A $69 billion rout after customers redirected spending to AI hardware, delaying IBM's software and mainframe deals. The July 22 outlook will decide whether the miss was timing or something deeper.
IBM's board walked into an uncomfortable choice after learning the company's second quarter had fallen well short of internal forecasts.
Directors could wait for the scheduled earnings release and let executives explain the miss in detail. Or they could warn investors right away and accept a violent market reaction.
IBM (IBM) shares fell more than 25% on July 14, wiping out roughly $69 billion in market value in the company's worst one-day decline ever. The stock closed July 17 at $212.67, leaving IBM's market capitalization around $202.5 billion.
The preliminary data were disappointing. They were not bad enough on their own to justify a selloff of that size.
IBM expects second-quarter revenue of $17.2 billion, up 1% from a year earlier, with operating earnings of $2.93 a share. Wall Street analysts had penciled in roughly $17.86 billion in sales and $3.02 per share, according to LSEG data cited by Reuters.
The cause for the shortfall is what rattled investors.
Customers shifted spending toward servers, storage, and memory needed for artificial intelligence infrastructure. That hardware jumped to the front of corporate purchase queues, delaying large IBM software and mainframe-related deals.
That dynamic points to a risk that extends beyond one rough quarter.
Artificial intelligence can weaken IBM without replacing the company outright. It can do so by shrinking the technology budgets clients once allocated to traditional software, consulting, and mainframe systems.
"This quarter we faltered," CEO Arvind Krishna wrote in his letter to investors.
IBM entered the quarter with real momentum.
First-quarter revenue rose 9% to $15.9 billion. Software revenue climbed 11%, infrastructure jumped 15%, and IBM Z revenue surged 51%. Management kept its forecast for more than 5% constant-currency revenue growth in 2026 and roughly $1 billion of additional annual free cash flow.
The preliminary second-quarter numbers changed that story fast.
Software growth slowed to 5%. Consulting revenue was flat. Infrastructure revenue fell 7%, worse than management's earlier guidance for a low single-digit decline during the first IBM z17 mainframe launch cycle.
IBM said the infrastructure shortfall came from weak Z system performance and related transaction-processing software.
Large acquisitions had not failed, and newly bought businesses were not falling apart. Krishna said HashiCorp and Confluent performed well. Red Hat revenue growth accelerated to 11%.
Wholesale infrastructure demand for IBM was strong.
The distributed infrastructure business, which covers power systems and storage, grew 37% and ended the quarter with a backlog above $500 million. IBM said the z17 program was still running at roughly 130% of the comparable z16 cycle, despite the quarterly stumble.
The problem was that IBM could not convert enough of that demand into the sales mix and timing that would satisfy investors.
Customers moved late-quarter capital spending toward supply-constrained servers, storage, and memory ahead of expected price increases. Many large transactions did not close on schedule because the company did not react fast enough, IBM said.
Some of those deferred deals may still close in later quarters.
If they do, the second-quarter miss would look mainly like a timing problem, not a signal of lasting demand destruction. A company that markets itself as a trusted guide through complex technology transitions should understand how its biggest customers are allocating their budgets.
IBM's board reportedly pressed Krishna before deciding to release the early warning. The move might eventually help the company rebuild confidence through transparency. For investors, the choice initially read as a sign that management had lost visibility into its own sales pipeline.
That is how a small revenue miss sparked a massive stock-market reaction.
The quarter tested more than IBM's financial guidance. It raised doubts about whether the company could manage a shift from traditional enterprise computing to artificial intelligence with any reliability.
The first problem is budget competition.
Building AI systems requires processors, memory, storage, networking equipment, and data center capacity. When these components run short, or when prices look likely to rise, buyers may snap them up before approving less urgent software or consulting projects.
That appears to have happened in the closing weeks of IBM's second quarter.
Banks and other large enterprises focused on infrastructure purchases, throwing off the timing of software and mainframe deals IBM expected to close. IBM's warning ranks among the clearest signals yet that AI spending may squeeze other business technology expenditures, Reuters reported.
The 37% gain in distributed infrastructure shows customers bought some IBM servers and storage. That strength was not enough to offset weakness in the higher-margin mix of mainframes and transaction-processing software.
That reveals a weak spot in IBM's broad portfolio.
The company sells hardware, software, and consulting services that are designed to work together. That mix can deepen customer relationships when technology budgets expand.
Things get harder when clients have strong preferences for one category over another.
IBM can win a storage sale. It might still lose or delay the software and consulting revenue that was supposed to accompany it. Spending on AI infrastructure can stimulate demand in one IBM division while hurting another at the same time.
The second problem is product replacement.
AI-powered coding tools and automation technology could reduce what companies pay for traditional software development, application maintenance, and consulting services. Those tools could also modernize legacy systems without the same level of reliance on the vendors that built and maintain them.
IBM is especially exposed because its hybrid-cloud strategy sits across two computing eras.
Red Hat helps customers run applications on private systems and public clouds. IBM consultants help large enterprises update their technology while protecting the mission-critical software and data that still lives on mainframes.
That strategy depends on clients continuing to pay IBM to bridge their legacy operations with modern systems.
If companies need help securely implementing complex new systems, AI could make that bridge more valuable. It could also lower the bridge's value if automation makes migration easier, or if customers spend their available money elsewhere on infrastructure vendors, cloud platforms, and independent AI developers.
That is the hidden problem behind IBM's warning.
IBM is not just trying to sell artificial intelligence. It is competing for its customers' dollars with artificial intelligence.
The company has shown it can generate AI demand.
IBM announced a generative AI book of business above $12.5 billion as of the end of 2025. That definition mixes software transaction revenue with new annual contract value from software subscriptions and consulting contracts. It does not mean $12.5 billion of reported quarterly or annual AI revenue.
IBM still needs to convert those contracts and signings into sustainable growth that can offset weakness elsewhere.
The company is leaning into the software businesses and technology it has acquired, which hold the most significant revenue potential going forward.
Red Hat is still posting double-digit growth. Krishna said HashiCorp and Confluent performed well during the quarter. Those businesses strengthen IBM's position in hybrid cloud, infrastructure automation, and data management.
IBM is also placing a big bet on quantum computing.
The company says it plans to invest more than $10 billion over five years and expects to deploy a large-scale fault-tolerant quantum computer in 2029. IBM has also announced plans for a quantum wafer foundry, backed by $1 billion in proposed federal incentives and $1 billion in corporate contributions.
Those projects could become strategically important over time.
They will not fix the next couple of quarters.
Investors looking at IBM after the selloff face a question about whether the existing software business can carry long-term objectives while AI reshapes client spending in real time.
IBM did not revise its annual outlook in the preliminary-results letter. Management said it would disclose full-year estimates when it reports final second-quarter earnings on July 22.
Before the warning, IBM had projected more than 5% constant-currency revenue growth and $1 billion of incremental free cash flow for 2026. A large cut would suggest management believes the spending disruption goes beyond a one-quarter timing issue.
A second question is whether the delayed deals will actually close.
Krishna said many significant transactions did not close within IBM's expected schedule. Investors will have to decide whether those clients simply postponed or whether they redirected the money permanently.
A delayed sale can boost revenue over several quarters.
A cancelled project would mean AI has shifted customer priorities in a more fundamental way.
The z17 was still ahead of the z16 program on a comparable basis. The division had not yet delivered the quarterly results management expected. Investors should get more clarity on shipments, capacity expansion, and revenue from the transaction-processing software that accompanies those systems.
A fourth issue is the quality of software growth.
Red Hat's 11% growth, along with HashiCorp and Confluent's performance, shows IBM's modernization portfolio is real. Shareholders need to separate organic performance from growth driven by acquisitions and see whether new products can compensate for decline in mature software categories.
IBM generated $4.8 billion in free cash flow during the first half. Debt stood at $66.4 billion at the end of the first quarter, up $5.1 billion from year-end as the company financed the Confluent purchase.
That balance sheet does not signal an imminent crisis.
It does mean IBM has less room for repeated execution mistakes as it integrates acquisitions, maintains its dividend, and funds expensive AI and quantum computing programs.
IBM's second-quarter revenue and adjusted earnings both came in below Wall Street expectations.
The stock's 25% drop erased roughly $69 billion in market value, the largest one-day decline on record for the company.
Customers redirected late-quarter spending toward scarce servers, storage, and memory, which delayed large IBM transactions.
Red Hat, HashiCorp, Confluent, and distributed infrastructure provided meaningful bright spots.
AI could pressure IBM from two sides: consuming technology budgets and automating work that traditional software and consulting vendors used to do.
IBM's July 22 outlook will help determine whether the miss was a temporary blip or evidence of a deeper transformation problem.
IBM's historic decline does not mean the company has become irrelevant.
Its biggest customers still depend on IBM systems for critical financial, governmental, and industrial workloads. Red Hat is still growing. Demand for distributed infrastructure is strong. The z17 program remains ahead of its predecessor on a comparable basis.
The warning changes the burden of proof.
Krishna has positioned IBM as the link between old corporate systems and the next generation of cloud computing, artificial intelligence, and eventually quantum computing.
That middle position looked attractive when customers were increasing multiple technology budgets at the same time.
The risk grows when AI forces customers to choose which investments to fund first.
IBM may find that the infrastructure required for AI consumes the budget before its software and consulting businesses can monetize the transition. At the same time, more advanced AI tools could put long-term pressure on the older software and services that are funding IBM's transformation.
A recovery case exists. If the delayed transactions close, Red Hat keeps growing at double-digit rates, and the company holds its full-year cash-flow target, then the size of the selloff could look extreme relative to a temporary disruption in buying patterns.
If IBM cuts yearly projections, transaction-processing demand keeps weakening, or major customers build AI capabilities on their own without buying the broader IBM portfolio, the bear case gets stronger.
The question is not whether IBM is involved in artificial intelligence.
The question is whether the company can generate enough AI-related revenue before the technology eats into the businesses that are funding IBM's future.
This story was originally published by TheStreet on Jul 19, 2026, where it first appeared in the Investing section.
Prepared with AlphaScala editorial tooling from the source reporting linked above. Indexable analysis may include a cited Alpha Score value. Publishing checks screen each story before release. Educational coverage, not personalized advice.