International Business Machines (IBM) has spent three years quietly convincing Wall Street it belonged in the artificial intelligence trade.
However, on July 14, the market decided otherwise in a single session. IBM shares closed at $217.07, falling 25.21% that day alone.
This came after the company pre-announced a second-quarter miss that nobody saw coming. That was its worst drop since 1968, even worse than Black Monday in October 1987.
Roughly$68 billion in market value vanished, and Big Blue is now worth about $204 billion.
Here’s the part that should interest investors more: Micron Technology (MU) went up the same day, on the same news.
Why IBM’s earnings warning erased $68 billion in one session
IBM told investors its preliminary second-quarter revenue came in at $17.2 billion, against the $17.86 billion Wall Street expected.
The company also reported adjusted earnings of $2.93 a share versus the $3.01 consensus, according to Forbes.
The miss itself was small. But the reason behind it was not.
More AI Stocks:
- Micron stock jumps as investors look beyond GPUs in AI chip trade
- The entire chip market is waiting for TSMC’s July earnings report
- Major AI chip stock plunges after blockbuster $26.5 billion Nasdaq debut
CEO Arvind Krishna wrote to investors that clients redirected budgets toward servers, storage, and memory in the final weeks of June to lock down supply before prices climbed.
IBM misjudged how much money would move.
The CEO admitted the company “did not anticipate the magnitude of the capex reprioritization” and conceded that “this quarter we faltered,” Axios reported.
Krishna also said numerous large deals failed to close on the timelines IBM expected. This failure drove most of the shortfall, International Business Times reported.
What “capex reprioritization” means for your portfolio
A company has a fixed technology budget, and when one line item gets more expensive, another line item dies.
Chief information officers did not get bigger budgets in June. They got scarier hardware prices.
So CIOs bought the servers and the memory first, because you cannot run an AI workload on a purchase order.
Then, they pushed the software renewals and the consulting engagements into the next quarter or out of the plan entirely.
Even with this, IBM’s software segment still grew 5%, with Red Hat up 11%. Although consulting was roughly flat, infrastructure fell 7%, The Motley Fool reported.
The damage was concentrated, not universal. That distinction matters because IBM did not lose to a competitor; it lost to a spreadsheet.
Why IBM’s fall is the validation Micron has been waiting for
Micron makes high-bandwidth memory, or HBM.
HBM is a stacked version of DRAM that sits directly beside an AI processor and feeds it data fast enough to keep the chip busy. Every serious AI server needs it.
For most of its history, Micron was a stock that ate its own gains every few years. Prices boomed, capacity flooded in, margins collapsed, repeat.
Related: Veteran analyst drops massive Micron valuation prediction
IBM just handed the company evidence that this cycle is different.
When enterprises let software deals that are nearly closed die rather than risk missing a memory allocation, that is not elastic demand. That is a customer base with no substitute and no patience.
Micron shares reflected it immediately. The stock traded around $980.55 midday on July 14, up 4.7% from the prior close.
This was after KeyBanc analyst John Vinhraised his target to $1,750, Barron’s reported.
The supply math behind Micron’s pricing power
The reason CIOs panicked in June is that the memory capacity they need was already spoken for.
Samsung, SK Hynix (SKHY), and Micron control more than 95%of global DRAM production.
All three companies prioritized converting capacity to HBM because it earns three to five times the revenue per wafer of conventional memory, Stocks Down Under noted.
Three facts investors should hold onto
- The order book is closed. Micron’s memory capacity is sold out well into 2027, and the company has secured about $100 billion in long-term customer agreements.
- The shortage has a stated end date, and it’s far away. On the June 24 earnings call, CEO Sanjay Mehrotra said market tightness is “locked in to persist beyond calendar 2027.”
- Prices are still climbing. KeyBanc forecasts DRAM up 15% to 20% in the third quarter and another 15% in the fourth, with HBM prices potentially more than doubling next year.
That combination is why Micron’s last quarter looked the way it did. Fiscal third-quarter revenue hit $41.46 billion, up about 346% from a year earlier, with gross margin at a record 84.9%.
The software stocks that fell with IBM, and why
When IBM warned of its business’s struggles, investors did not think it was a problem unique to IBM. Instead, they realized that other software companies would soon face the exact same troubles.
Microsoft (MSFT) closed at a 1.55% drop, Oracle (ORCL) fell 2.74%, and Accenture (ACN) dropped 2.86% on July 14, according to Bloomberg.
Salesforce (CRM) dropped about 4% in early trading before recovering, while Microngained 5% in the afternoon and SanDisk (SNDK) rose nearly 6%.
The pattern is the trade. Money left the application layer and showed up in the hardware layer within the same session.
HSBCdowngraded IBMto reduce from hold with a $191 price target, which was the sharpest bear call of the day, The Motley Fool noted.
What has to stay true for the memory trade to keep working
Zero-sum budgets are good for Micron only while the shortage lasts, and shortages have a well-documented habit of ending.
Four things need to hold:
- Supply stays behind demand. New fabs in Idaho, New York, and Virginia are being built precisely to fix this, and when they ramp, the pricing argument weakens.
- Contract prices stick. Micron’s strategic customer agreements are take-or-pay, so customers can walk away but forfeit what they’ve committed.
- AI capex keeps growing. If hyperscalers pause, memory has no second buyer of that size.
- The legal overhang stays quiet. A June 25 class action lawsuit accuses Micron, Samsung, and SK Hynix of restricting supply to inflate prices by as much as 700% since 2022.
Not everyone believes the cycle broke.
Michael Burry shorted Micron at $1,051.87 on July 1, noting the stock has suffered 34 drawdowns of more than 30% over the past 42 years.
What investors should watch between now and July 22
IBM’s full results and conference call land on July 22, and that is the next real decider event for both sides of this trade.
For IBM holders, the question is simple: Did those deals slip, or did they die? Krishna needs to put a number on it on July 22, and a vague answer suggests the software drought isn’t over.
For Micron holders, every delayed software deal is good news. The July 22 call will show whether June was a one-month thing or the new normal.
Practical framing for both:
- Position size matters more than conviction here, since a 25% single-day move in a 115-year-old blue chip is a reminder that nothing is too big to reprice.
- Watch software peers into the July 22 results to see whether this was a one-day repricing or the start of a broader derating.
- Micron already trades on an assumption of extended tightness, so the risk is not that demand disappoints; it’s that supply arrives early.
IBM lost $68 billion because it sold the wrong half of the AI stack.
Micron sells the half that got paid first, and until new capacity shows up, that is unlikely to change.
Related: Wall Street flees software plays for triple-digit chipmaker boom
