IBM’s stock dropped roughly 23% after the company warned that Q2 revenue would come in at $17.2 billion, up just 1% year over year and short of analyst expectations. Software grew 5%, consulting was flat, and infrastructure fell 7%, but the story underneath those numbers is more specific than a broad miss. Customers pulled capital out of software and services deals in the final weeks of June to stockpile servers, storage, and memory ahead of expected price increases, a shift CEO Arvind Krishna admitted IBM neither anticipated nor adapted to quickly enough. IBM’s distributed infrastructure business<\/a> actually posted its best quarter ever, with 37% revenue growth and a $500 million backlog, which makes the software underperformance harder to explain away as a market-wide slowdown.<\/p>\n The CIOs who matter here are not the ones still debating AI budgets in the abstract. They’re the ones who have already committed capital to hardware and are now sitting on infrastructure with a software roadmap that hasn’t caught up. IBM’s quarter is a clean signal about sequencing: organizations are front-loading compute, storage, and memory, and that pull-forward crowds out the consulting and software line items that vendors like IBM depend on to monetize the AI story. If your organization is in that infrastructure-first posture, your software vendors are about to feel it, and IBM won’t be the last to warn.<\/p>\n The more important question is whether IBM’s software miss is a timing problem or a positioning problem. Krishna’s own letter concedes execution failures, not just market forces, and Zeus Kerravala at ZK Research makes the sharper point: vendors with a genuinely differentiated AI story are still closing software deals even in this capex environment. The infrastructure spending cycle is real, but it’s not a universal excuse. IBM’s consulting revenue was flat, not growing, which means the company isn’t capturing the follow-on work that a credible AI platform should be generating as customers build out infrastructure.<\/p>\n The budget cycle this sets up is the one to watch carefully. Infrastructure investments of this scale typically force a software reckoning six to twelve months later, when organizations need to actually run workloads on what they’ve bought. That’s the window where IBM should be able to recover, but only if it arrives with a software and services motion that customers find compelling enough to fund after a quarter of deferred spending. If IBM’s Q3 software numbers don’t accelerate meaningfully, the execution explanation stops being sufficient and the positioning explanation becomes the only one left standing.<\/p>\n Capex pull-forward happens when buyers accelerate planned hardware purchases into the current period to lock in supply or pricing before conditions worsen, think of it as panic-buying at the enterprise scale. The cash spent isn’t new spending, it’s future spending arriving early, which means the quarters that follow can look artificially soft on hardware while software and services, which weren’t pulled forward, appear disproportionately weak. For CIOs reading vendor earnings, a pull-forward quarter is a distortion, not a trend, but the sequencing effects on software budgets are real and can persist for multiple quarters.<\/p>\nWhat this means for your business<\/h2>\n
Concept deep-dive: Capex pull-forward<\/h2>\n