IBM Q2 2026: Software growth holds as mainframe drop and 4-5% guide cap rebound

International Business Machines Corporation (NYSE: IBM) reported second-quarter 2026 results on July 22 that confirmed the numbers behind its record-breaking July 14 pre-announcement and rebased full-year expectations. Revenue rose 1% to $17.2 billion, GAAP diluted earnings per share came in at $2.27, and operating (non-GAAP) diluted EPS reached $2.93, up 5% year over year. The company now expects full-year constant-currency revenue growth of 4% to 5%, a step down from the “at least 5%” trajectory the market had been modelling, while reiterating free cash flow will rise by about $1 billion year over year. The central tension for the stock is no longer whether the print matched consensus; it is whether the second-half deal conversion, mainframe cycle transition and $10.5 billion year-to-date acquisition spend can rebuild the confidence that vanished when IBM stock lost roughly a quarter of its value in a single session on July 14, the worst trading day in the company’s 110-year listed history.

What did the second-quarter print actually confirm after the record 25% July 14 collapse?

The July 22 release closed the loop on the preliminary numbers that Arvind Krishna, IBM chairman, president and chief executive officer, issued through an SEC filing on July 14. That filing landed a revenue miss of roughly $660 million against the $17.86 billion FactSet consensus and an operating EPS miss against the $3.01 to $3.02 range compiled by FactSet and LSEG. IBM stock fell about 25% intraday, exceeding the previous worst single-day decline of 23.7% recorded on October 19, 1987. By the time IBM reported the full quarter after the close on July 22, consensus had been reset. Revenue of $17.162 billion and operating EPS of $2.93 matched the pre-announced range and left the debate for the earnings call: whether the shortfall was, as management framed it, a timing issue in a discrete slice of client capital spending or a broader signal about enterprise software demand as clients redirect budgets toward AI infrastructure. Kavanaugh, in the release, described the quarter’s headwinds as late in the period, and emphasised that IBM continued to prioritise productivity, portfolio strengthening and free cash flow generation. That framing sets the tone for every metric investors will now scrutinise into the third quarter.

Why does the IBM Z segment’s 42% collapse matter more than the headline consulting flatness for the full-year setup?

The Infrastructure segment fell 7% to $3.835 billion, with Hybrid Infrastructure down 10%. Inside that number, IBM Z revenue declined 42%, a scale of drop that reflects the natural late-cycle position of the z16 platform ahead of the z17 ramp, but which nonetheless removed several hundred million dollars of high-margin revenue from a single quarter. Offsetting this was Distributed Infrastructure, which rose 37% at what IBM described as a record pace and now carries an order backlog of nearly $500 million across Power and Storage. Consulting was essentially flat at $5.327 billion, up 1% at constant currency, with both Strategy and Technology and Intelligent Operations reporting the same shape. The Consulting flatness is analytically important because it removes any offset from the segment that historically absorbed macro softness at IBM. However, the Infrastructure mix shift is the more consequential story for the second half. If the Distributed Infrastructure momentum sustains and the z17 refresh compounds it, the fiscal 2026 exit rate on Infrastructure could look materially different from the second-quarter print, even without a change in the software growth trajectory.

How does the Red Hat, Data and watsonx growth track compare with Transaction Processing’s continued erosion?

Software revenue of $7.761 billion grew 5%, or 4.6% at constant currency, and remained the largest single segment. Within that, Hybrid Cloud (Red Hat) grew 11%, Data grew 19% (18% at constant currency) and Automation grew 4% (3% at constant currency). Transaction Processing declined 8%, or 9% at constant currency, extending the pattern that has been visible for several quarters and reflects both the mainframe cycle timing and the gradual monetisation ceiling on legacy processing licences. Segment profit reached $2.502 billion at a margin of 32.2%, up from 31.1% a year earlier. The mix is the story. Red Hat, watsonx, HashiCorp and Confluent, the four names IBM highlighted as anchoring its high-growth portfolio, produced the growth. Transaction Processing continued to weigh. Management’s argument that clients paused software spending in the late weeks of the quarter to redirect budgets toward supply-constrained infrastructure is testable, and the third-quarter print will provide the first data point. If Red Hat holds its double-digit trajectory and Data sustains the high-teens growth rate, the Transaction Processing drag becomes a manageable overhang inside a segment growing 5%. If those two growth engines slow, the software thesis narrows sharply.

What does the revised 4% to 5% constant-currency full-year guidance signal about H2 deal conversion confidence?

IBM’s revised full-year outlook reads as calibrated rather than defensive. Constant-currency revenue growth of 4% to 5% implies second-half growth broadly consistent with the first-half exit rate rather than a dramatic H2 rebound. Currency is now expected to be neutral to reported growth for the year. Free cash flow guidance of approximately $1 billion of year-over-year improvement remains unchanged, and management flagged expected pre-tax income margin expansion for the full year. The message is that the Q2 shortfall was not, in management’s view, structural, but the revised guide does not attempt to make up all of the lost ground in the second half. That distinction matters. A guide that assumed full recovery would have set up a second binary event at the Q3 print. The revised guide instead sets a bar that can be cleared through steady conversion of the pipeline management referenced on the call as delayed rather than lost. J.P. Morgan reaffirmed its Overweight rating, most recently at a $250 price target, citing the improved full-year outlook. Baird retained a Neutral rating. RBC Capital reiterated Outperform at $270. Jefferies analyst Brent Thill maintained Buy while cutting the price target from $320 to $260, an 18.75% reduction that captures how the buy-side is now valuing the same growth algorithm at a lower multiple. Wedbush analyst Dan Ives kept his $350 target and framed the Q2 miss as an execution issue on deal timing, not a fundamental change in enterprise AI demand for IBM’s stack. The 25-analyst consensus target compiled by S&P Global sat at $248.77 on July 23.

How does $10.5 billion in year-to-date acquisitions square with a $6.3 billion cash drawdown and steady free cash flow?

Cash, restricted cash and marketable securities ended the second quarter at $8.2 billion, down $6.3 billion from year-end 2025. IBM invested $10.5 billion in acquisitions year to date, weighted toward the HashiCorp close, and total debt rose $0.7 billion to $62 billion, of which $13 billion sits inside IBM Financing. Free cash flow of $2.5 billion in the quarter was down $0.3 billion year over year; first-half free cash flow of $4.8 billion was flat. The company returned $1.6 billion to shareholders in dividends and left buybacks paused in the period. The board also declared a regular quarterly cash dividend of $1.69 per share, extending an unbroken quarterly payment record since 1916. The capital allocation question sits at the centre of the equity narrative. IBM has demonstrated that the software portfolio can absorb material acquisition capital and continue to generate approximately $1 billion of incremental annual free cash flow, but the cash draw and debt build have narrowed the balance sheet cushion. If HashiCorp, Red Hat and watsonx continue to compound at the current pace, the mathematics work. If growth in any of those three moderates, the capital allocation choice will look tighter than at any point since the Red Hat close in 2019.

Why are analysts holding overweight views despite Jefferies cutting its price target by $60?

The pattern of analyst commentary after July 22 was consistent. Ratings mostly held. Price targets came down. J.P. Morgan, RBC Capital and Wedbush kept their bullish stances, with Wedbush’s $350 target implying more than 65% upside from the July 23 traded range near $207. Jefferies, Baird and others cut targets while retaining constructive or neutral stances. That combination reflects the specific shape of the disappointment. The buy-side is treating the miss as a timing issue on deal conversion inside a portfolio that still has three of the four largest software franchises growing at double digits or high single digits. What has changed is the multiple investors are willing to pay while waiting for the mainframe cycle to inflect, the delayed software deals to close and management’s characterisation of the pipeline to be validated. IBM stock closed the July 22 session at approximately $213 and traded around $207 during the July 23 session, roughly 35% below the June 2026 all-time high of $329.23. The reset has, in analyst framing, pulled the risk-reward closer to symmetric, without any of them declaring the recovery complete.

What does the securities fraud investigation into IBM’s pre-announcement disclosures add to the risk picture?

Following the July 14 pre-announcement, an investigation was launched into whether IBM management characterisations of the pipeline in prior public statements were consistent with the deal conversion rate that materialised in the quarter. The company confirmed it is cooperating. The investigation has not produced formal findings, and Business News Today has not identified any regulatory determination on the matter. It nonetheless adds an editorial layer to every forward-looking statement management makes on subsequent calls. Investors will parse commentary about the second-half pipeline, quantum computing timelines and the pace of Red Hat, watsonx and Confluent growth against that backdrop. Krishna and Kavanaugh both used the July 22 release to emphasise that the shortfall was concentrated in a specific slice of the portfolio rather than a broad demand deterioration, and that the company continues to invest in specialised technical and client-facing talent, including Forward Deployed Engineers, to capture enterprise-scale AI deployment budgets. Whether that framing holds through the third quarter is the practical resolution point for both the investigation’s relevance and the guidance’s credibility.

What execution proof points would validate management’s second-half rebound narrative before the Q3 print?

Several measurable items sit between the July 22 release and the third-quarter print. Red Hat’s growth rate is the single most important line. A double-digit constant-currency figure preserves the software thesis. Data segment sustainability at anywhere near the current 18-19% growth extends it. Distributed Infrastructure’s ability to convert its near-$500 million backlog into revenue at scale would validate the Infrastructure mix shift. The IBM Z decline rate against the z17 refresh will indicate whether the mainframe cycle is now approaching its inflection. Free cash flow conversion tracking toward the reaffirmed $1 billion year-over-year improvement would confirm that pre-tax income margin expansion is holding despite the revenue miss. On the innovation side, Krishna reiterated that IBM will invest more than $10 billion in quantum computing over the next five years and remains on track to deliver its first large-scale fault-tolerant quantum computer by 2029, while the letter of intent with the U.S. Department of Commerce to build Anderon, described as the world’s first pure-play quantum wafer foundry, sits as a longer-dated optionality. Lightwell, a new capability that has already made more than 7,500 open source patches available in its first two weeks, gives IBM an incremental angle on the enterprise open source security market that leverages both IBM and Red Hat trust. None of these individually resolves the debate the July 14 pre-announcement created. Collectively, they represent the map by which the market will decide whether the July 22 print marked the trough of confidence or an intermediate stop.

Key takeaways from IBM’s Q2 2026 print, revised full-year guidance and post-crash execution setup

• International Business Machines Corporation (NYSE: IBM) reported Q2 2026 revenue of $17.162 billion, up 1% year over year, and operating (non-GAAP) diluted EPS of $2.93, up 5%, confirming the July 14 pre-announcement numbers that produced the worst single-session decline in the company’s 110-year listed history.
• Full-year 2026 constant-currency revenue growth guidance was rebased to 4% to 5%, below prior market expectations of at least 5%, while free cash flow is still expected to rise by approximately $1 billion year over year and pre-tax income margin expansion was reiterated.
• Software revenue rose 5% to $7.761 billion with Red Hat up 11%, Data up 19% and Automation up 4%, offset by Transaction Processing declining 8% and reflecting the continued mainframe-linked monetisation pattern.
• Infrastructure fell 7% to $3.835 billion with IBM Z down 42% ahead of the z17 refresh and Distributed Infrastructure up 37% at record pace, carrying an order backlog of nearly $500 million across Power and Storage.
• Consulting was essentially flat at $5.327 billion, removing the historical offset that segment has provided during periods of software softness and increasing the analytical weight of the second-half software conversion narrative.
• Cash and marketable securities ended the quarter at $8.2 billion, down $6.3 billion from year-end 2025, after $10.5 billion of year-to-date acquisition spend and $1.6 billion of Q2 dividends; total debt rose $0.7 billion to $62 billion and the board declared the September dividend at $1.69 per share, continuing a quarterly payment record since 1916.
• Analyst response was mixed but mostly constructive: J.P. Morgan Overweight at $250, RBC Capital Outperform at $270, Wedbush analyst Dan Ives at $350, Jefferies analyst Brent Thill cut price target from $320 to $260 while retaining Buy, and the 25-analyst S&P Global consensus target sat at $248.77 on July 23.
• A securities fraud investigation into IBM’s pipeline disclosures ahead of the July 14 pre-announcement remains active and adds scrutiny to every forward-looking management statement without altering current guidance credibility on its own.
• The measurable proof points into Q3 are Red Hat sustaining double-digit growth, Data holding high-teens growth, Distributed Infrastructure converting the near-$500 million backlog and the IBM Z decline rate moderating against the z17 refresh; delivery on those lines would validate the timing framing, while slippage on any two of them would reopen the debate the July 14 pre-announcement created.
• Quantum computing continues as a longer-dated optionality, with the $10 billion five-year investment commitment, the Anderon pure-play quantum wafer foundry letter of intent with the U.S. Department of Commerce, and the 2029 large-scale fault-tolerant quantum computer target sitting as strategic anchors well beyond the near-term guidance debate.