Why Western Digital & Seagate Stocks Are Dropping: Toshiba’s AI HDD Push Explained

 

Two of this year's strongest technology stocks were repriced in a single session on October 2. Seagate (STX) and Western Digital (WDC) both closed down roughly 10%, with Western Digital trading as low as $398.51 intraday, about 13.8% below the prior close. The trigger was not a demand warning, a guidance cut, or a customer loss. It was a Nikkei report that Toshiba intends to double its hard disk drive capacity for AI data centers.

 

For more than a year, the bull case on both names has rested on a single premise: nearline drives are scarce, cloud operators have locked up output years in advance, and the three remaining manufacturers therefore hold unusual pricing power. That premise produced non-GAAP gross margins of 52.7% at Seagate and 54.4% at Western Digital. What the market traded on Friday was the probability that the premise gets diluted after 2027, not the end of AI storage demand. On a price chart the two look identical. In the fundamentals they are not remotely the same thing.

 

 

 

The investment is modest in dollar terms but pointed in direction. Toshiba is reported to be spending roughly 60 billion yen, close to $380 million, on its Philippines facility to double AI data-center drive capacity by fiscal 2027, its first major HDD investment in about five years. The number that unsettled investors is the medium-term target: lifting share measured by storage capacity from just over 10% to 30%.

 

The selloff was tightly contained within hard drives. The Roundhill Memory ETF fell only 0.3% that day while the Nasdaq 100 ETF rose 0.6% and SanDisk slipped about 1.8%. What was repriced was the industry's supply curve, not AI storage demand.

 

Both businesses were still accelerating when the news hit. Seagate grew fiscal 2026 revenue 34% to $12.195 billion, with nearline exabyte shipments jumping from 497 to 695. Western Digital grew revenue 36% to $12.9 billion over the same period, and its blended price per terabyte rose from high single digits in the prior quarter to the high teens in the June quarter.

 

Wall Street largely called the drop overdone. Morgan Stanley, Rosenblatt and Citi all kept positive ratings, arguing that Toshiba's addition is too small to close the industry gap and that Toshiba does not make its own heads and media, leaving the expansion hostage to external component suppliers.

 

Toshiba's new output cannot reach the market before fiscal 2027 at the earliest, while the vast majority of Seagate's nearline exabytes are already allocated into calendar 2028. The real test comes during the 2028 to 2029 contract renewal cycle, not this quarter.

 

 

 

Benzinga's premarket tracking showed Seagate down 11.22% at $839.44 and Western Digital down 7.95% at $425.80 before the open, at a moment when Nasdaq futures were up 1.31% and S&P 500 futures up 0.97%. Losses widened once trading began. Stocktwits reported Seagate down 12% and Western Digital down 11% intraday, with both settling near a 10% decline.

 

The divergence matters. 24/7 Wall St. observed that the Roundhill Memory ETF fell just 0.3% while the Nasdaq 100 ETF gained 0.6%. Had the market been pricing a peak in AI storage demand, DRAM and NAND names would not have escaped untouched. What was reassessed was the competitive structure among three drive makers and the long-term price curve that structure supports.

 

Positioning amplified the move. Even after the decline, Stocktwits data put Seagate up roughly 203% year to date and Western Digital up roughly 139%. When a valuation already embeds years of tight supply, any challenge to that assumption gets priced with force well beyond its near-term economic weight.

 

 

According to Nikkei, Toshiba plans to invest about 60 billion yen expanding its Philippines plant, doubling AI data-center drive capacity by fiscal 2027, with new products offering up to 40% more storage per drive. Details compiled by Stocktwits add that Toshiba's share by storage capacity stands just above 10% against a medium-term target of 30%, that it aims to mass produce 65-terabyte-class drives in 2030 and eventually 100-terabyte-class products, and that automating inspection and clean-room processes could cut the extra staffing an expansion normally requires by about 40%.

 

One precondition deserves attention. Since the consortium led by Japan Industrial Partners took it private, Toshiba has been delisted from the Tokyo Stock Exchange and no longer carries listed-company disclosure obligations. The source for this expansion is therefore a press report rather than a corporate filing. Investors adjusting models should file it under reported by credible media but not officially confirmed, not under settled fact.

 

The share figures also need care. TrendForce data from July put Western Digital and Seagate each above 40% of the global HDD market on a unit basis, with Toshiba around 17%. Measured by storage capacity, Toshiba sits much lower, just above 10%, which is a direct reflection of a thinner high-capacity product line. Moving from 10% to 30% of capacity requires more than assembly lines. It requires the drives to be competitive at the top of the range.

 

 

 

AI storage demand is often reduced to the idea that training needs fast storage. The actual architecture is tiered. Solid-state drives handle throughput- and latency-sensitive training and inference work, while hard drives carry long-term retention of training corpora, model checkpoints, inference logs and cold archives. TrendForce notes that within the traditional tiered architecture, nearline drives have long dominated cold storage on the strength of exceptionally low cost per gigabyte, and that the rapid expansion of inference applications is pushing cold-data demand higher still.

 

Nearline refers to the layer between online storage and offline archive: data that does not need millisecond response but must remain retrievable at any time. These are almost exclusively 3.5-inch, helium-filled, high-capacity enterprise drives running from 24TB up through the 40TB class. The customer base is highly concentrated among a handful of hyperscale cloud operators. Seagate's annual report shows OEM customers accounting for 81% of revenue in fiscal 2026, with the data center share of revenue rising from 75% to 80%.

 

What AI changed is not the type of storage but the absolute volume and the slope of its growth. Training requires retaining raw data, cleaned datasets and intermediate state. Once inference scales, every call generates context, output and audit records that have to land somewhere. The overwhelming majority of that data does not need flash performance but does need affordable retention, which is precisely the nearline role. Cost per terabyte for flash remains materially higher in cold-storage use, and TrendForce has cautioned that providers moving cold data onto QLC SSDs must update data systems, verify software compatibility and track costs closely.

 

 

None of the three makers expanded aggressively in recent years, leaving the supply curve close to rigid. TrendForce reported in January that Seagate CEO Dave Mosley described nearline capacity as fully allocated through calendar 2026, with orders for the first half of 2027 opening in the following months. By the July earnings call, that language had become the vast majority of nearline exabytes allocated into calendar 2028. TrendForce supply-chain work cited by TechRadar showed nearline lead times stretching from a few weeks to beyond 52 weeks.

 

That is the true source of pricing power. In a market booked out years ahead, a manufacturer does not need a technology gap to command a premium. It only needs to raise price at renewal to a level customers will still accept, because supply assurance has outranked unit cost as the buyer's priority. Toshiba's plan stings not because of how many drives it might take, but because it could hand buyers another option at the next negotiating table.

 

 

 

Seagate's fiscal fourth quarter and full year results show GAAP gross margin of 52.3% and non-GAAP gross margin of 52.7% for the quarter, with full-year figures of 45.6% and 46.1%. Earnings call coverage adds that non-GAAP gross margin expanded 570 basis points sequentially, operating margin reached 44.6%, the company shipped 218 exabytes in the quarter for 34% year-over-year growth, and HAMR-based products made up roughly 40% of the nearline exabyte run rate at the end of fiscal 2026.

 

The annual numbers capture the slope better. The 10-K reports fiscal 2026 revenue of $12.195 billion, up about 34%, driven primarily by higher nearline exabyte shipments and favorable pricing actions. Nearline shipments rose from 497 exabytes to 695, while non-nearline edged down from 98 to 94. Free cash flow reached a record $3.1 billion and the company retired $1.4 billion of debt. Management guided fiscal first quarter 2027 revenue to roughly $4.1 billion with non-GAAP EPS of $7.30.

 

 

Western Digital's fiscal fourth quarter release reported revenue of $3.75 billion, up 44% year over year, GAAP gross margin of 54.1%, non-GAAP gross margin of 54.4%, operating cash flow of $1.39 billion and free cash flow of $1.28 billion. The 8-K filed the same day sets out fiscal first quarter 2027 guidance of $4.1 billion in revenue plus or minus $100 million, non-GAAP gross margin of 55% to 56%, and EPS of $4.00 plus or minus $0.15.

 

The operating detail from the call matters more. CFO Kris Sennesael said the company shipped 231 exabytes in the quarter, up 22%, that gross margin expanded 1,310 basis points year over year, that the blended average price increase per terabyte improved from high single digits in the prior quarter to the high teens, and that cost per terabyte fell about 8% year over year. For the full year, revenue grew 36% to $12.9 billion, gross margin expanded 970 basis points to 49.1% and operating margin expanded 1,290 basis points to 37.3%. The company's fiscal Q4 earnings presentation traces a non-GAAP gross margin path of 41.3%, 43.9%, 46.1%, 50.5% and 54.4% across five consecutive quarters.

 

Put the two sets side by side and the sensitivity becomes obvious. Price contributed far more to margin than cost reduction did, and price depended entirely on scarcity. Introduce a third-party supply variable and the same leverage runs in reverse.

 

 

 

Toshiba's target points at fiscal 2027, which means physical output cannot reach customers much before 2028. Until then, the two American producers have their capacity under contract. Seagate said in late July that the vast majority of nearline exabytes are allocated into calendar 2028, while Western Digital stated earlier this year that it was essentially sold out for calendar 2026 and holds commercial agreements with core customers extending into 2027 and 2028.

 

That is the basis for Morgan Stanley's view that the reaction was overdone. The bank noted that storage upcycles have historically ended with a supply shock rather than a sudden demand drop, so any credible capacity announcement invites scrutiny, yet argued that the gap between HDD supply and demand through calendar 2028 still looks wider than Toshiba's planned addition. It also pointed out that Toshiba lacks leading-edge capacity and heat-assisted magnetic recording, the process Seagate has been scaling for its highest-capacity drives, and reiterated Overweight ratings on both names.

 

Rosenblatt framed it differently but landed close. The firm read the plan as an attempt to reclaim share toward the stated 30% ambition after years of losses, rather than a broad supply surge, while still calling it a credible medium-term risk because extra capacity gives large customers incremental leverage in long-term agreements stretching into 2029 through 2031.

 

 

Citi analyst Asiya Merchant identified a structural limit. According to Investing.com, unlike Seagate and Western Digital, Toshiba does not manufacture its own media and heads and depends on outside suppliers. Doubling drive output requires those suppliers to raise component production in tandem, and that dependency could cap the net exabytes that actually reach the market.

 

Supply-chain evidence supports the point. TrendForce has tracked expansion among Japanese component makers including TDK, JX Advanced Metals, Resonac, Nitto Denko and HOYA, with TDK stating in May that it would add capacity for HDD heads and suspensions to meet nearline demand. Those expansions are underway, but such lines typically ramp over years, and they serve all three drive makers rather than prioritizing Toshiba's doubling ambition.

 

In short, Nikkei reported an intention and a capital commitment, not a locked-in supply increment. Between intent and the first platter leaving a clean room sit product qualification, customer adoption and component availability.

 

 

The two trade as one position but their exposures are not identical. Having separated its flash business, Western Digital is now a pure-play hard drive company with no NAND business to cushion the news, which helps explain an intraday decline exceeding 13% on October 2. Seagate also sells enterprise nearline SSDs and storage systems, but nearline drives remain the core, with data center accounting for 80% of fiscal 2026 revenue.

 

Technology roadmaps are the more durable differentiator. Seagate has pushed HAMR to roughly 40% of its nearline exabyte run rate, while Western Digital began shipping next-generation 40TB ePMR drives in the June quarter and targets having that platform exceed 50% of nearline exabytes by the third quarter of calendar 2027. Whichever path drives cost per terabyte down faster over the next few years will determine bargaining power at renewal. By comparison, TrendForce indicates Toshiba is advancing MAMR with 30TB to 34TB drives and plans 40TB-class products in 2027, leaving it a step or two behind at the top of the capacity range.

 

For traders tracking these swings, MEXC lists Ondo-issued tokenized US equities, including Seagate and Western Digital, which can be traded directly against stablecoins and offer a read on sentiment while traditional markets are closed. Tokenized equities are minted and redeemed by a third-party issuer, and eligibility, trading hours and liquidity differ from the underlying shares.

 

Hard drive stocks tend to do their damage within a few hours, and trading Western Digital's tokenized stock with USDT lets you adjust before the US open rather than after it.

 

 

 

The first is the one the selloff actually traded: supply catching up after 2028. It is real but deferred, and it bears on renewal pricing rather than current earnings. The second is execution risk, meaning whether Toshiba can double output under component constraints, something no official document currently allows anyone to verify. The third is valuation risk, since after more than doubling this year both stocks carry amplified sensitivity to any negative headline, a dynamic that has nothing to do with the drive industry itself.

 

One variable gets less attention than it deserves: the cadence of cloud capital expenditure. Nearline demand is concentrated among a few buyers, so a delay in AI infrastructure spending would hit order books faster than any Toshiba expansion.

 

 

In a continued tightness case, Toshiba's addition is absorbed by demand growth, contract pricing holds firm through 2028, and both companies keep gross margins above 50%. This is the implicit base case behind the Morgan Stanley and Rosenblatt views.

 

In a gradual loosening case, Toshiba delivers on schedule with component supply keeping pace, buyers gain leverage in the 2028 and 2029 renewal rounds, and price growth per terabyte compresses from the current high teens toward mid single digits. Revenue could still grow, but margin expansion stops and the multiple has to retreat from cycle highs.

 

In a demand-side case, cloud storage procurement slows and the shortage disappears before new supply arrives. This carries the lowest probability and the largest impact, and the evidence for it would come from hyperscaler capital expenditure guidance rather than from any drive maker's capacity announcement.

 

 

The nearest test is earnings. Based on historical reporting patterns, both Seagate and Western Digital are expected to report fiscal first quarter 2027 results in late October, with exact dates not yet formally confirmed. What matters is not whether the quarter beats, but whether management's language around 2028 and beyond shifts: whether customers are still actively asking to extend planning horizons past 2029, and which way renewal pricing is moving.

 

Over the medium term, component maker capacity is the leading indicator of whether Toshiba's plan converts into shipments. Quarterly outlooks from suppliers such as TDK sit closer to the real supply curve than any drive maker's statement. Whether Toshiba issues a formal confirmation of the plan is itself a signal worth watching.

 

 

For James Mitchell, the significance of October 2 is that the market finally put a risk price on the hard drive oligopoly's tacit discipline. The outsized returns of the past year came not from a technology breakthrough but from the simple fact that nobody was adding capacity. Seagate's price per terabyte and Western Digital's 1,310 basis points of gross margin expansion are both functions of a rigid supply curve. The moment one participant signals it will break ranks, even for $380 million, the weakest link in that argument gets retested. From a risk management standpoint, holders of these positions should treat competitive structure, rather than AI demand, as the primary monitoring variable.

 

Two readings look wrong. The first equates a capital commitment with immediate supply. Toshiba's target points at fiscal 2027 while the two American producers have capacity contracted into calendar 2028, leaving at least a two-year buffer during which reported results will not change because of this story. The second treats the path from 10% to 30% of capacity share as given. The component constraint Citi identified is structural: not making your own heads and media means expansion speed is set by an upstream industry serving all three customers. More importantly, this plan currently exists only as a press report, and Toshiba has carried no disclosure obligation since going private at the end of 2023, so it should not be modeled as confirmed industrial fact.

 

What deserves the most attention next is contract language rather than quarterly figures. Seagate said in late July that the vast majority of nearline exabytes are allocated into calendar 2028. If management on the late-October call continues to describe customers extending planning horizons and discussing 2029, pricing power has not moved. If the framing shifts toward intensifying competition or more challenging negotiations, the buyer-leverage effect Rosenblatt flagged has begun to show. The matching high-frequency indicator is component ramp progress, which tracks real supply more faithfully than any manufacturer statement. On technology, the rate at which HAMR and ePMR push cost per terabyte lower will determine who still holds pricing power after 2028.

 

Across asset classes, the episode carries a lesson that extends to crypto and fintech. Every excess return built on constrained supply eventually faces the same question: is the constraint physical or elective. Bitcoin's supply cap is written into the protocol and cannot be moved by capital expenditure. The hard drive ceiling was only a shared choice by three firms in a particular cycle, and a single investment was enough to shake confidence in it. When an industry's margins rest on collective restraint rather than structural barriers, the valuation contains an implicit discipline premium, and that premium never gives notice before it disappears. For investors who size positions through a cycle and risk-management lens, that is reason enough to decouple position size from recent performance.

 

 

 

Western Digital closed down about 10% on October 2, trading as low as $398.51 intraday, roughly 13.8% below the prior close. The direct cause was a Nikkei report that Toshiba plans to invest about 60 billion yen expanding its Philippines plant to double AI data-center hard drive capacity by fiscal 2027. The concern is that future supply may catch up with demand and erode current pricing power, not that storage demand has weakened. As a pure-play hard drive company since separating its flash business, Western Digital has no other segment to cushion the move.

 

 

Available evidence does not support that. The Roundhill Memory ETF fell only 0.3% that day, the Nasdaq 100 ETF rose 0.6% and SanDisk slipped about 1.8%, leaving the selloff concentrated in two hard drive names. Both Seagate and Western Digital grew revenue more than 40% year over year in their most recent quarters at record gross margins, and nearline capacity at both remains contracted into roughly 2028. The supply curve was repriced, not the demand curve.

 

 

A nearline drive sits between high-speed online storage and offline archive: data that does not require millisecond response but must stay retrievable, typically on 3.5-inch helium-filled high-capacity enterprise drives. AI training and inference generate enormous volumes of data that are rarely accessed yet must be retained, including raw corpora, model checkpoints, inference logs and archives. Cost per terabyte for flash remains materially higher in those use cases, so nearline drives form the cold-data foundation of AI data centers.

 

 

On the reported timeline, the doubling targets Toshiba's fiscal 2027, which means output would reach the market no earlier than around 2028. Before then, the vast majority of Seagate's nearline exabytes are allocated into calendar 2028 and Western Digital holds agreements with core customers running into 2027 and 2028. The meaningful window is the 2028 to 2029 contract renewal cycle rather than current spot pricing.

 

 

Morgan Stanley, Rosenblatt and Citi all maintained positive ratings and treated the decline as overdone. Morgan Stanley argued the supply-demand gap through calendar 2028 still exceeds Toshiba's planned addition, noted that Toshiba lacks HAMR technology, and said it would buy both on the dip while reiterating Overweight ratings. Rosenblatt read the plan as share reclamation rather than a supply flood, while calling it a credible risk to agreements running into 2029 through 2031. Citi emphasized that Toshiba does not make its own heads and media.

 

 

Their exposures differ. Western Digital is a pure-play hard drive company without a flash business to offset pricing swings, giving it more leverage to drive economics. Seagate also sells enterprise nearline SSDs and storage systems, though data center still accounted for 80% of fiscal 2026 revenue. On technology, Seagate has HAMR at roughly 40% of its nearline exabyte run rate, while Western Digital is backing 40TB ePMR with a target of more than 50% of nearline exabytes by the third quarter of calendar 2027. Whichever reduces cost per terabyte faster will hold more bargaining power.

 

 

The nearest checkpoint is fiscal first quarter 2027 results, expected in late October based on historical reporting patterns, with dates not yet formally confirmed. The signal lies in how management describes contract terms for 2028 and beyond rather than in the quarter itself. Over the medium term, track capacity ramps at upstream component suppliers such as TDK, which determine whether Toshiba's plan converts into shipments, and watch for any formal statement from Toshiba, since the current source remains a press report.

 

 

The information above is provided for general market information and analysis only and does not constitute investment advice, financial advice, legal advice, tax advice or a recommendation to trade. Prices of crypto assets, equities, tokenized equities and other related financial assets can fluctuate sharply, and past performance, technical indicators and on-chain data do not guarantee future results. Share prices, financial figures, analyst views and industry plans referenced here may change at any time, and some information originates from press reports rather than official announcements, so the latest formal disclosures from the relevant companies and regulators should be treated as authoritative. Tokenized equities are minted and redeemed by a third-party issuer, and their eligibility, trading hours, liquidity and legal characteristics differ from traditional securities and may not be available to users in all jurisdictions. Readers should conduct their own research and make decisions based on their own financial circumstances, investment objectives and risk tolerance, consulting a qualified professional where appropriate. The MEXC Crypto Pulse team accepts no liability for any direct or indirect loss arising from the use of this information.

 

 

James Mitchell specializes in technical analysis, market trends, and trading strategies for both Bitcoin and altcoins. Based in London, he has over 10 years of experience in financial markets. Before joining MEXC Learn, James worked as a senior analyst at a leading European investment firm, where he developed expertise in risk management and quantitative trading. His transition to cryptocurrency markets began in 2017, and he has since become recognized for his data-driven approach. He holds a Master's degree in Financial Economics from the London School of Economics. His analytical approach combines traditional technical analysis with on-chain metrics to provide readers with actionable insights.

 

His areas of expertise span technical analysis, market trends and cycles, trading strategies, Bitcoin and altcoin analysis, and risk management.

 

 

 

Want the fastest access to MEXC's latest updates? Join our official Telegram group now!

MEXC Platform: Download App | Zero Fee Trading | Buy USDT Now

Join MEXC Community: X (Twitter) | Telegram | Discord

Get Insights: MEXC Learn | Terms of Service | MEXC Blog | MEXC News

Account Verification: Understand KYC | How to Complete KYC

External Content Platforms: Substack | Medium | Paragraph | LinkedIn | X(News)