AI Spending Knocks IBM Mainframe Sales Down 42% in Shock

On Wednesday, IBM turned last week’s warning into official numbers, and the real signal was not that the mainframe is dead. It was that AI spending has become powerful enough to knock IBM mainframe sales off their expected path.

IBM reported quarterly results that still showed a large, profitable company, according to TechCrunch. Those are not broken-company numbers. The problem was sharper: IBM warned investors early, pointed to weakness in its infrastructure business, and exposed pressure on its cash-cow mainframe franchise.

That is why the stock reaction was so violent. After CEO Arvind Krishna told investors the quarter would be worse than expected, investors heard something more dangerous than one weak quarter. They heard pressure on a historically resilient franchise during the same AI boom that had helped lift IBM’s stock under Krishna.

Last week’s warning turned a mainframe miss into a trust problem

IBM did something unusual before the official earnings release. Krishna gave investors an early warning that the quarter would be worse than expected and that the company’s “infrastructure” category was especially weak.

That mattered because mainframes are not just another hardware line for IBM. CFO Jim Kavanaugh told investors that IBM earns $3 in software revenue for every $1 of mainframe hardware it sells. A missed hardware cycle can therefore ripple into software attachment, margins, and longer customer commitments.

The company’s explanation was specific. “Tens” of customers that were expected to buy a new mainframe during the quarter did not do so. On paper, that sounds small. In IBM’s world, it isn’t. TechCrunch notes that mainframes cost hundreds of thousands to millions of dollars, and maintenance contracts and software can generate many millions more.

“We see no evidence of clients moving off the mainframe,” Krishna said.

XOOMAR analysis: that quote is the hinge of the whole story. IBM is asking investors to treat the mainframe shortfall as delayed demand, not demand destruction. The stock drop shows the market is not ready to grant that assumption for free.

Wednesday’s numbers show why IBM mainframe sales carry outsized weight

The quarter was not weak because IBM stopped making money. It still generated substantial profit. The damage came from the mismatch between expectations and where the weakness landed.

Krishna’s explanation was that some customers redirected budget after facing price increases of 15% to 30% for data center gear and PCs. He connected those increases to the AI build-out boom, which has pushed up costs for components such as memory. TechCrunch also notes that enterprise hardware makers including Dell and HP have warned about rising component costs, and Apple has said the same.

That does not mean every postponed IBM deal turned into an AI server purchase. The source does not say that. It says customers bought other hardware after facing extreme price increases, and those prices were affected by the AI build-out.

XOOMAR analysis: this is a subtler problem than “AI replaced the mainframe.” AI appears to have changed the priority stack inside corporate hardware budgets. If a customer has to absorb sudden cost inflation on data center gear and PCs, a mainframe refresh can slide, even if the system still matters.

IBM says AI delayed deals, but investors need proof they come back

Krishna’s argument is that corporate technology budgets did not disappear. They were pulled toward urgent hardware needs after price shocks elsewhere.

“When they were faced with that issue, then they decided to move budget to those areas where they were having that extreme price,” Krishna said.

That framing helps IBM because it preserves the central mainframe story. The customer still wants the system, according to management. The purchase just moved. Krishna also said some of those customers have already bought their new mainframes this quarter.

The weak spot is timing. A delayed enterprise hardware purchase is safe only if it eventually closes. If the next quarter shows those “tens” of customers returning, IBM’s argument gets stronger. If the delay stretches, investors will read the miss differently: not as a budget hiccup, but as a reset in how customers rank IBM mainframe sales against other infrastructure needs.

XOOMAR analysis: the mainframe’s role is not being judged in isolation anymore. It is being judged against whatever spending demand AI creates across the rest of the hardware stack. That is a harder comparison than past technology debates because it hits budget calendars immediately.

Decades of mainframe death calls meet a harsher AI budget cycle

TechCrunch makes the historical point plainly: the tech industry has predicted the death of the mainframe for many decades. IBM’s latest defense leans on that history without saying it too loudly. Mainframes have survived repeated waves of replacement talk because large enterprises tend to move core systems slowly.

But AI spending creates a different kind of pressure. It does not need to fully replace the mainframe to hurt IBM. It only needs to delay enough high-value purchases to dent infrastructure revenue and the software revenue that follows.

That is the lesson for hardware markets. Price and budget pressure can hit even products customers still intend to buy. We have seen separate hardware stories where pricing and timing mattered, including XOOMAR’s coverage of Steam Deck sales plunging after Valve’s brutal price hike. The IBM case is enterprise-scale, but the budget logic is familiar: when hardware costs jump, buyers make trade-offs.

AI also keeps pulling attention toward new devices and compute strategies, a theme visible in our reporting on how an Apple lawsuit threatens OpenAI hardware at the worst time. IBM’s issue is different, but the timing pressure is similar. Hardware decisions are getting squeezed by the AI cycle.

The same quarter sends different signals to IBM, CIOs, and AI vendors

For IBM, the message is simple: this was temporary. The company says clients are not leaving the mainframe, and some delayed purchases have already returned this quarter.

For investors, the message is harsher. A sharp decline in a cash-cow business is not easily waved away, especially after IBM warned that infrastructure weakness was weighing on results. The stock reaction showed little patience for a timing explanation.

For enterprise buyers, the lesson is budget discipline. XOOMAR analysis: if AI-related cost pressure forces companies to postpone core infrastructure, the risk is not visible right away. It shows up later, when delayed refresh cycles create operational or financial friction.

For AI vendors, IBM’s quarter is useful evidence without needing to prove replacement. Every postponed mainframe purchase suggests AI-linked spending pressure can reorder enterprise budgets, even in categories that remain deeply embedded.

IBM’s next test is turning delayed mainframe orders into reported revenue

IBM now has a clean test ahead. Management said some delayed customers have already bought new mainframes this quarter. Future results need to show that last quarter’s weakness really was a timing issue.

The evidence that would support IBM’s case is straightforward:

• Recovered orders: delayed mainframe customers show up in later quarters.
• Stabilized infrastructure revenue: the weak “infrastructure” category stops dragging results.
• Protected software attachment: the $3 of software revenue per $1 of mainframe hardware remains intact.
• Clearer AI positioning: IBM explains why mainframes still fit into customer plans while AI spending absorbs budget attention.

The evidence that would weaken IBM’s case is just as clear: more delays, weaker software follow-through, or another warning tied to infrastructure.

AI probably has not killed the mainframe. Krishna may be right on that narrow point. But it has changed the budget politics around IBM mainframe sales. IBM’s job now is to prove that AI caused a short-term shock, not a lasting reset in enterprise hardware priorities.

The Bottom Line

• IBM’s weak infrastructure result raised doubts about one of its most dependable profit engines.
• AI spending may be changing enterprise budget priorities faster than expected.
• The stock reaction shows investors are worried about trust and visibility, not just one bad quarter.