IBM just told every enterprise technology buyer something more important than an earnings miss: your AI budget isn’t shrinking; it’s moving.
On July 14, IBM warned investors that preliminary second-quarter revenue of $17.2 billion would fall well short of the roughly $17.9 billion Wall Street expected, with adjusted earnings of $2.93 a share missing the $3.01 consensus. The stock dropped as much as 23% before the opening bell, its worst single-day reaction in years. But the headline number isn’t the story CIOs should be tracking. The reallocation behind it is.
Software Grew And Infrastructure Didn’t.
Software revenue rose 5%, and consulting held roughly flat. Infrastructure, the mainframe and hardware business, fell 7%, and that segment carried the miss. CEO Arvind Krishna’s explanation, laid out in a letter to investors, was blunt: in the closing weeks of June, clients diverted quarterly capex toward servers, storage, and memory, racing to lock in supply-constrained infrastructure ahead of expected price increases. Software and consulting deals expected to close on schedule didn’t. Krishna’s own words were unusually direct for a shareholder letter: “These conditions require our teams to execute perfectly, and this quarter we faltered.”
The news confirms what many CIOs have felt anecdotally over the past two quarters: enterprise AI spending hasn’t cooled; it has shifted. Budgets were reallocated to compute, storage, and memory capacity before prices climb further. If your infrastructure team has been fighting procurement delays or unexpected price pressure on GPUs, storage, or memory this year, IBM’s numbers are the first hard evidence that you’re not alone, and that this is now large enough to move a $270 billion company’s stock by double digits.
AI Investment Continues
While the hardware business absorbed the hit, IBM didn’t retreat from its AI positioning. Lightwell, the $5 billion commitment IBM and Red Hat made in May to secure open-source software using frontier AI models, reached general availability on July 8, with early adopters including Bank of America, JPMorgan Chase, Goldman Sachs, and Visa. Krishna also reaffirmed a five-year, $10 billion investment in quantum computing tied to a new domestic chip foundry backed by CHIPS Act incentives. Neither initiative slowed down because of a soft quarter. That’s a data point for CIOs weighing vendor stability: IBM is treating this as a demand-timing problem, not a strategy problem, and it’s continuing to fund the platforms it expects enterprises to depend on for AI governance and security over the next several years.
Rethink IT Budget Discussions
Three things are worth raising with your CFO and your board.
First, expect continued volatility in infrastructure pricing. If clients are pulling forward hardware purchases at IBM’s scale, similar dynamics are likely playing out across cloud providers and chipmakers. Lock in capacity commitments where you can forecast reliably.
Second, don’t mistake a vendor’s infrastructure miss for a sign of weakening AI demand. IBM’s software business, the part of the portfolio most directly tied to AI-enabled products, still grew. The market is punishing execution and timing, not the underlying thesis.
Third, watch how vendors fund strategic AI bets during a rough quarter. A vendor that keeps investing through a miss, as IBM has with Lightwell and quantum, is signaling where it expects the multi-year value to land. That’s useful intelligence when you’re deciding which platforms to build long-term dependencies on.
IBM holds its full call on July 22, and that’s when the real picture, and full-year guidance, will come into focus. Until then, the lesson for executives outside IBM’s shareholder base is the more durable one: the AI infrastructure race hasn’t gotten smaller. It’s gotten more expensive, and it’s rewriting budget lines faster than most planning cycles can keep up with.
