toshiba’s old memory business is rationing chips, while the disk business is being asked to ration fewer disks. Friday did not kill the substitution. It asked who gets to sell it.

“Our performance is being driven by robust cloud data center demand and disciplined execution, and we see the momentum continuing in 2027,” Dave Mosley, Seagate’s chair and chief executive, said in the July 29, 2026 release. He also said Seagate sees “durable long-term demand for mass capacity storage.” That is a demand sentence. The sellers were answering a supply sentence.

Who spoke, who did not, and what the Street actually argued

The Nikkei dispatch we reviewed does not say Toshiba declined to comment. AskTraders wrote on Oct. 2 that Toshiba had not confirmed the plan, and that Seagate, Western Digital, and Toshiba did not immediately respond to requests for comment. Silence is not a denial. The only primary text on the plant that morning was Nikkei’s.

The summer filings already described demand. Irving Tan, Western Digital’s chief executive, said in the Aug. 5 release that fiscal-fourth-quarter revenue rose 44% year over year, margins expanded, and earnings per share more than doubled, on “customers’ growing storage demand.” He said the company was entering fiscal 2027 with “continued confidence in the durability of demand and with increasing visibility into our business.” Finance chief Kris Sennesael put the next quarter’s midpoint at $4.1 billion of revenue, a 55.5% non-GAAP gross margin, and non-GAAP earnings per share of $4.00. Seagate’s July 29 guide was the same $4.1 billion revenue midpoint, plus or minus $100 million, and non-GAAP diluted earnings per share of $7.30, plus or minus $0.20.

Friday’s sell-side defense was about timing, and it is a secondary account of notes this desk did not receive. ZeroHedge said Citi argued that Toshiba does not make its own media and heads, so a doubling of exabytes needs outside suppliers to expand too. Morgan Stanley, in that write-up, still saw the supply gap through calendar 2028 as wider than Toshiba’s addition and called the drop a dip to buy, noting that Toshiba lacks leading-edge capacity and HAMR. Rosenblatt read it as share reclaim rather than a sprint to 30%, with contract risk in 2029 through 2031. Startup Fortune described the same three points and also printed closes of 14.6% and 13.5%. Nasdaq’s closes are 10.21% and 10.22%.

Intraday wires were not the close. Investing.com, on Yahoo Finance at 12:18 UTC, had Seagate down 11% and Western Digital down 7%. Nasdaq’s lows were $791.75 and $396.57, 16.3% and 14.3% under Thursday’s closes. The same ZeroHedge summary said Evercore has most of Seagate’s nearline exabytes allocated into 2028, with Western Digital still negotiating toward 2031, and that an August Bernstein roadshow described full factories, no new drive units, and nearline growth near 25% a year via HAMR rather than a new plant.

ZeroHedge named Goldman Sachs analyst James Schneider on Seagate chief financial officer Gianluca Romano’s Communacopia remarks: margins supported through fiscal 2027, a HAMR crossover by the end of calendar 2026, and a later shift toward buybacks. Citi’s SanDisk work, after Micron’s NAND prices rose faster than modeled, shows the same firm can like flash and disks at once, for different bottlenecks. On parts, TrendForce, citing MoneyDJ’s reading of Nikkei, said Nitto Denko plans 28 billion yen for head-suspension components, a 40% capacity lift by the end of fiscal 2028 versus fiscal 2025, in a niche above 90% share. Nikkei separately said Western Digital plans about $1 billion in Japan through 2030 for technology and research, not a rush of extra units.

Two tapes, one headline

Wires that said Seagate was up about 244%, and Western Digital about 170%, before Friday were rounding Thursday’s 243.4% and 168.5% from the Dec. 31, 2025 closes. Dividends are excluded.

Prices are Nasdaq daily figures retrieved Oct. 3, 2026, for STX and WDC. Fundamentals are the two earnings exhibits. Seagate’s 48.5% is calculated here from $3.629 billion and $2.444 billion. Western Digital’s “up 44%” is the company’s wording; the dollars imply 43.8% before rounding. Guides are not results. Nasdaq’s quote page showed 52-week ranges of $209.00 to $1,145.00 and $112.52 to $799.87.

The paths inside the day were not the same trade. Seagate opened at $822.56, 13.0% under Thursday, hit $791.75, and closed at $848.99, above the open. Western Digital opened at $423.20, only 8.5% down, and closed at $415.29, under the open. Volume was about 3.6 times the prior 20 sessions for both, 13.58 million shares versus about 3.77 million, and 24.77 million versus about 6.79 million. Most of the giveback from June was already done. From Seagate’s June 22 close of $1,094.04 to Thursday was 13.6%. From Western Digital’s June 18 close of $746.23 to Thursday was 38.0%. Friday added a shared 10% on top. By the bell the drawdowns from those highs were 22.4% and 44.3%.

An August FinanceFeeds case had marked $415 as a bear level against a $1,050 bull case. Friday’s close was $415.29. A Sept. 11 case set $725 against $290, and the close sat inside that band. Full-year revenue was $12.195 billion at Seagate, up 34% from $9.097 billion, and $12.919 billion at Western Digital, up 36% from $9.520 billion. Seagate’s free cash flow was $3.1 billion, about eight times the reported Toshiba program. The dollars were not the event. A broken refusal to add units was.

Where the premium actually sits

Split the curve into bytes already reserved and bytes still being argued over. Nikkei said that in January Seagate had called its 2026 nearline capacity sold out. Later notes, as relayed, push Seagate’s allocation toward 2028 and Western Digital’s open talks toward 2031. This desk has not read those purchase orders. The direction is consistent: the front of the curve was spoken for, and Friday’s news hits the open years. A Philippine line that still needs heads and media Toshiba does not make is a poor way to cancel a 2026 shipment and a better way to reopen a 2030 bid.

The comparable is a renegotiation, not a demand miss. A third supplier does not take today’s unit. It takes some of the right to name the next price, which is the 2029 to 2031 risk Rosenblatt was summarized as flagging. A buyer can still be short of NAND and still need the disk. Needing the disk and paying last summer’s scarcity price are different decisions. No new export rule showed up in the sources behind the move. A glut would have been a guide-down. Friday was a headline and two 10% closes.

Density cuts both ways. Seagate’s public path is HAMR, under the Mozaic name, more exabytes without more spindles. The Bernstein summary put that nearline pace near 25% a year, with factories full. Toshiba is proposing more lines and up to 40% more capacity per drive. The per-drive increase is something the incumbents already do, because it defends margin. The scare is the units.

What has to happen next

The next fact that matters is what both companies say about exabytes past 2028 when they report the quarter guided at $4.1 billion. If the allocation language holds, Friday was a duration mark. If either call shortens the sold-out window, or walks gross margin back toward last year’s 40% area from this year’s mid-50s, the premium was already shorter than the summer filings implied. Price here has been a margin story, and margin has been full factories plus more terabytes per drive. A line that is still a plan does not change those two sentences.

The second tell is whether Seagate or Western Digital answer with their own unit capacity. The August roadshow summary said they would not, and that a factory takes two years they did not intend to spend. If a later call adds units, the discipline story is over whether or not Toshiba reaches 30%. If they do not, Toshiba is bidding for share inside a market the larger two still will not expand in spindles, and the head-and-media constraint is the schedule. Nitto Denko’s reported build, if it holds, is a fiscal 2028 add in a component niche above 90% share. That valve, not the word “double,” is the timeline.

The third tell is the contract, not a press release about a cheaper drive. Seagate’s nearer exabytes are the ones described as allocated. Western Digital’s open talks, out toward 2031 in the Evercore summary, are where a new bidder can be mentioned. That fits Friday’s tape without proving it: Western Digital had already given back more of the June peak, and it sold off after the open, while Seagate was bought off the low. AI demand is why the margins got to the mid-50s. Toshiba’s plan attacks how long they last. The 10.21% and 10.22% closes were the price of putting a date on a scarcity that 2026 had treated as open-ended. The date is not next week.

FAQs

Why did Seagate and Western Digital shares fall on Oct. 2, 2026?

Nikkei Asia reported that Toshiba will roughly double hard-disk capacity for AI data centers within fiscal 2027 and spend about 60 billion yen ($380 million) in the Philippines. Seagate closed down 10.21% at $848.99 and Western Digital down 10.22% at $415.29, on Nasdaq’s official record. The move priced a shorter scarcity premium, not a report that data-center customers had canceled drive orders.

How large is Toshiba’s plan relative to Seagate and Western Digital?

The yen outlay Nikkei published, about $380 million, is a fraction of either company’s recent quarter. Seagate’s fiscal 2026 free cash flow was $3.1 billion. Western Digital just reported $3.747 billion of quarterly revenue. The market reaction was about a 10% supplier saying it wants a path toward 30% and will add unit capacity, not about the dollar size of the check.

Did Toshiba, Seagate, or Western Digital comment on Friday?

Nikkei’s report did not carry a Toshiba confirmation. AskTraders wrote on Oct. 2 that Toshiba had not confirmed the plan and that Seagate, Western Digital, and Toshiba did not immediately respond to requests for comment. Both U.S. companies had, in summer earnings releases, described cloud and AI storage demand as durable into fiscal 2027. Those comments predate the Nikkei story.

Were the stocks still up for 2026 after the drop?

Yes. Using Nasdaq closes, Seagate finished Friday up 208.3% from the Dec. 31, 2025 close of $275.39, and Western Digital up 141.1% from $172.27. Through Thursday the figures were 243.4% and 168.5%, which is the source of the “about 244% and about 170%” lines in some briefs. The calculation is price only, without dividends.

What did the latest reported quarters actually show?

Both companies reported a quarter ended July 3, 2026. Seagate’s revenue was $3.629 billion, up 48.5%, with non-GAAP diluted EPS of $5.71 and a non-GAAP gross margin of 52.7%. Western Digital’s revenue was $3.747 billion, up 44%, with non-GAAP diluted EPS of $3.56 and a non-GAAP gross margin of 54.4%. Each guided the next quarter to about $4.1 billion of revenue. Those are filings, not Friday updates.

Why did some analysts call a 10% drop overdone?

The accounts published by ZeroHedge and Startup Fortune describe a supply-timing case, not a claim that demand is stronger than feared. Citi, in that telling, stressed that Toshiba does not make its own media and heads. Morgan Stanley pointed to a supply gap that still runs past Toshiba’s addition through 2028. Rosenblatt framed a share reclaim and a contract risk in 2029 to 2031, not a near-term flood of drives.