A software contract completed after a quarter-end can make the next reporting period look healthier without changing the customer’s longer-term spending plan. That distinction sits at the center of IBM’s third-quarter recovery: contracts are arriving, but their timing complicates the evidence for a durable acceleration.
The company’s October 21 results show improvement after a difficult second quarter. The useful question is how much of that improvement can repeat once the catch-up work has passed through the accounts. It requires separating contract execution, product mix and cash allocation, rather than treating an earnings beat as a single verdict on enterprise technology demand.
A late signature can strengthen one quarter without changing the run rate
IBM reported revenue of $17.1 billion and software revenue of $7.8 billion, with software up 8% from a year earlier, rounded as presented in its earnings release. In a Reuters interview, finance chief James Kavanaugh said delayed software contracts were beginning to close and raised the expected annual software growth range to about 7–8%, from 6–8%.
There is substance in recovering a delayed sale. It suggests that at least some customers deferred commitments rather than abandoned them. But a contract moved from June into September helps the comparison between those quarters twice: it weakens the first and strengthens the second. That timing effect cannot be assumed to recur every period.
Reuters’ subsequent analysis cites BofA analysts estimating that about $250 million of the software improvement came from deals delayed from the second quarter. That is an attributed analyst assessment, not a separately audited measure of organic growth. It supports caution about extrapolation without establishing that the remainder of the business is weak.
The prior disruption also had an execution component. AP’s July reporting described management’s explanation that customers were shifting spending toward hardware and that large deals had missed expected closing dates. A recovery can therefore reflect better selling execution, a changing budget mix, or both. These explanations have different implications for the next quarter’s baseline.
The software portfolio is moving at several speeds
IBM’s release reports Red Hat growth of 11%, data growth of 19% and an automation decline of 3%. Transaction processing rose 9%. These are reported growth rates, rather than a mixture of reported and constant-currency figures.
The divergence makes the portfolio more informative than a broad claim that software has recovered. Renewal-heavy products, new data capabilities and large enterprise agreements need not move together. A customer may complete an essential transaction-processing purchase while postponing another project. Equally, weakness in one product category does not invalidate demand elsewhere.
Hardware adds a different cycle: IBM Z revenue fell 20%, while distributed infrastructure grew 31%, according to the same release. A mainframe installation and the software attached to its operation can be recognized on different schedules. It would therefore be misleading to combine these changes into an undifferentiated measure of AI enthusiasm or enterprise caution.
For the business model, the important inference is that sales quality depends on which products are expanding and how customers are committing. A faster-growing category can support future cross-selling, but that becomes an economic benefit only if adoption, retention and the cost of serving customers support it. The quarterly growth table alone cannot establish those outcomes.
Free cash flow does not settle the acquisition bill
For the first nine months, IBM reported $11.8 billion of operating cash flow and $7.8 billion of its non-GAAP free cash flow, alongside $11.4 billion invested in acquisitions. Its published reconciliation adjusts operating cash flow for IBM Financing receivables and deducts net capital expenditure. It does not deduct acquisition spending.
These measures answer different questions. Free cash flow helps assess cash generated by the operating business under management’s definition. Acquisitions describe another use of capital intended to enlarge that business. A positive free-cash-flow figure does not mean acquisitions have already paid for themselves or that every use of cash is covered within the same period.
This distinction is particularly relevant when purchases help shape the software portfolio. Acquired revenue can be commercially valuable, yet the purchase price and integration burden still matter for the return on capital. Conversely, subtracting an entire acquisition outlay and calling the remainder a recurring operating result would also confuse investment decisions with current trading performance.
The durability test sits beyond the catch-up quarter
The strongest favorable interpretation is that late contracts validate customer demand while stronger product lines provide continuing growth. A more cautious interpretation is that catch-up temporarily flatters a mixed portfolio. Both fit parts of the disclosed evidence.
IBM kept its overall annual constant-currency revenue growth outlook at 4–5%, while lifting its cash-flow expectation. Those remain management forecasts. The unchanged group range also limits the claim that every business has entered a uniformly faster growth phase.
Subsequent disclosure can distinguish the interpretations: whether new contracts replace the recovered ones, whether weaker categories stabilize, and whether cash generation supports integration and distributions without repeated financing strain. The recovery is real in the reported quarter. Its persistence is a separate proposition that still has to earn support.