To The Moon

$Western Digital(WDC)$  

Mathematical Money | October 7, 2026

Storage got hit twice in three sessions and most of the coverage I read treated it as one story. It wasn't. There were two separate stories on two different days, and the second one is the one that matters.

Here's the damage first, and it caught $WDC$ and $STX$ in near-identical shape. Western Digital went from $462.56 last Thursday to $415.29 on Friday, a 10.2% drop, bounced 6.3% on Monday, then fell another 6.9% on Tuesday to close at $411.04. Seagate did the same thing but worse, $945.57 down to $848.99, up to $887.09, then down 9.2% to $805.63. Three days, Western Digital off 11.1% and Seagate off 14.8%, and both of them closed below the panic low they'd made on Friday.

Friday was a capacity story

Nikkei reported that Toshiba will double its data-centre hard drive capacity by fiscal 2027 against its 2025 level, spending roughly 60 billion yen to expand its plant in the Philippines. It's their first serious HDD expansion in about five years. Toshiba holds a little over 10% of the market by capacity and has said it wants 30% over the medium term, with next-generation drives up to 40% denser and 65TB-class units in mass production by 2030.

The market sold it, thought about it over the weekend, and bought it back on Monday. The bear argument was fair enough. New capacity from a 10% player takes years to actually arrive, and the industry is in shortage right now, so the near-term pricing picture doesn't change much.

Tuesday was a supply-chain story, and it's the real one

On Tuesday it came out that Seagate and Toshiba are both bidding for TDK's magnetic-head business, several billion dollars, with Toshiba in talks since the spring and Seagate coming in over the summer with a higher offer.

Magnetic heads are the component that reads and writes the platter, and TDK is the only independent maker of them left. It supplies all three drive makers. TDK wants out because the business is low-margin and swallows capital to keep pace with recording technology, and they'd rather put that money into batteries and sensors.

Here's the bit that connects the two days, and it took me a while to see it. Seagate and Western Digital make their own heads in-house and only buy from TDK when demand spikes. Toshiba buys all of its heads from TDK. Toshiba also needs TDK's technology for its HAMR roadmap, which is the recording method that would let it build the high-capacity drives AI data centres actually want.

So Friday's announcement and Tuesday's auction are the same story. Toshiba cannot double output, or go from 10% share to 30%, without locking down that head supply. Friday looked like a press release you could shrug at. Tuesday showed the mechanism sitting underneath it, which is why the Monday bounce got erased completely.

It also explains why Seagate fell harder than Western Digital on Tuesday. Seagate is the one with the chequebook out, bidding billions for a business whose output it mostly already makes itself. That's money spent to stop a competitor rather than to add capacity, and the market reads that as a worse use of capital than growth spending. Western Digital isn't bidding at all, so whoever wins, the only independent supplier in its chain ends up owned by a rival.

The check I'd suggest doing yourself

This is the part I think is genuinely useful, and it works on any sector selloff.

When a whole industry drops, go and look at what the customers did. Over those same three sessions $MU$ fell 4.7%, which is modest, and it doesn't make hard drives anyway. And $NTAP$ — NetApp, which buys drives and sells storage systems built around them — rose 6.2% and closed at an all-time high on Tuesday. Up on both of the days that hammered the drive makers.

A company that buys drives hitting a record while the two companies that make them fall double digits tells you something quite specific. Nothing in those three sessions said AI storage demand is weakening. If demand were the problem then NetApp would be falling hardest of the four, not printing a record. What the market priced was a possible shift in who gets to keep the margin, and cheaper drives are straightforwardly good news if you happen to be the one buying them.

Being able to spot that distinction quickly matters, because a demand shock and a pricing-power shock need completely different responses. One means you were wrong about the end market. The other means you were right about the end market and wrong about who captures it, which is a much more fixable mistake.

What the headlines bury

Almost every article I read noted that Western Digital is still up around 140% this year. True, it's up 138.6%. Steady on, though, because that framing hides the thing you'd actually want to know.

Western Digital closed at $746.23 on 18 June and now sits at $411.04, which is 44.9% below its June high. Monthly closes since then went $544.84, $450.55, $454.46 and now $411.04. One up month in four. For comparison, Seagate is 26.4% off its own peak and Micron 13.8%.

So this news landed on a stock already four months into a decline, not on one sitting at a high. Of the four names I looked at Western Digital is the most damaged by a wide margin, and a year-to-date figure measured off a low base conveniently obscures that. I'd check the drawdown from the high before I'd check the YTD number on anything that has tripled.

Where I sit in this

I hold January 2028 calls at the $320 strike, deep in the money so most of what I paid is intrinsic value rather than time premium. They're down about 14%. I also closed the short calls I'd written against them when they got cheap on Tuesday, so the long side is currently uncapped and earning nothing, which is a decision I'll have to revisit this week.

The expiry is the part I'd defend. Whatever this question resolves to, it resolves on a 2027-to-2030 timetable: capacity arriving, HAMR drives shipping, head supply settled. A six-month option here would be a bet on sentiment turning. January 2028 is a bet on the structure, and the structure is the thing actually in dispute.

The uncomfortable position is a short $400 put expiring this Friday. With the stock at $411.04 that's 2.7% of room, which sounds fine until you check the volatility. Sixty-day realised is running at 81.9% annualised, about 5.2% a day, so two sessions gives you a standard deviation of roughly 7.3%. A 2.7% buffer against a 7.3% move is somewhere around a one-in-three chance of assignment. Same lesson I keep relearning: a buffer means nothing until you measure it against how much the thing actually moves, and this one moves plenty.

If it does get assigned I'd own the stock at an effective $394.71, which is a second entry rather than a problem, given I'm already long the January 2028 calls. Small size either way.

What would change my mind

Not the price, and not the Toshiba capacity number on its own.

Who wins TDK. If Toshiba takes it, the capacity plan is funded and supplied, the HAMR roadmap has its critical component, and the pricing premium the two incumbents have enjoyed starts compressing on a real timetable. If Seagate takes it, Toshiba's 30% ambition has a hole straight through the middle of it and the duopoly holds, paid for with several billion dollars of Seagate's money. Those two outcomes point in opposite directions for the same stock, which makes the auction the single most important item on my watch list for this name.

Two things I'd like other views on.

Does anyone else routinely check the customers when a sector gets sold off? I only started doing it properly this year and it has caught me out twice already, in a good way. Once on a semicap selloff where the equipment buyers never flinched, which told me the story was about market share rather than capex, and I'd have sold into it otherwise.

And on the structure question: when you hold deep in-the-money LEAPS and buy back the short calls into a drop, do you rewrite them straight away or wait? Writing again immediately collects more premium while volatility is high, but it caps you at exactly the point a bounce is most likely. I've gone both ways and I'm still not sure which one is the discipline and which one is the greed.

#WDC #STX #semiconductors #options #LEAPS #AIinfrastructure

Stop guessing. Start calculating.

Live to fight another day. 🤙