2026-07-16

IBM’s 25% stock crash: what happened on July 14, 2026

By FinScope Research · Data as of 2026-07-16

IBM closed the July 14, 2026 regular session at $217.07. That was $73.16 below its previous close of $290.23, a one-day loss of 25.2%. The stock touched $213.22 intraday, down 26.5% from the prior close, per market reporting.

In FinScope’s daily IBM series, which begins on 1962-01-02, that makes July 14 the worst one-day drop on record. It surpassed IBM’s 23.5% loss on Black Monday in 1987 by roughly 1.7 percentage points. CNBC likewise reported it as IBM’s worst trading day ever.

The immediate trigger was a preliminary second-quarter warning released before IBM’s scheduled earnings report. But the market was reacting to more than a modest revenue and earnings miss. Delayed large deals and management’s admission that it failed to adapt quickly made investors question the assumptions behind IBM’s outlook.

The preliminary Q2 numbers

IBM released selected preliminary results in CEO Arvind Krishna’s letter to investors on the morning of July 14. The company is still closing its books, so the final figures may differ slightly.

Metric	IBM preliminary result	Consensus	Difference
Revenue	$17.20B	$17.86B	-$0.66B, about -3.7%
Adjusted EPS	$2.93	$3.02	-$0.09, about -3.0%
Total revenue growth	+1%	—	Grew, but missed expectations
Software revenue	+5%	—	Growth
Consulting revenue	0%	—	Flat
Infrastructure revenue	-7%	—	Decline

Consensus figures are LSEG estimates cited by Reuters. Revenue and adjusted EPS were both higher than a year earlier — just not as high as analysts expected. Infrastructure fell 7%, and software ran weaker than IBM’s own plan.

The surprise mattered as much as the miss. IBM’s full earnings release was scheduled for July 22. Management chose to acknowledge the shortfall more than a week early, which damaged confidence in the previous outlook.

How the AI boom hurt IBM this quarter

IBM’s explanation was counterintuitive. Customers did not stop spending on AI — they redirected quarterly capital budgets toward servers, storage, and memory to secure supply ahead of expected shortages and price increases. That left less room for IBM mainframes and related software, particularly transaction processing.

Krishna also said industry-wide cybersecurity concerns slowed client decisions. Several large deals failed to close on IBM’s expected schedule, and his letter used unusually direct language: “this quarter we faltered.” He acknowledged that IBM did not adapt or move quickly enough — not merely that the environment turned against it.

That admission carried weight. A delayed deal can close in a later quarter. A lasting shift of enterprise IT budgets toward AI hardware would be more serious, because it would force investors to lower their assumptions for IBM’s software and consulting growth. The 25% selloff reflects how hard it is to tell those two cases apart before the full report.

The five worst IBM trading days since 1962

The ranking below is computed from FinScope’s adjusted daily USD closes for IBM, which now include the July 14 session. The dataset begins on 1962-01-02.

Rank	Date	One-day return
1	2026-07-14	-25.2%
2	1987-10-19	-23.5%
3	2000-10-18	-15.5%
4	1999-10-21	-15.0%
5	2026-02-23	-13.1%

July 14 was worse than IBM’s Black Monday loss — and unlike 1987, the trigger was one company’s warning rather than a market-wide crash. In the observed FinScope history it ranks first outright, not merely among the largest.

A record daily loss is not the same as maximum drawdown

IBM’s worst observed maximum drawdown remains -69.4%, spanning the 1987-08-20 peak to the 1993-08-16 trough. July 14 was the worst single day, but not the deepest full drawdown cycle.

At the July 14 close, IBM stood about 34.1% below its $329.23 all-time high from 2026-06-02. As of the July 16 close ($219.05), that gap was 33.5%. And recovering the one-day loss alone requires a 33.7% gain from $217.07 to return to the previous $290.23 close — a 25.2% rebound would not be enough, because the gain applies to a smaller base.

That distinction separates a dramatic daily return from the longer loss an investor actually has to sit through. See What is maximum drawdown? and What is recovery time? for the definitions.

What to watch on July 22

IBM plans to report full second-quarter results and discuss its full-year outlook on July 22. The main questions go beyond a few cents of EPS.

First: were the missed large deals delayed, canceled, or reduced? The composition of the 5% software growth and the weakness in transaction processing will matter. So will management’s explanation of whether the 7% infrastructure decline is a temporary z17 product-cycle effect or evidence of a longer budget shift. Any change to full-year revenue or free-cash-flow guidance could reshape how this selloff gets interpreted.

A lower price does not automatically make IBM cheap. Equally, a 25% market-value loss does not mean the underlying business deteriorated by exactly 25% in a day. The fuller July 22 report should narrow the gap between those two readings.

Data basis and limitations

All price figures — the one-day return, the historical ranking, the drawdown from peak, and the required rebound — are computed from FinScope’s recorded adjusted daily series for IBM, which now includes the July 14–16 sessions. The intraday low is from market reporting, not the daily series.

IBM’s Q2 figures are preliminary; the company stated that final numbers may differ. This article covers what was known immediately after the July 14 selloff, and the July 22 report may change some figures and the outlook.

How to view it in FinScope

Open the IBM page to see the current decline from peak, maximum drawdown, trough date, and recovery status, updated daily. For the concepts, read How to read underwater periods, What is recovery time?, and How FX changes drawdowns.

Sources

Methodology: one-day returns and drawdowns use adjusted daily closes; drawdowns are peak-to-trough declines from the running all-time high. See the methodology page.