Two years of the IT Spending Pulse have read like a barometer of the broader economy. Confidence ran high in December 2024, driven by ambitious plans for growth and AI. It was replaced by austerity by April 2025, when newly announced tariffs and recession fears pushed cost control level with growth as the top priority. By mid-2025, the picture had steadied into a cautious rebalancing, with buyers edging back toward strategic initiatives.
Our latest survey, conducted in May 2026, marks the end of that oscillation. (See the sidebar, “Survey Methodology.”) Confidence has not merely returned, it has surpassed where it stood before the tariff shock. Buyers are no longer spreading their bets. Instead, they are concentrating spending on a single priority—AI—and paying for it both by expanding their budgets and by pulling back in other areas.
Survey Methodology
The result is an IT agenda defined less by recovery than by conviction. Buyers know what they want to fund, and they are reorganizing their portfolios around it.
Confidence Returns, and Budgets Follow
The clearest signal of the shift is the budget itself. IT leaders now expect overall spending to grow 5.8% in 2026 compared with 2025—the strongest outlook the survey has recorded and, as of mid-2025, a full 2.2 percentage points higher than the 3.6% they were planning for 2026. (See Exhibit 1.)
That figure is notable not only for its size but its trajectory. Budget expectations bottomed out at 2.4% in the immediate aftermath of the tariff announcements in early 2025. Firms then spent through the second half of the year more freely than they had planned, bringing actual 2025 growth to 4.6%—back in line with their original, pretariff optimism. The 2026 outlook now clears that bar comfortably.
Underlying the numbers is a steady climb in AI maturity. Since our mid-2025 survey, the share of firms describing themselves as high-maturity AI adopters—those with AI embedded in enterprise strategy and scaled across multiple functions—has risen by 10.8 percentage points, while medium-maturity firms grew by 4.7 points. As capabilities improve, so does the willingness to invest behind them.
As capabilities improve, so does the willingness to invest behind them.
AI Draws Spending Away from Everything Else
Confidence alone does not explain how buyers are spending. What distinguishes this edition is the sheer concentration of investment.
Across the full technology stack, buyers expect net spending to decline in most product categories over the next 12 months. The exceptions are telling: 66% of respondents expect to increase spending on AI and machine learning (ML), including GenAI and AI agents—more than any other category and up 21 points from mid-2025. The next largest gainers are the enablers that AI depends on for running AI applications at volume and keeping data environments secure: cloud services (up 11 points) and security infrastructure (up 3 points). (See Exhibit 2.)
Almost all other tech investments are losing ground. CRM and ERP, which had shown brief momentum in earlier surveys, both turned negative again as companies reprioritized around building AI capability rather than modernizing systems of record.
IT services saw a meaningful reversal, swinging down 19 points—a sharp change in sentiment from H2 2025. That shift reflects reprioritized budget growth, not retrenchment: buyers aren’t pulling work back in-house or cutting ties with providers. Over the past three years, 44% of respondents outsourced somewhat or much more, versus just 4% who brought meaningfully more in-house, and outsourcing levels have held steady or grown for the large majority of buyers. (See Exhibit 3)
When asked where they intend to expand versus consolidate suppliers, buyers pointed to net consolidation in every category but one. AI and ML stand alone, with a net 34% of buyers planning to expand their roster of providers. Cloud, analytics, and security see continued spending but through a stable or shrinking set of vendors—a pragmatic move to simplify portfolios. (See Exhibit 4.)
From Experimentation to Disciplined Execution
If earlier surveys captured the enthusiasm of broad AI experimentation, this one captures its discipline. Buyers are narrowing where they apply GenAI and AI agents, concentrating on a smaller set of high-impact use cases—analytics, customer service, security and risk prevention, and internal communications chief among them. The average number of use cases per respondent has actually fallen across every maturity tier, even as adoption of those priority use cases has climbed. The strategy has shifted from spraying pilots across the organization to executing on what is core. (See Exhibit 5.)
That discipline is showing up in returns. Weighted-average ROI measured to date on GenAI and AI agent deployments has risen to 13.8%, up 2.6 points from 11.2% in our mid-2025 survey. Much of that gain reflects a changing mix of respondents—more firms now sit in the higher-maturity segments that report stronger returns—itself a signal of the market moving from promise to payoff.
Buyers are narrowing where they apply GenAI and AI agents, concentrating on a smaller set of high-impact use cases.
A Widening Gap Between Leaders and Laggards
Beneath the aggregate optimism runs a familiar divide, and it is growing. Firms with high AI maturity report measured ROI of roughly 19%, more than double the 8% to 9% reported by little-to-no- and low-maturity adopters. Early experimentation, particularly among less mature firms, still produces a meaningful share of negative or unclear returns; discipline and scale are what convert investment into value.
Maturity confers advantages beyond returns. Higher-maturity firms maintain far better visibility into their AI and token costs, often maintaining team- and user-level visibility. (See Exhibit 6.)
Counterintuitively, these firms use that visibility to be more permissive, adopting liberal token strategies that encourage developers to use AI tools freely rather than throttling them. They can afford to: their power users—the top 10% of AI adopters—reportedly generate outsized productivity without consuming a disproportionate share of spending, a signature of effective token usage. Less-mature firms reportedly lack that visibility and tend toward harder limits and more cautious, tactical management.
The takeaway is that the leaders are not simply spending more. They are actively tracking where spending is going and making active decisions to ensure tokens are spent effectively.
Concentrated Optimism
For two years, IT buyers oscillated between the exuberance of late 2024 and the tariff-driven austerity of early 2025.
Budgets are growing again, and above pre-tariff levels. But the more important story is not how much buyers are spending; it is how narrowly. AI has become the organizing principle of the IT agenda, funded by fresh budget growth and by pull-backs in other areas, with cloud and security carried along as its essential enablers.
The firms positioned to win with AI will be the ones disciplined enough to concentrate their bets on what’s truly core to the business—and rigorous enough to track how that spending is paying off.
