IBM’s Weak Q2 Warning Shows How AI Chip Spending Is Reshaping Enterprise Tech

IBM’s preliminary second-quarter update has become one of the clearest signals yet that the AI infrastructure boom is reshaping enterprise technology spending in uncomfortable ways.

In a letter to investors, IBM CEO Arvind Krishna said the company expects Q2 revenue of US$17.2 billion, up 1 percent year-on-year but below market expectations. Software revenue rose 5 percent, consulting was flat and infrastructure revenue fell 7 percent.

The most interesting part of the letter is the explanation. Krishna said that in the last few weeks of June, clients shifted quarterly capital spending toward servers, storage and memory purchases to secure supply-constrained infrastructure before expected price increases. In other words, customers chose hard infrastructure over some of IBM’s expected deals.

We already covered how AI spending is beginning to look like an inflation story and how memory costs are now pressuring smartphone shipments. IBM’s warning shows the same pressure moving through enterprise technology budgets.

IBM said the quarter was hurt by weaker-than-expected Z performance and the associated software stack, particularly Transaction Processing. The company also said several large deals failed to close on expected timelines, while clients were distracted by fast-moving cybersecurity concerns.

Krishna’s language was unusually direct. He said IBM did not adapt and move quickly enough. That matters because enterprise technology companies often explain misses through broad macroeconomic language. Here, IBM is acknowledging both external pressure and execution problems.

The result was sharp market concern because IBM is supposed to be one of the companies benefiting from enterprise AI, hybrid cloud and mainframe modernisation. A weak quarter raises a bigger question: are customers still spending on AI, or are they redirecting money to the physical infrastructure needed before software projects can scale?

AI is changing budgets. Companies that once planned neatly across software, consulting, hardware and cloud are now being forced to secure chips, servers, memory and storage before prices rise or supply tightens. That can delay other purchases, even when those purchases are still strategically important.

This is why the IBM update matters beyond IBM. It suggests that the AI boom is not lifting every tech company evenly. Some vendors sell the scarce infrastructure itself. Others sell software, services or systems that compete for the same budget. When customers panic-buy hardware, the timing of software and consulting deals can shift.

IBM still pointed to strengths. Red Hat revenue growth accelerated to 11 percent. Distributed Infrastructure had its best reported performance, up 37 percent, helped by demand for Power and Storage. IBM also highlighted its Lightwell initiative and continued investment in quantum computing.

That mixed picture is why the story is interesting. IBM is not being left out of AI entirely. It is being squeezed by the order in which customers are spending. Infrastructure first, software and services later, is not a comfortable rhythm for every enterprise vendor.

For the wider tech market, IBM’s warning is a useful reality check. AI demand is real, but it is messy. It can boost chipmakers and memory suppliers while disrupting software sales cycles, consulting plans and legacy enterprise expectations.

The AI economy is not one smooth rising tide. It is a reallocation of budgets, power and urgency. IBM just gave investors a sharper view of what that feels like from inside the enterprise stack.