Rethinking IT Budgets in the Age of AI

Enterprise technology spending is shifting as organizations divert capital toward the infrastructure required to support artificial intelligence workloads. IBM reported weaker-than-expected mainframe performance and delayed large deals, while pointing to customers redirecting capital expenditure toward servers, storage and memory. While the company saw a 7% decline in direct infrastructure revenue, its distributed infrastructure revenue increased by 37%, reflecting growing demand for the technology needed to support AI workloads.

Technology budgets are being strained

Despite strong growth in spending, this is not a “rising tide lifts all boats” market trend, according to Gartner. Worldwide IT spending is forecast to reach $6.37 trillion in 2026, up 14.2% from the previous year, with data center systems and IaaS among the fastest-growing segments. However, technology budgets are being strained by inflation, supply shortages, rising hardware and memory costs, AI funding initiatives and shifting priorities.

For CIOs, that creates a difficult portfolio management challenge. AI investment is accelerating, but it is competing for attention and resources alongside cybersecurity, modernization, infrastructure resilience and operational improvement initiatives that remain essential to running the enterprise. The challenge facing technology leaders is how to evaluate those competing priorities as AI becomes a larger part of the technology setting.

Leaders must distinguish between technology that merely keeps the organization operating and technology that genuinely improves its competitive position. AI has made the tolerance for undifferentiated IT expenditure considerably lower.

The pressure to demonstrate value

The growing focus on AI is increasing pressure on CIOs to demonstrate clear business value from technology investments across the portfolio. Simon Ratcliffe, fractional CIO at Freeman Clarke, said he believes AI has introduced a new strategic priority while existing operational responsibilities remain intact. Security, compliance, resilience, applications and infrastructure still require investment, but CIOs are now being asked to create funding capacity for AI initiatives whose long-term economics may still be developing.

Adrian Murray, founder and CEO of Fisent Technologies, a developer of enterprise GenAI workflow software, said he sees this as a broader shift in how organizations should think about all technology investments. He argued that AI initiatives need to be evaluated as part of broader enterprise capabilities.

The rise of AI forces a shift from managing technology as isolated, opportunistic projects to building a unified automation fabric that acts as core enterprise infrastructure. Determining which investments create reusable capabilities across the organization and which remain limited to individual use cases will be a primary challenge for CIOs in the months to come.

Protecting enterprise foundations

AI may be getting a lot of the attention, but its success depends on other, less flashy investments. As organizations expand AI adoption, attention is increasing on the underlying capabilities that determine whether those AI initiatives can succeed at scale. Felix Van de Maele, CEO and co-founder of enterprise data intelligence and governance platform Collibra, argues that organizations need to protect investments in areas such as data platforms, governance, security and systems of record because those capabilities provide the context AI systems require.

Ratcliffe said he sees a similar pattern in infrastructure spending. He argued that IBM’s results look “less like infrastructure being abandoned and more like expenditure being rapidly reordered around the infrastructure needed for AI.”

Money is moving toward compute, storage, data and AI-enabling capabilities, while expenditure that cannot demonstrate urgency, differentiation or measurable value is increasingly vulnerable.

The same dynamic applies beyond infrastructure. Murray pointed to API enablement, platform engineering, structured data pipelines, orchestration and observability as examples of capabilities that become increasingly important as organizations move AI from experimentation into production.

Managing the practical reality of funding

The question of IT funding remains one of the most practical challenges for CIOs, especially when executives call for greater AI investment. The challenge is that the technology itself — including enabling infrastructure and data pipelines — is only one component. Organizations must also budget for process redesign, data management reform and evolving governance. These areas, therefore, each require their own investments.

Yet Van de Maele argued that many enterprises underestimate those requirements, focusing heavily on models and infrastructure while treating supporting capabilities as secondary.

The biggest misconception is that the AI budget is the model and compute budget. Leaders picture the spend as models, infrastructure, and talent, and treat data and governance as overhead, when it’s really the reverse.

Addressing this gap has led some teams to get creative. While some organizations may create dedicated AI budgets, industry observers suggest many are reallocating existing resources across the technology portfolio.

Most AI budgets are not new money; they are old money with a more fashionable job description, Ratcliffe said. He described what he called “budget laundering,” where existing automation, analytics or modernization projects are reframed as AI initiatives because the AI label can make funding easier to secure.

This may work in the short term, but Ratcliffe argued that a more sustainable approach involves shared ownership: central IT funding reusable capabilities such as data, security and governance, while business units fund specific AI applications tied to measurable outcomes.

A disciplined approach to portfolio decisions

The increasing importance of AI does not remove the fundamental challenge of IT portfolio management: deciding where limited resources will create the greatest enterprise value. If anything, AI raises the stakes of those decisions by creating new investment opportunities — while simultaneously increasing the importance of the underlying capabilities required to support them.

Fortunately, CIOs should be prepared for this calculation. Ratcliffe recommends using the same decision-making matrix that’s been in place for years, describing it as “equally valid today” and arguing that “the fact that AI is now an item on the agenda should not change it.”

His version of this involves evaluating initiatives based on business value, time to value, strategic dependency, risk of deferral and reversibility.

Murray said the firm takes a similar approach with its automation scorecard, which evaluates projects across four factors: financial impact, data readiness, deployment speed and the creation of shared capabilities.

These approaches provide different ways of looking at the same challenge: understanding how individual technology investments contribute to broader business capabilities.