Product1 distinct publisher3 min readUpdated
Preliminary second-quarter results show software up 5 percent, infrastructure down 7 percent and consulting flat, with Arvind Krishna citing capex reprioritization and deals that did not close.
The Product Desk · Product desk

Compiled by The Product DeskSomething wrong?How this is made
IBM released selected preliminary second-quarter 2026 results on July 14 alongside a letter from Arvind Krishna explaining what he called a Software and Infrastructure performance shortfall, and warned that final figures could be slightly different once reporting closes [1]. Revenue was $17.2 billion, up 1 percent, with software up 5 percent, consulting flat (up 1 percent at constant currency), and infrastructure down 7 percent [2][3][4][5].
The interesting part is the mechanism IBM describes. In the last few weeks of June, according to Krishna, clients shifted quarterly capex toward servers, storage and memory to secure supply-constrained infrastructure ahead of expected price increases, which changed buying patterns [6]. IBM expected some supply chain effect but not the magnitude of the reprioritization [7], and Krishna adds that clients were distracted by rapidly evolving, industry-wide cybersecurity concerns [8]. That is a vendor describing a fixed budget being resequenced rather than a budget being cut, and IBM was on both sides of it: Distributed Infrastructure posted what the letter calls its best performance in reported history, up 37 percent with growth in Power and Storage and roughly $500 million of backlog exiting the quarter [9], while the infrastructure segment overall fell 7 percent [5][14].
Where the money did not go was the mainframe and everything attached to it. Krishna attributes the miss to Z performance and the associated software stack, primarily Transaction Processing [10]. In April, IBM had told investors z17 would wrap in the second quarter and that infrastructure would decline low single digits for the year starting then, off what it described as the strongest start to a mainframe program in its history [11]; the reported quarterly decline was 7 percent [5][13]. IBM still says z17 is at nearly 130 percent program-to-program versus z16, with clients representing 85 percent of installed MIPs maintaining or growing capacity [12]. The growth inside software came from elsewhere: Red Hat accelerated sequentially to 11 percent, and HashiCorp and Confluent are cited as strong performers [15][16].
Krishna does not rest the explanation on buyer behaviour. He writes that the conditions required perfect execution and that IBM faltered, did not move quickly enough, and that numerous large deals failing to close on expected timelines drove the majority of the shortfall [17]. That is a timing claim, and it is testable in ninety days.
The cost line held up better than the top line. GAAP gross margin was 57.7 percent, down 100 basis points, and operating gross margin 59.4 percent, down 70 basis points [18]; GAAP pre-tax margin fell 90 basis points to 14.4 percent while operating pre-tax margin rose 30 basis points to 19.2 percent, helped by productivity initiatives [19][20]. GAAP EPS of $2.27 was down 2 percent against operating EPS of $2.93, up 5 percent, a seven-point spread [21][22]. Year to date, operating cash flow was $7.8 billion and free cash flow $4.8 billion, about 62 percent conversion [23][24].
The forward-looking items in the letter are commitments rather than revenue: Lightwell, a $5 billion program with more than 20,000 engineers acting as an enterprise clearinghouse for open source vulnerabilities, reached general availability on July 8 with early adopters including JPMorganChase, Citi, Goldman Sachs, Visa and Mastercard [25], and IBM announced a letter of intent with the U.S. Department of Commerce to build a quantum system called Anderon [26].
Watch three things next quarter: whether the deferred large deals close and Transaction Processing recovers, whether the roughly $500 million Distributed Infrastructure backlog converts [9], and whether the June rush into supply-constrained hardware repeats once the expected price increases land [6].
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
When IBM discussed expectations with investors in April, it noted it would be wrapping on the launch of z17 in the second quarter and, given the strongest start to a mainframe program in its history, expected Infrastructure revenue to decline low-single digits for the year beginning that quarter.
After the introduction of Mythos, IBM and Red Hat teams launched Lightwell, a $5 billion commitment backed by new frontier AI capabilities and more than 20,000 engineers, creating a trusted enterprise clearinghouse to address open source software vulnerabilities; general availability was announced on July 8 and early adopters include Bank of America, BNY, Citi, Goldman Sachs, JPMorganChase, Mastercard, Morgan Stanley, Royal Bank of Canada, State Street, Visa and Wells Fargo.
On July 14, 2026, IBM released selected preliminary second-quarter 2026 financial results with a letter from Arvind Krishna to IBM investors, noting the company was still working to close its financial reporting and that final results could be slightly different; Krishna said he wanted to explain the Software and Infrastructure performance shortfall.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
First-party and specific, but preliminary and uncorroborated
The single source is the primary document itself: an IBM newsroom letter with named, quantified financials, which is authoritative for what the company reported. Two things cap the score. The figures are explicitly preliminary and pre-close, and the causal explanation for the miss, the strength bullets and the Lightwell and quantum commitments are all company assertions with no third-party corroboration, no segment-level breakout of Z or Transaction Processing revenue, and no GAAP-to-non-GAAP reconciliation in the supplied material.
Real customer activity, all self-reported
There are concrete uptake datapoints rather than vapor: a 37 percent Distributed Infrastructure quarter with roughly $500 million of backlog, z17 at nearly 130 percent program-to-program with 85 percent of installed MIPs maintaining or growing capacity, Red Hat at 11 percent growth, and Lightwell reaching general availability on July 8 with eleven named large financial institutions as early adopters. Every one of these is disclosed by the vendor, none is verified independently, and Lightwell's named adopters come with no usage, seat or spend detail. Anderon is only a letter of intent.
Candid on the miss, promotional on the pivot
On the financials the letter is unusually plain: it names the shortfall, admits IBM faltered and says slipped deals drove most of it, which pulls the gap toward zero. The overstatement sits in the surrounding framing, where an unquantified $5 billion Lightwell commitment, more than 20,000 engineers, a $10 billion quantum plan and a letter of intent described as the world's first pure-play quantum wafer foundry are presented alongside the miss with no verification, pricing or schedule. External causes (memory and server supply, unnamed cybersecurity concerns) also carry explanatory weight the source does not evidence. Net: modestly overstated relative to what is demonstrated.
Issuer-controlled channel, pre-earnings framing
The sole source is the company's own newsroom, the author is the CEO, the audience is named as investors, and the document lands eight days before IBM's scheduled July 22 earnings call while explaining a revenue and guidance miss. The letter selects which metrics to disclose (non-GAAP margins and EPS, program-to-program percentages, growth rates without dollar bases) and defers full-year expectations to the call. There is no independent publisher in the cluster to offset that positioning.
Reported facts solid, interpretation single-sourced
Confidence is high that IBM said what the ledger records and that the reported figures are as stated, because the source is the primary document. Confidence is materially lower on the interpretation layer: one publisher, no independent verification of the capex-shift narrative or the Lightwell and quantum claims, preliminary numbers subject to change at close, and no updated guidance until July 22. Derived items such as the mix shift and the free cash flow conversion ratio follow from the letter's own arithmetic but cannot be decomposed further.
security
The disclosure pipeline is triaging itself: 20,700 new CVEs, 10% more exploitation1 distinct publisher
invest
The chips never move: Washington's fix for the Southeast Asia compute loophole1 distinct publisher
security
Akrites switches on in September with 20-odd members and a one-to-10 engineer donation band1 distinct publisher
science
IBM's bottleneck was cold volume, and its answer is an 8-foot box you can bolt to another one1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 17, 2026