IBM Trims 2026 Revenue Outlook After Q2 Miss and 42% Mainframe Sales Drop

IBM's Mainframe Sales Just Fell 42%. Here Is What That Number Does and Does Not Mean.

On July 22, 2026, IBM reported second-quarter revenue of $17.16 billion, up about 1% from a year earlier^[1][3]. That fell short of the roughly $17.58 billion Wall Street had expected^[1]. Adjusted profit came in at $2.93 per share, just below the $2.97 analysts had forecast^[1].

IBM also cut its outlook for the rest of the year. It now expects "constant-currency" revenue growth of 4% to 5% for 2026, down from a prior forecast of more than 5%^[1]. Constant currency strips out the effect of exchange-rate swings, so it is meant to show how the underlying business is doing, not how the dollar moved.

The single biggest reason for the miss was IBM's Z line of mainframe computers, the large, high-reliability machines that banks and governments use to run core transactions. Z revenue fell 42% from a year earlier^[1][3]. The company's chief financial officer, James Kavanaugh, said that drop alone cut more than five percentage points off IBM's total growth, far worse than the one or two points the company had planned for^[1][7].

The report was not a total surprise. On July 14, IBM had issued a rare early warning that the quarter would come in weak. That single announcement sent IBM's stock down about 25% in one day, the worst one-day drop in the company's history^[2][8].

What Everyone Agrees On

Several facts are not in dispute. IBM's software revenue rose 5%, to $7.76 billion, though it too missed analyst estimates^[1][3]. The company's infrastructure segment, which includes the mainframe business, fell 7% to $3.84 billion^[3][7].

It is also agreed that IBM kept one important number unchanged: its guidance for full-year free cash flow, the cash left over after running the business, which it still expects to rise by about $1 billion this year^[5][12]. First-half free cash flow came in near $4.8 billion^[5][12].

And everyone agrees on the cause, at least in broad strokes. In late June, customers shifted some spending toward AI-related servers, storage, and memory, ahead of expected price increases^[3][6]. That pulled money away from other purchases in the same quarter, including mainframes.

Where the agreement ends is on what that shift means for the future. Was it a one-time delay, or the start of something longer?

The Mainframe Cycle, Explained

To understand the 42% figure, it helps to know how IBM sells mainframes. The company releases a new model every few years — the current one, called z17, launched in mid-2025. Sales spike right after launch, as customers rush to upgrade, then taper off as the cycle matures^[5][11].

That means a 42% year-over-year decline can happen even in a healthy product cycle, if the comparison quarter was an unusually strong one. IBM says that is exactly what happened here: it is comparing this quarter against the quarter when z17 first launched, not against some steady baseline^[3][11]. One outlet, the Canadian financial site NAI500, described the drop as "sequential," meaning quarter-over-quarter, but IBM's own press release and its filing with the Securities and Exchange Commission both describe the comparison as year-over-year, against the same quarter in 2025^[3][4][11].

IBM argues the broader cycle still looks strong. Kavanaugh said the first five quarters of z17 sales are running nearly 30% ahead of the equivalent stretch for the prior model, z16 — a pace the company describes as its best mainframe cycle ever^[5].

How IBM Sees It

IBM's leadership, led by CEO Arvind Krishna and CFO Kavanaugh, frames the quarter as a timing problem, not a sign of customers leaving. The company says dozens of large infrastructure deals slipped past the end of the quarter rather than being canceled outright^[3][6]. Krishna said on the earnings call that roughly a third of those delayed deals had already closed by the time of the call, in the third quarter, which he offered as evidence the business was deferred rather than lost^[12].

Kavanaugh put it directly: "We see no evidence of clients moving off a mainframe^[6]." IBM also points to its unchanged free cash flow guidance as proof the underlying business remains healthy, whatever the quarterly mainframe number looks like^[5][6].

IBM has an obvious stake in this framing. The company is trying to rebuild investor confidence after a record one-day stock crash, and it needs its long-term growth story around hybrid cloud and AI to stay intact^[2][6]. That does not make the explanation wrong, but it is worth knowing why IBM is motivated to tell it.

How Skeptics See It

Not everyone is convinced the weakness is temporary. Bearish analysts and some traders argue that heavy corporate spending on AI hardware could be crowding out budgets for IBM's older product lines in a way that does not simply reverse next quarter^[9]. They point out that software revenue also missed expectations, and that IBM lowered its software growth target too, suggesting the problem may be broader than one hardware cycle^[1][3].

The Motley Fool, a retail-investor commentary site, described the results as "weak, but not terrible," a framing meant to reassure long-term holders while still acknowledging real damage^[9]. Other skeptics go further, warning that a company growing revenue by just 1% while cutting its outlook can be what investors call a "value trap" — a stock that looks cheap on paper but keeps getting cheaper because the underlying business is slipping^[9].

Underneath both views sits the same undisputed pressure: corporate technology budgets are finite. When a company pours more money into AI computing, something else usually gets pushed down the list, and this quarter, IBM's traditional hardware and software were among the things pushed down^[6][14].

How the Coverage Split

News outlets largely agreed on the numbers but differed sharply in tone. CNBC and Reuters stuck close to the facts, laying out the earnings miss, the guidance cut, and management's explanation with relatively little editorial framing^[1][6].

Fox Business emphasized the shock of the sell-off, describing IBM as sending "a shockwave through the tech industry," a framing that foregrounds the market drama over IBM's cycle-timing defense^[10]. TechCrunch took a more skeptical angle, with a headline saying IBM "insists" AI is not killing the mainframe — a word choice that signals doubt about whether management's reassurance will hold up^[7].

Coverage outside the United States, including the Canadian site NAI500 and the India-based site Bitcoinworld, tended to focus on the numbers and the AI-budget explanation rather than offering a distinct national or political angle^[11][14]. NAI500's description of the 42% drop as "sequential" appears to be a factual error rather than a nuance other outlets missed, since IBM's own press release and regulatory filing describe it as a year-over-year figure^[3][4][11].

What remains unresolved is not a matter of fact-checking but of time. The revenue miss, the guidance cut, the 42% mainframe decline, and the record stock drop on July 14 are all agreed upon^[1][2][3][8]. Whether the delayed deals and diverted AI spending come back to IBM in future quarters, or represent a lasting shift away from its legacy products, will only become clear in the results still to come^[6][9].