IBM’s second-quarter results place a sharper operational burden on its AI strategy. The company is not presenting artificial intelligence solely as a future revenue category through watsonx, Red Hat and consulting services. It is also positioning AI and automation as mechanisms to improve how it develops software, sells, markets and manages its supply chain.
That distinction matters after a quarter in which IBM produced modest overall revenue growth but fell short of its own expectations in important areas. Management’s response is therefore less a declaration that AI experimentation has ended than a demand that its technology investments now yield measurable improvements in productivity, profitability and sales execution.
A revised growth target and a more explicit internal mandate
IBM reported second-quarter revenue of $17.2 billion, up 1% year on year. Software revenue rose 5%, while consulting was broadly flat and infrastructure revenue declined 7%. The company reported operating, non-GAAP earnings per share of $2.93, up 5%, while operating pre-tax income margin improved slightly to 19.2%.
For the full year, IBM now expects constant-currency revenue growth of 4% to 5% and continues to expect free cash flow to increase by roughly $1 billion from 2025. The revenue range is a reduction from the more-than-5% constant-currency growth outlook it retained after the first quarter. At the same time, IBM said it expects improved pre-tax margin expansion for the year.
The combination gives AI a dual role in the company’s plan. First, IBM wants to sell technology and services that help enterprises build, deploy, manage and secure AI applications across hybrid environments. Second, it intends to use AI internally to reduce the cost and time involved in delivering those offerings.
IBM identified three internal areas for accelerated productivity: software development, sales and marketing effectiveness, and supply-chain optimisation. These are broad categories rather than quantified programmes, but their inclusion in an earnings release is significant. They link AI adoption directly to financial objectives such as margin expansion and free-cash-flow growth, rather than treating it as a separate innovation initiative.
The quarter exposed the importance of execution
The urgency reflects an uneven second quarter. In a letter issued before the full results, chief executive Arvind Krishna said several large transactions did not close within the expected timeframe. He also said clients had shifted spending towards servers, storage and memory, partly to secure supply-constrained infrastructure before anticipated price increases. Cybersecurity concerns added to the disruption in buying patterns.
Those factors affected IBM’s software and infrastructure performance, particularly its IBM Z mainframe business and the related Transaction Processing software category. IBM Z revenue fell 42% in the quarter, reflecting a difficult comparison after the early cycle of the z17 platform launch. Transaction Processing revenue declined 8%.
The contrast within infrastructure illustrates why IBM’s productivity programme cannot substitute for commercial responsiveness. Distributed Infrastructure revenue grew 37%, led by Power and Storage, and the company exited the quarter with an order backlog of nearly $500 million in that area. However, strength in the products customers were prioritising did not fully offset weakness in the mainframe-related business and delayed deals.
Automation can shorten internal workflows, improve sales targeting and help teams respond faster to changing customer demand. It cannot, by itself, ensure that large enterprise projects are approved or that client technology budgets are allocated to software rather than hardware. IBM’s ability to translate the programme into results will therefore depend on whether its operating changes improve both efficiency and deal conversion.
Portfolio breadth remains central to the strategy
IBM is relying on portfolio breadth to address enterprise AI demand. In software, Red Hat grew 11%, while Data revenue increased 19%. IBM also cited continued contributions from HashiCorp and Confluent, acquisitions intended to expand its capabilities in hybrid cloud infrastructure, automation and data management.
This portfolio logic is important because enterprise AI deployments rarely depend on a single model or application. Companies must integrate data, secure infrastructure, govern software components, run workloads across cloud and on-premises systems, and adapt existing business processes. IBM’s strategic argument is that it can provide technology and services across this stack, including through Red Hat’s open hybrid-cloud platform and its consulting organisation.
Its new commercial initiatives point in the same direction. IBM has highlighted Lightwell, an automated vulnerability-remediation offering built with Red Hat, as an example of a faster path from internal innovation to a market-facing service. The company also continues to invest in longer-horizon opportunities, including a planned investment of more than $10 billion in quantum computing over five years.
Yet the second-quarter results suggest that near-term credibility will rest more on established businesses than on distant technology bets. Red Hat growth, data software, automation, distributed infrastructure and consulting signings are more immediate indicators of whether IBM can convert AI demand into recurring revenue.
From promise to metrics
IBM’s internal use of AI should be assessed through a practical set of measures. Investors and customers will look for sustained margin improvement without a deterioration in service quality, evidence that software development becomes faster or more productive, and signs that sales execution improves when enterprise purchasing conditions change. They will also watch whether revenue growth strengthens across software and consulting rather than relying on selective infrastructure demand.
The company has retained its free-cash-flow objective and expects better margin expansion, which gives management clear financial benchmarks. But it has not disclosed a standalone productivity target, a projected cost-saving figure or a timetable for the operational changes. That leaves room for IBM to demonstrate progress through subsequent quarterly performance rather than promises.
The broader lesson is that enterprise AI strategies are entering a less forgiving phase. The competitive question is no longer simply whether a company has AI products, partnerships or research programmes. It is whether those assets improve customers’ operations and the company’s own ability to develop, sell and deliver at scale.
For IBM, the second quarter made that transition more concrete. AI and automation are now being asked to support innovation, protect margins and make commercial execution more resilient. The company has a diversified portfolio and several areas of growth, but the revised outlook shows that productivity gains must be accompanied by more consistent demand capture if the strategy is to meet its financial ambitions.
