IBM’s 25% Share-Price Fall Is a Warning, But Not Yet a Verdict

A 25% fall in IBM’s share price on 14 July left little room for measured judgment. After the largest one-day drop in IBM’s history, commentators jostled to explain what’s gone wrong and what it means for the technology sector. Many concerns are justified, but some criticism has bordered on the terminal and others assume the problem will pass once supply constraints loosen.

The facts were uncomfortable. IBM warned that second-quarter revenue would see a 1% increase, well below expectations, with growth in software slowing, infrastructure revenue declining, and several large transactions failing to close. CEO Arvind Krishna acknowledged that IBM hasn’t adapted quickly enough as customers redirect budget toward servers, storage, memory and other scarce AI infrastructure.

These are material shortcomings. They’re particularly awkward for a company whose valuation had come to reflect confidence in steadier execution and dependable software growth. But they should be judged against IBM’s longer operating record, portfolio direction and investment strategy, rather than treated as proof that the company’s transformation vanished in a single quarter.

Our assessment draws on several years of tracking IBM’s business performance, leadership direction, product strategy and acquisitions. That record shows a company with a clearer identity than some of the harsher commentary allows, but also one that has less room to explain away weak execution.

The Fall Was Justified, But Some Conclusions Were Not

IBM misread the speed at which customer priorities were changing and didn’t adjust its sales response quickly enough. Delayed transactions raise questions about forecasting and pipeline management. Wider market pressure helps explain the shortfall, but it doesn’t excuse IBM’s response.

The more-severe interpretations still go too far. From 2Q25 through 1Q26, IBM reported total revenue growth of 8%, 9%, 12% and 9%. Software revenue rose 10%, 10%, 14% and 11% over the same four quarters, alongside stronger margins, improving cash generation and progress spanning the Red Hat, Data, Automation and IBM Z segments.

IBM’s product strategy is credible, and its acquisitions fit a recognizable architecture. Red Hat provides the hybrid application platform, HashiCorp strengthens provisioning and control, and Confluent adds the real-time data capabilities that AI agents increasingly require. IBM Z supports secure and regulated workloads.

However, IBM needs commercial proof. It must still demonstrate sustained organic growth, disciplined integration and returns that justify the capital committed. Red Hat needs a steadier rhythm, HashiCorp competes with the hyperscale cloud providers and Confluent brings near-term costs. IBM’s $5 billion Project Lightwell commitment and planned investment of more than $10 billion in quantum computing add longer-term options, although neither is a near-term earnings remedy.

IBM’s transformation wasn’t just slideware. Nor is the company beyond reproach. There’s no guarantee that growth will resume, but it makes it difficult to claim that IBM’s entire AI and software story evaporated overnight.

The more defensible conclusion is a sharp execution setback after several stronger quarters. Well-run companies can still make consequential mistakes. The real test is whether management corrects them quickly enough to prevent an uncomfortable quarter from becoming a habit.

The Picks-and-Shovels Phase Is Still Setting the Pace

The old “picks and shovels” adage, which places equipment manufacturers as the best bets in a new market, still appears to hold in AI. For now, the clearest economics sit with suppliers of AI processors, memory, servers, storage and data centre capacity as enterprises secure the infrastructure before deciding which software and operating models will create lasting value.

IBM’s warning adds weight to that view. Customers appear to have brought forward purchases of scarce equipment ahead of expected price increases, delaying spending on software, mainframes and transformation programmes. In the near term, AI is displacing other priorities rather than simply expanding the total technology budget.

Suppliers of chips and memory components are therefore in a stronger position: they sell what buyers believe they must secure now. Buyers are delaying the harder questions about utilization, integration, governance and return on investment until the hardware arrives and someone has to make it useful.

Infrastructure alone doesn’t create a business outcome, as enterprises still need data management, security, automation, integration and application modernization. IBM’s broader proposition sits in those layers, helping customers connect AI to existing systems and operating processes.

IBM isn’t wholly outside the picks-and-shovels category. IBM Z, distributed infrastructure and storage are part of the foundations for AI used by large enterprises. However, the company’s growth assumptions depend on customers moving from capacity acquisition to practical deployment. That transition may take longer than IBM would like. Technology budgets have a habit of becoming finite just when industry narratives assume otherwise.

Mainframe Weakness Should Be Judged Through the Product Cycle

There are questions about IBM’s modelling of its product cycle for IBM Z. Revenue from IBM Z fell 14% in the first quarter of 2025 as the z16 cycle ended, then accelerated following the launch of z17, rising 70%, 61%, 67% and 51% in the subsequent four quarters. The franchise didn’t disappear and reappear in a few months; it moved through the familiar rhythm of a hardware transition.

The timing of this cycle is unfortunate: some demand was probably brought forward during the early refresh, leaving tougher comparisons just as customers began prioritizing memory, servers and storage. However, IBM should’ve modelled these dynamics more accurately. It can’t invoke cyclicality as though it were an unexpected meteorological event. But the evidence doesn’t yet show structural erosion of the mainframe franchise.

The z17 is also more than just a routine refresh. IBM has positioned it as providing resilience, quantum-safe security, energy efficiency and AI inferencing for regulated, high-volume workloads. The question is whether that demand remains durable once the initial surge fades, not whether the platform suddenly ceased to matter.

The More Serious Issue Is the Quality of AI Revenue

It’s inaccurate to say IBM has no AI business. Its position spans data, automation, OpenShift, governance, security, consulting and mainframe inferencing. The legitimate question is the scale, incrementality and repeatability of the revenue.

Much of IBM’s book of business for generative AI has been associated with its Consulting unit, but growth in that segment has remained modest. This raises the possibility that some existing consulting, modernization and transformation engagements are being recast as AI-related work rather than representing entirely incremental expenditure. That’s not the same as saying IBM’s AI position is artificial. A modernization programme can become more valuable because it includes AI without increasing the total contract value. Such work may also be essential to deployment even when it doesn’t produce a clean, separately identifiable AI revenue line.

The distinction still matters. Suppliers should separate AI-influenced activity, genuinely new AI revenue and the conversion of that work into repeatable software consumption. However, IBM is hardly alone here. Almost every project now seems to acquire an AI label somewhere between the contract signing and the earnings report.

IBM’s recent decision to reduce the separate disclosure of its AI book of business increases the burden of proof elsewhere. Investors will judge progress through software revenue, annual recurring revenue, growth in data and automation revenue, consulting backlog conversion and margins. IBM says AI is embedded throughout the business. This may be true, but once a separate measure is removed, the remaining metrics must offer more detail.

A Coherent Strategy Doesn’t Guarantee a Fairly Valued Share

IBM must demonstrate sustained organic growth, disciplined integration and returns that justify the capital committed. The company can remain strategically coherent, financially sound and well-managed while its shares are also valued too highly. The warning challenged assumptions of sustained software acceleration, stronger AI conversion and smooth execution.

The 25% fall in share price is easier to understand when seen as the removal of a premium built on high expectations. IBM’s competitive position didn’t deteriorate by 25% in a day. Investor confidence simply moved faster than IBM’s enterprise technology portfolio.

The company has suffered a significant execution setback at a point when its valuation allowed little room for error. Its strategy, infrastructure foundations, cash generation and acquisitions remain credible. The next few quarters must show that delayed transactions can be recovered, software growth can regain momentum, its consulting business can turn AI-related activity into incremental revenue, and the z17 franchise can remain durable beyond the early surge. IBM’s 2Q26 results on 22 July should clarify whether delayed deals are recoverable and if the spending shift is temporary or more structural.

IBM’s explanation is plausible. The market, having briefly misplaced its enthusiasm and then found its anxiety, will now require proof rather than reassurances.