IBM's surprise preliminary second-quarter warning on July 14, 2026, gave investors a sharper look at a problem spreading through enterprise technology budgets: AI infrastructure is not just adding new spending. In some cases, it is pulling money away from software, consulting and traditional hardware plans.
The company said it expects second-quarter revenue of $17.2 billion, up 1 percent from a year earlier, and operating earnings of $2.93 a share. Both figures came in below Wall Street expectations cited by the Associated Press, which reported analyst estimates of $17.86 billion in revenue and $3.01 a share.
IBM shares fell 25.2 percent on Tuesday, according to AP market coverage, making the company the biggest drag on the Dow even as the broader market rose after cooler inflation data. The move mattered beyond one ticker because IBM is a large supplier to banks, governments and big companies that are now deciding how much of their technology budgets must go into AI-ready infrastructure.
The numbers
In a letter to investors, IBM said software revenue rose 5 percent, consulting was flat, and infrastructure revenue fell 7 percent. The company also reported year-to-date net cash from operating activities of $7.8 billion and free cash flow of $4.8 billion.
The shortfall was concentrated in areas that had been expected to hold up better. IBM said its Z mainframe performance and related transaction-processing software were weaker than planned, even though the new z17 program was still ahead of the prior z16 cycle by some internal measures.
Why investors care
The important part is IBM's explanation. The company said clients shifted late-quarter capital spending toward servers, storage and memory purchases to secure supply-constrained infrastructure before expected price increases. That is a practical consequence of the AI buildout: companies may need more compute capacity and memory now, but the money has to come from somewhere.
That makes IBM's warning useful for readers watching the broader AI trade. If more corporate buyers front-load hardware purchases or delay large software deals, the winners and losers of the AI boom could be less obvious than the headline demand suggests. Chipmakers, memory suppliers and data-center vendors may benefit, while some software and services contracts face longer approval cycles.
The caveat
This is still preliminary. IBM's final second-quarter results are scheduled for July 22, 2026, at 5 p.m. ET, and the company said the final numbers could differ slightly. It also pointed to strengths, including faster Red Hat growth, strong distributed infrastructure demand and continued investment in AI security and quantum computing.
One quarter does not prove that enterprise software demand is breaking. But it does show that AI spending can create budget friction, not just new growth. For investors and technology buyers, the next question is whether IBM's warning was a company-specific execution miss or an early sign that AI infrastructure costs are forcing harder tradeoffs across corporate IT plans.
What to watch next
The July 22 earnings call should clarify whether IBM expects the delayed deals to close later in the year, whether infrastructure weakness is temporary, and how management is changing its forecast for full-year demand. Watch for similar comments from software, consulting and cloud companies as earnings season continues.
