IBM’s newest earnings, introduced this week, have highlighted a shift underway in enterprise know-how spending: Organizations are diverting extra capital towards the infrastructure required to assist synthetic intelligence workloads.
The corporate reported weaker-than-expected mainframe efficiency and delayed giant offers, whereas additionally pointing to prospects redirecting capital expenditure towards servers, storage and reminiscence. Whereas it noticed a 7% decline in direct infrastructure income, IBM’s distributed infrastructure income elevated by 37%, reflecting rising demand for the know-how wanted to assist AI workloads.
The broader market image suggests this shift is occurring alongside continued development in total know-how spending. Gartner forecasts worldwide IT spending will attain $6.37 trillion in 2026, up 14.2% from the earlier 12 months, with information heart techniques and IaaS among the many fastest-growing segments. But whereas enterprises are growing know-how investments, not each space of the portfolio will profit equally.
“Regardless of the sturdy development in spending, this isn’t a ‘rising tide lifts all boats’ market development,” mentioned John-David Lovelock, distinguished vice chairman analyst at Gartner, within the agency’s July 27 report. “Know-how budgets are being strained by inflation, provide shortages, rising {hardware} and reminiscence prices, AI funding initiatives and shifting priorities.”
For CIOs, that creates a troublesome portfolio administration problem. AI funding is accelerating, however it’s competing for consideration and assets alongside cybersecurity, modernization, infrastructure resilience and operational enchancment initiatives that stay important to working the enterprise.
The query going through know-how leaders is the way to consider these competing priorities as AI turns into a bigger a part of the know-how panorama.
AI raises the stakes for know-how investments
The rising give attention to AI is growing strain on CIOs to exhibit clear enterprise worth from know-how investments throughout the portfolio.
Simon Ratcliffe, fractional CIO at fractional IT management agency Freeman Clarke, mentioned he believes AI has launched a brand new strategic precedence whereas current operational tasks stay intact. Safety, compliance, resilience, purposes and infrastructure nonetheless require funding, however CIOs are actually being requested to create funding capability for AI initiatives whose long-term economics should be growing.
“The change I see is that CIOs are being pressured to tell apart rather more clearly between know-how that merely retains the group working and know-how that genuinely improves its aggressive place,” Ratcliffe mentioned. “AI has made the tolerance for undifferentiated IT expenditure significantly decrease.”
That strain is more likely to have an effect on a variety of know-how investments, not simply AI initiatives. Adrian Murray, founder and CEO of Fisent Applied sciences, a developer of enterprise GenAI workflow software program, mentioned he sees this as a broader shift in how organizations ought to take into consideration all know-how investments. He argued that AI initiatives must be evaluated as a part of broader enterprise capabilities.
“The rise of AI forces a shift from managing know-how as remoted, opportunistic tasks to constructing a unified automation cloth that acts as core enterprise infrastructure,” Murray mentioned.
Figuring out which investments create reusable capabilities throughout the group and which stay restricted to particular person use circumstances will probably be a major problem for CIOs within the months to return.
The rising significance of enterprise foundations
AI could also be getting quite a lot of the eye, however its success will depend on different, much less flashy investments. As organizations increase AI adoption, consideration is growing on the underlying capabilities that decide whether or not these AI initiatives can succeed at scale .
Felix Van de Maele, CEO and co-founder of enterprise information intelligence and governance platform Collibra, argues that organizations want to guard investments in areas reminiscent of information platforms, governance, safety and techniques of file as a result of these capabilities present the context AI techniques require.
“I might resist framing it as AI versus the whole lot else as a result of the tasks most tempting to raid for AI funding are sometimes those AI will depend on,” Van de Maele mentioned.
Ratcliffe mentioned he sees an analogous sample in infrastructure spending. He argued that IBM’s outcomes look “much less like infrastructure being deserted and extra like expenditure being quickly reordered across the infrastructure wanted for AI.”
“Cash is shifting towards compute, storage, information and AI-enabling capabilities, whereas expenditure that can’t exhibit urgency, differentiation or measurable worth is more and more weak,” he defined.
The identical dynamic applies past infrastructure. Murray pointed to API enablement, platform engineering, structured information pipelines, orchestration and observability as examples of capabilities that turn out to be more and more vital as organizations transfer AI from experimentation into manufacturing.
These foundational investments increase a sensible query: how are organizations creating room for AI-related spending whereas rearchitecting and aligning the capabilities wanted to assist it?
Rethinking how the IT portfolio will get funded
The query of IT funding stays one of the vital sensible challenges for CIOs, particularly when executives name for better AI funding. The problem is that the know-how itself — together with enabling infrastructure and information pipelines — is just one part. Organizations should additionally funds for course of redesign, information administration reform and evolving governance. These areas, subsequently, every require their very own investments.
But Van de Maele argued that many enterprises underestimate these necessities, focusing closely on fashions and infrastructure whereas treating supporting capabilities as secondary.
“The most important false impression is that the AI funds is the mannequin and compute funds,” he mentioned. “Leaders image the spend as fashions, infrastructure, and expertise, and deal with information and governance as overhead, when it is actually the reverse.”
Addressing this hole has led some groups to get artistic. Whereas some organizations might create devoted AI budgets, business observers counsel many are reallocating current assets throughout the know-how portfolio.
“Most AI budgets are usually not new cash; they’re previous cash with a extra trendy job description,” Ratcliffe mentioned.
He described what he known as “funds laundering,” the place current automation, analytics or modernization tasks are reframed as AI initiatives as a result of the AI label could make funding simpler to safe. This will likely work within the brief time period, however Ratcliffe argued {that a} extra sustainable strategy entails shared possession: i.e., central IT funding reusable capabilities reminiscent of information, safety and governance, whereas enterprise items fund particular AI purposes tied to measurable outcomes.
A extra disciplined strategy to portfolio selections
In the end, the growing significance of AI doesn’t take away the elemental problem of IT portfolio administration: deciding the place restricted assets will create the best enterprise worth. If something, AI raises the stakes of these selections by creating new funding alternatives — whereas concurrently growing the significance of the underlying capabilities required to assist them.
Happily, CIOs needs to be ready for this calculation. Ratcliffe recommends utilizing the identical decision-making matrix that is been in place for years, describing it as “equally legitimate right now” and arguing that “the truth that AI is now an merchandise on the agenda mustn’t change it.” His model of this entails evaluating initiatives based mostly on enterprise worth, time to worth, strategic dependency, threat of deferral and reversibility.
Murray mentioned the agency takes an analogous strategy with its automation scorecard, which evaluates tasks throughout 4 components: monetary influence, information readiness, deployment velocity and the creation of shared capabilities.
These approaches present alternative ways of trying on the similar problem: understanding how particular person know-how investments contribute to broader enterprise capabilities.
