IBM’s Big Mainframe Bet Takes a Hit as AI-Boom Costs Shake Customer Spending

IBM said its latest quarter fell far short of Wall Street’s expectations, and the surprise was sharp enough that the company warned investors before reporting results. The immediate shock came from a 42% plunge in mainframe sales, a decline IBM says was driven not by demand collapse but by customers delaying purchases while AI-fueled hardware inflation pushed up other technology costs.

The episode matters because IBM’s mainframe business remains a critical profit engine: a single hardware sale can lead to years of software and services revenue. Even as the company posted $17.2 billion in quarterly revenue and nearly $2.2 billion in profit, the miss triggered a record stock drop and forced IBM to cut its full-year outlook.

For a company that has spent years trying to convince investors it can thrive in the cloud and AI era, the quarter was a reminder that its old-school systems business still has enormous influence over its financial results. IBM’s leadership insists the setback is temporary and that demand for mainframes has merely been pushed back, not lost.

Why IBM’s quarter rattled investors

IBM’s latest report was not a collapse in the ordinary sense. The company still produced healthy revenue, strong gross profit, and significant margins. But public markets were looking for more, and the gap between expectations and reality was large enough to overshadow the underlying profitability.

Before the official earnings release, CEO Arvind Krishna and the board took the unusual step of telling investors that results would be weaker than previously anticipated. That early warning gave the market time to react, and IBM shares fell dramatically, registering the steepest one-day decline in the company’s history.

The warning helped explain the surprise, but it did not soften it. IBM also revised its full-year guidance downward, signaling that the weakness would not be confined to one quarter. For investors, that raised a more important question: was this simply a timing issue, or a sign that the company’s hardware mix is becoming more vulnerable in an AI-distorted market?

What caused the mainframe slump?

IBM said the mainframe decline came from customers delaying upgrades rather than abandoning the platform. According to the company, “tens” of clients that were expected to buy new systems during the quarter decided to hold off.

That may sound modest, but in the mainframe world, a relatively small number of delayed orders can make a huge difference. These systems are expensive, often costing hundreds of thousands of dollars and sometimes reaching into the millions once software, support, and long-term maintenance are added.

IBM’s executives said the customers who postponed mainframe purchases did not simply sit on their budgets. Instead, many redirected spending toward other infrastructure and personal computing equipment, where prices had risen sharply.

Arvind Krishna said the issue was less about customers losing interest in the mainframe and more about them being forced to prioritize other equipment after encountering steep price increases elsewhere in the market.

That explanation points to an unusual spillover effect from the AI boom. Massive demand for data center components has pushed up the cost of memory, servers, and other hardware. In that environment, some enterprise buyers appear to be making tradeoffs, buying what they need most urgently and postponing other major purchases.

How the AI boom hit IBM from two directions

The AI boom has helped IBM in recent years by boosting spending across the broader data center ecosystem. But the same surge has also created cost pressure that is complicating buying decisions for enterprise customers.

IBM is not alone in seeing the impact. Rival hardware makers, including Dell and HP, have already warned that rising component costs have pushed them to raise prices. Apple has also pointed to similar pressures. In other words, AI is creating both a tailwind and a cost shock across the hardware market.

IBM’s problem is especially acute because its mainframe business is tightly connected to its software business. The company said it typically earns about $3 in software revenue for every $1 of mainframe hardware sold. That means one delayed hardware purchase can ripple through several layers of IBM’s income statement.

For a business that relies on long software contracts and ongoing support, the danger is not just one missed sale. It is the deferral of a whole revenue stream that may have been expected to arrive in the same reporting period.

Why a few delayed orders matter so much

They matter because the mainframe market is low-volume but high-value. Unlike consumer electronics, where sales are measured in millions of units, IBM’s mainframe buyers are large enterprises with procurement cycles that can move the company’s quarterly numbers in a major way.

When those customers delay decisions, the impact is magnified by the attached ecosystem of software licenses, services, and long-term support agreements. That makes the mainframe business unusually sensitive to timing shifts, even when total customer demand remains intact.

• Mainframes are bought infrequently but in very large transactions.
• Each hardware sale can unlock recurring software revenue.
• Delayed upgrades can distort one quarter while preserving long-term demand.
• Price pressure elsewhere in IT budgets can change purchase timing.

IBM says the problem is temporary

IBM’s leadership tried to reassure investors that the weakness does not reflect a structural decline in the mainframe franchise. Krishna and Chief Financial Officer Jim Kavanaugh said the deferred purchases should eventually convert into completed deals, and that some customers have already bought new systems in the current quarter.

The company also argued that it has not seen signs of customers abandoning the platform for competing architectures. That point is crucial, because IBM has spent decades defending the relevance of the mainframe against predictions of its demise.

In the company’s view, the quarter was distorted by budget pressure and the timing of purchases, not by a long-term shift away from the platform. If that is correct, the hit should soften as customers work through their spending priorities.

IBM executives said they see no evidence that clients are moving away from the mainframe, describing the drop as a postponement rather than a permanent loss of business.

Still, investors have reasons to remain cautious. Companies often describe demand shocks as temporary until the data proves otherwise. The challenge for IBM will be showing that the delayed orders do, in fact, convert into revenue soon enough to restore confidence in its growth trajectory.

How is the mainframe still central to IBM?

The mainframe remains central to IBM because it is not just a hardware product; it is a platform that anchors a much wider revenue engine. Although the company has diversified into hybrid cloud, consulting, and AI infrastructure, the mainframe still plays an outsized role in both profits and strategic positioning.

Many large organizations continue to rely on mainframes for transactions that demand high reliability, security, and performance. Banks, insurers, governments, and large retailers have long used them to process mission-critical workloads, and those workloads are not easily migrated.

That is one reason the market has repeatedly underestimated the platform’s longevity. Mainframes have outlasted multiple waves of supposed replacement technologies, from client-server systems to cloud-native services. IBM has been able to monetize that durability through hardware refresh cycles and recurring software sales.

What makes the business so profitable?

What makes the business so profitable is the blend of high-ticket hardware and sticky software. Once a customer commits to a mainframe environment, IBM can often generate recurring income through operating software, security tools, support contracts, and upgrades.

This makes the business harder to replace than many observers assume. Customers may delay purchases, renegotiate timing, or extend existing systems, but they often remain tied to the platform for operational reasons and because migration costs are high.

That also explains why IBM’s executives are so confident that the latest quarter was a timing issue. If customers stay within the ecosystem, the company can still recover the delayed sales and preserve the software relationship that follows them.

What the numbers show

IBM’s financial results showed a company still generating substantial cash, but one that fell well below analyst expectations. The headline figures underscore how much weight a single product category can carry inside a diversified technology giant.

Those figures help explain the mixed investor reaction. IBM was still profitable, but the market cares about momentum, and the guidance revision suggested the company’s momentum had slowed at precisely the wrong time.

What does this mean for the wider hardware market?

IBM’s results offer another sign that the AI infrastructure build-out is reshaping enterprise spending in unexpected ways. The boom is not only creating demand for chips, servers, and data center equipment; it is also raising costs enough to force customers to delay or reshuffle capital budgets.

That dynamic has implications well beyond IBM. Hardware vendors across the enterprise stack may continue to face pressure if memory and component prices remain elevated. Customers with finite budgets may choose the most urgent purchases first, leaving other infrastructure projects waiting in line.

For IBM, the concern is especially visible because its mainframe cycle is periodic and capital-intensive. A sudden spike in the cost of competing infrastructure can alter when customers decide to refresh their systems, even if the underlying need has not changed.

In that sense, IBM’s quarter may be a warning sign for the broader market: AI enthusiasm can distort purchasing patterns in both directions, lifting some segments while starving others of budget.

Why the mainframe death prediction keeps failing

The mainframe has been declared obsolete many times over the past several decades, yet it remains embedded in some of the world’s most important computing environments. The reason is not nostalgia. It is continuity, scale, and the difficulty of replacing systems that sit at the core of business operations.

Large organizations often prioritize stability over novelty when the systems involved handle payments, records, or other mission-critical workloads. Replacing a mainframe can be costly, risky, and time-consuming, especially when software and operational processes are deeply integrated.

That reality has given IBM a durable business even as the rest of the technology industry evolved rapidly. The latest quarter does not change that history, though it does highlight how dependent the company still is on the timing of a relatively small number of large customer decisions.

Could AI eventually change that?

Could AI eventually change that? It might, but not in the immediate way many observers assume. AI could influence procurement cycles, hardware pricing, and infrastructure architecture, but that does not automatically mean enterprises will abandon mainframes.

For now, IBM argues the opposite: that AI-related spending pressure is delaying mainframe purchases without displacing them. If that proves true, the company may simply be experiencing a noisy quarter rather than a lasting structural break.

Even so, investors will watch the next several quarters closely. If delayed orders turn into sales, IBM’s explanation will look persuasive. If they do not, the story could shift from temporary budget pressure to something more troubling about the platform’s future.

What investors should watch next

Investors will be looking for evidence in three areas: whether mainframe sales rebound, whether software revenue follows those sales, and whether IBM can stabilize its full-year guidance after the cut.

The company’s strongest case is that enterprise customers still need the platform and have not signaled a migration away. The weakest case is that AI-era pricing pressure is not a one-off but a recurring drag on large IT purchases.

1. Track mainframe sales in the next quarter for signs of recovery.
2. Watch software revenue tied to mainframe installations.
3. Monitor whether hardware inflation eases or continues.
4. Compare IBM’s results with other enterprise hardware makers.

For now, IBM is asking investors to view the quarter as a temporary distortion caused by a highly unusual technology spending environment. Whether the market believes that story will depend on what happens next.

The company may be right that AI is not killing the mainframe. But the quarter shows that AI can still wound it.