IBM is finding out that the AI boom does not lift every old technology boat at the same time. The company has cut its full-year revenue growth outlook after a difficult quarter in which customers appeared to move budget toward AI infrastructure and away from parts of IBM’s more traditional enterprise stack.
Reuters reported through MarketScreener that IBM now expects annual revenue growth of 4 percent to 5 percent, down from previous guidance of more than 5 percent. The company earned $2.2 billion on $17.2 billion in revenue, while adjusted earnings of $2.93 per share came in slightly below expectations.
The most painful number is in infrastructure. Mainframe sales fell 42 percent, and that matters because IBM does not only make money when it sells the machines. As TechCrunch noted, IBM also earns meaningful software revenue around those systems, so a weak mainframe cycle can ripple through the rest of the business.
IBM’s explanation is worth paying attention to. Management says customers have not necessarily abandoned IBM, but some large deals slipped as companies redirected spending toward servers, storage and AI-related infrastructure. In plain language, the AI spending wave is forcing enterprise buyers to choose what they fund now and what waits.
That is an uncomfortable place for IBM because it has been trying to tell a software, hybrid cloud, AI and consulting story around Red Hat, watsonx and enterprise automation. The quarter shows that even companies selling AI tools can be hurt when clients prioritise physical AI infrastructure first.
The market had already reacted harshly after IBM warned earlier about the shortfall, with the stock suffering one of its worst single-day drops in decades. The latest results did not make the story worse than feared, but they did confirm that this is not a small timing issue.
This connects with the broader AI infrastructure pressure we have been watching across the industry. OpenAI’s reported $750 billion compute plan shows how much capital is moving into power, chips and data centres. IBM’s quarter shows the other side of that same shift: budgets moving away from older priorities to pay for the AI buildout.
IBM is not dead, and mainframes are not suddenly irrelevant. Banks, airlines, governments and large institutions still run critical systems on them. But this quarter is a reminder that AI is not just creating new winners. It is also changing the order in which enterprises spend, and that can hurt even well-established vendors.