IBM expects second-quarter revenue of $17.2 billion and adjusted earnings per share of $2.93, below consensus estimates of $17.86 billion and $3.01. The revenue gap is $660 million, while adjusted EPS is $0.08 below the estimate.
The numbers are not enormous in isolation. What changed the reading was IBM’s preliminary disclosure before its scheduled July 22 earnings report, coupled with management’s statement that numerous large deals did not close during the quarter.
That leaves investors with a more useful question than whether IBM missed by eight cents: were those contracts delayed by a late-quarter reshuffling of IT budgets, or did customers decide some projects no longer deserve immediate spending? IBM’s explanation points toward the first possibility, but the answer will determine whether this is a timing issue at one company or an early warning for enterprise technology more broadly.
A small miss with a larger implication
Large enterprise contracts can have an outsized effect on a quarter because deals often close near period-end. A handful of delayed decisions can move meaningful revenue from one reporting period to the next without necessarily changing the customer relationship or the long-term project.
That is the charitable interpretation of IBM’s preliminary result. If the unclosed deals remain in its pipeline and close later, the second-quarter shortfall would mainly reflect timing. IBM would still have lost the revenue and profit contribution from the quarter, but the underlying demand would be intact.
The less comfortable interpretation is that late June exposed a real budget constraint. Companies may be willing to fund some technology projects while delaying others, especially when they face supply concerns around infrastructure or rising costs for equipment needed to support AI workloads. In that case, a missed quarter is not simply an accounting-calendar nuisance. It can signal that the order of corporate IT priorities has changed.
IBM has not yet established which explanation is correct. Its July 22 call will need to address the status of the delayed contracts, whether customers have merely pushed decisions into a later period, and whether the late-June pattern continued into July.
Infrastructure spending moved to the front of the line
IBM said clients shifted late-June capital spending toward supply-constrained servers, storage, and memory ahead of expected price increases. Cybersecurity concerns also diverted attention and budgets. Those explanations describe a purchasing queue, not a disappearance of technology spending.
When hardware capacity is scarce or expected to become more expensive, it can become the urgent purchase. A chief information officer can postpone a consulting project or delay a software commitment more easily than an equipment order tied to capacity needs, supply availability, or a planned AI deployment. Cybersecurity can create a similar urgency when management attention shifts toward a perceived operational threat.
The result is a potentially awkward period for vendors serving enterprise budgets. Customers may still be spending heavily on technology, but the dollars are directed toward the projects that cannot wait. That can make aggregate IT spending look healthier than the results of a software provider, consultant, or company dependent on a few large deal closures.
IBM’s preliminary segment trends fit that uneven picture. Software revenue rose 5%, consulting was flat, and infrastructure revenue fell 7%. The figures do not prove that AI infrastructure spending caused the infrastructure decline, nor do they establish a broad enterprise-software slowdown. They do show that IBM did not experience a uniform drop in demand across its businesses.
Software growth is especially relevant because it limits the claim that customers broadly stopped buying enterprise software. Yet growth in one segment does not erase the concern around delayed large deals. A business can post positive software growth while still losing higher-value or late-quarter transactions that investors expected to close.
The revenue mix also deserves attention. IBM’s disclosed pattern suggests that spending pressure may land unevenly on businesses that customers can defer. Research, modernization, consulting work, and major platform decisions often involve longer approval processes than an urgent capacity purchase. A delay can therefore weaken a quarter even when the customer has not abandoned the broader project.
IBM’s preliminary explanation also mentions mainframe purchases as an area that could be deferred. That matters because the company’s infrastructure business declined 7%, and the strongest company-specific countercase is a worse-than-expected mainframe cycle rather than a broad reset in software demand. Investors should resist treating the infrastructure result as a clean measure of AI spending pressure on all enterprise technology.
The software read-through is real, but unconfirmed
IBM’s warning created a market read-through for Microsoft, Salesforce, and ServiceNow, whose shares declined following the preliminary disclosure. That reaction appears to reflect concern that enterprise customers are becoming more selective about where AI-related budgets go. It is not evidence that demand at those companies has weakened.
The distinction is more than semantic. IBM has a business mix that includes mainframes and infrastructure, while Microsoft, Salesforce, and ServiceNow have different revenue models and customer exposures. IBM’s experience can identify a risk worth testing, but it cannot settle the question for companies with different products, contract structures, and sources of growth.
Microsoft has exposure to enterprise software and cloud services. Salesforce and ServiceNow are more closely associated with enterprise software platforms. Their results will carry greater weight than IBM’s warning because they can show whether customers are actually deferring software commitments, or whether IBM ran into a narrower combination of mainframe-cycle weakness and execution problems around large deals.
The market may be reacting to the possibility of a spending trade-off. If a company redirects funds toward servers, storage, and memory before expected price increases, that money remains inside the technology budget. But it may leave less room in the same planning cycle for consulting assignments, software expansion, or a large contract that requires executive approval before quarter-end.
That is a more targeted concern than a generic “enterprise IT is weak” thesis. It suggests that vendors with discretionary projects or a reliance on large, late-quarter transactions could face more scrutiny. It also suggests that suppliers of constrained hardware may benefit if customers continue accelerating purchases to secure capacity or avoid higher costs.
Still, the available evidence does not identify specific infrastructure winners, and it does not show that all software spending is under pressure. IBM’s own software revenue rose 5%. The relevant signal is not whether enterprise customers spend on technology at all. It is whether they increasingly fund infrastructure and cybersecurity first, then postpone projects that can wait.
The countercase is company-specific weakness
The most credible challenge to the budget-squeeze thesis is that IBM’s shortfall says primarily about IBM. Numerous large deals failed to close, infrastructure revenue fell 7%, and the company may be dealing with a worse-than-expected mainframe cycle. Those factors could explain much of the gap between IBM’s preliminary revenue expectation of $17.2 billion and the $17.86 billion consensus estimate without demonstrating a wider decline in enterprise software demand.
This countercase has support in the segment data. Software rose 5%, while consulting was flat and infrastructure declined. If pressure was concentrated in infrastructure and mainframe-related purchasing, IBM would offer a limited read-through for companies whose results rely more heavily on recurring software commitments or cloud demand.
There is also a mechanical reason not to overreact to one quarter. Large contracts that miss a reporting deadline can be delayed for administrative, budgetary, or approval reasons. IBM’s preliminary statement identifies failed quarter-end closures, but the research does not establish that customers cancelled the deals or reduced their long-term technology plans.
What would weaken the company-specific explanation is evidence that delayed deals are no longer active, that the budget reallocation continued after June, or that IBM sees pressure spreading beyond the areas tied to its infrastructure and mainframe exposure. The conclusion would also become more serious if Microsoft, Salesforce, or ServiceNow describe comparable delays in their own enterprise spending patterns.
Conversely, confirmation that IBM’s large deals remain in the pipeline, along with continued software growth and a limited impact from the late-June shift, would support the view that the quarter was disrupted rather than structurally impaired. That would not make the revenue shortfall disappear, but it would reduce the case for treating it as a sector-wide demand warning.
July 22 is the first real test
IBM’s July 22 earnings call needs to clarify several points that the preliminary figures cannot answer. Management’s view of the delayed deals matters most: whether they remain in the pipeline, what timing IBM expects, and whether the missed closures reflected customer deferrals or lost business.
Investors will also need a clearer segment outlook. Software demand should be separated from consulting and infrastructure trends, because the preliminary results were already uneven. A general statement about IT budgets would be less useful than an explanation of where spending shifted, which projects customers delayed, and whether mainframe weakness was a central driver.
Finally, management’s assessment of July will help distinguish a short-lived purchasing surge from a continuing budget reallocation. If the late-June rush toward servers, storage, and memory faded after immediate supply and pricing concerns passed, IBM’s second quarter may look like an unusually bad deadline problem. If customers continued directing capital toward AI infrastructure and cybersecurity, the market will have stronger grounds to reassess how much room remains for discretionary software, consulting, and mainframe spending.
