On July 14, 2026, IBM stock fell 25.2% after the company released disappointing preliminary second-quarter results. AP reported that it was IBM's worst day since at least 1972, according to FactSet. The broader market rose: the S&P 500 gained 0.4% and the Nasdaq added 0.9%.
That contrast matters. This was not a broad market selloff that happened to pull IBM down. Investors were reacting to an IBM-specific warning about revenue, customer budgets, and deals that did not close when expected.
What IBM reported
IBM said preliminary second-quarter revenue was $17.2 billion, up 1% from a year earlier. Adjusted earnings per share were $2.93, up 5%. Growth alone was not enough because both results fell below Wall Street's expectations.
| Second-quarter measure | IBM preliminary result | FactSet estimate | Difference |
|---|---|---|---|
| Revenue | $17.20 billion | $17.86 billion | -$660 million |
| Adjusted earnings per share | $2.93 | $3.01 | -$0.08 |
| Software revenue growth | +5% | — | — |
| Infrastructure revenue growth | -7% | — | — |
The revenue miss was about 3.7% of the consensus estimate. Yet the stock lost 25.2%. Those percentages measure different things. One compares a quarter's sales with an estimate; the other reprices the stream of money investors expect IBM to earn in the future.
Was AI the problem?
AI demand was part of the chain, but not because customers suddenly stopped spending on technology.
IBM said that in the final weeks of June, customers moved quarterly capital budgets toward servers, storage, and memory. They wanted to secure supply-constrained equipment before expected price increases. AP connected those expected increases to the AI boom.
For a customer with a fixed quarterly budget, urgent hardware purchases can leave less money or less attention for software, mainframe-related products, and consulting deals. IBM said its Z mainframe performance and the related software stack, especially Transaction Processing, came in below its expectations. Rapidly changing cybersecurity concerns also distracted customers.
The result was a timing and execution problem. IBM said numerous large deals failed to close on the schedules it had expected and that this drove most of the shortfall. A delayed deal may still become revenue later, but investors now have to ask how much was merely delayed, how much may be lost, and whether the same budget pressure can continue.
Why the stock reaction was so much larger
A $660 million revenue miss does not mechanically erase 25.2% of a company's value. The market reaction also reflected confidence.
IBM had expected Infrastructure revenue to decline by a low-single-digit percentage for the year as its z17 launch cycle matured. Instead, total Infrastructure revenue fell 7% in the quarter, and management said the outcome was worse than expected. It also acknowledged that the company did not adapt quickly enough.
There were stronger areas. Red Hat revenue growth accelerated to 11%, and Distributed Infrastructure grew 37% with an approximately $500 million backlog. That performance shows that the entire business did not weaken at once. But it did not answer the immediate question: can IBM turn strong demand in selected products into the total revenue and software growth investors had priced in?
The 25.2% fall suggests the market was not only marking down one quarter. It was raising the risk attached to future deal timing, mainframe-linked software sales, and management execution. That is an inference from the size of the price move, not a forecast of IBM's next quarter.
What happened to $1,000?
A 25.2% decline leaves 74.8% of the starting money.
| Calculation | Amount or return |
|---|---|
| Starting investment | $1,000 |
| Loss after a 25.2% drop | $252 |
| Money remaining | $748 |
| Gain needed to return to $1,000 | about +33.7% |
A $1,000 IBM position before the drop would be worth $748 after it. Returning from $748 to $1,000 requires a gain of about 33.7%, not 25.2%, because the recovery starts from a smaller amount.
This calculation does not predict that IBM will recover. It only shows the return required to reverse the one-day loss.
What remains unknown
IBM released selected preliminary results and said the final results could differ slightly. The regular earnings call is scheduled for July 22. The most important clarification will be whether delayed large deals are still expected to close, how the spending shift affects the full-year outlook, and whether weakness in Z and Transaction Processing lasts beyond one quarter.
The selloff is therefore best read as two events at once: a measurable earnings miss and a larger loss of confidence in when IBM will collect the money it expected. AI spending did not disappear. It moved, and IBM was caught on the wrong side of that move during the quarter.
This article explains the event and the loss calculation. It is not a recommendation to buy or sell IBM stock.