AI spending rises; only one in five firms has scaled it

26.09.2026

4 min read

Only 22 percent of organisations have scaled AI across multiple business units or adopted an AI-first approach, while 85 percent of functional leaders plan to increase AI spending. Gartner proposes an AI value portfolio with ongoing reviews, reallocation and the termination of underperforming initiatives.

Key Takeaways

Most organisations have yet to scale AI. Only 22 percent of organisations have scaled AI across multiple business units or adopted an AI-first approach, while 85 percent of functional leaders plan higher spending for 2026.
High performers measure ROI continuously. They report positive returns from 81 percent of their AI initiatives. Low performers cannot say whether 29 percent of their initiatives are paying off.
Willingness to invest remains high. 94 percent of CEOs intend to keep investing even if it does not pay off in 2026. Only 37 percent of McKinsey respondents see a contribution to EBIT.

Only one in five organisations scales AI

According to a Gartner survey published in early September, only 22 percent of organisations have successfully scaled AI across multiple business units or adopted an AI-first approach. Between January and April 2026, 1,303 respondents from organisations with annual revenue equivalent to at least 44 million euros took part.

At the same time, 85 percent of functional leaders plan to increase their AI budgets for 2026. In 2025, their functions spent an average of 12 percent of their budgets on AI.

In around 11 percent of organisations, leaders say they do not even know how much their function spent on AI in 2025. Most respondents name productivity as their main AI objective, accounting for an average of around 30 percent of AI spending.

What is an AI value portfolio? An AI value portfolio is a management approach in which an organisation treats all its AI initiatives as a connected portfolio and regularly reviews each initiative’s contribution to business value. Funds are reallocated and initiatives without sufficient returns are ended on that basis. Gartner describes this approach as a characteristic of high performers, who continuously measure the ROI of their AI initiatives.

High performers keep checking the numbers

Gartner categorises the organisations surveyed by how they manage AI spending. High performers continuously measure the ROI of their AI initiatives, manage AI as a value portfolio and regularly review project performance to reallocate funds or end initiatives.

This group reports positive returns from 81 percent of its AI initiatives. Low performers cannot even say whether 29 percent of their initiatives are paying off.

Tina Nunno, Distinguished Vice President and Gartner Fellow, says CEOs and CIOs need to establish which AI use cases actually deliver financial returns. Only then can they define clear objectives for each function and each executive.

CEOs take responsibility for AI

For its BCG AI Radar 2026, published in January, BCG surveyed 2,360 executives, including 640 CEOs. 72 percent of those CEOs describe themselves as the main AI decision-maker in their organisation, twice as many as in the previous year.

Half of the CEOs surveyed believe their job security depends on getting AI investment and strategy right. According to the survey, AI investment as a share of revenue is expected to rise from around 0.8 percent in 2025 to around 1.7 percent in 2026.

94 percent of CEOs say they will continue investing even if it does not pay off in 2026. If AI investment fails to deliver the desired financial impact over the next twelve months, only 6 percent of the companies surveyed intend to cut back.

In that scenario, 70 percent of the executives surveyed plan to stay the course or make strategic adjustments. 24 percent would increase funding.

Limited impact on EBIT

At an individual level, many respondents report a noticeable impact from AI. In McKinsey’s State of AI survey published in August, 80 percent of respondents report higher individual productivity and 50 percent report better decisions.

Respondents report less impact on company earnings. Only 37 percent report a contribution to EBIT, virtually the same proportion as in the previous year.

Around 6 percent of respondents qualify as high performers under McKinsey’s definition: they attribute at least 5 percent of EBIT to AI and report substantial impact. This group has not grown since the previous year.

Nevertheless, 60 percent of respondents expect higher AI investment in the coming year. At the same time, around 20 percent say AI operating costs, including token costs, have already limited their organisation’s use of AI.

AI returns in IT

Gartner breaks down individual use cases for IT. The most widely pursued are cybersecurity detection and response and IT service desk automation, each at 54 percent.

The highest proportions of positive returns are reported instead for intelligent IT asset and cost optimisation, at 40 percent, and synthetic data generation, at 28 percent. The two most widely pursued use cases are outside the three areas with the highest proportions of positive returns.

Gartner had already published a forecast for agentic AI projects in June 2025. It predicted that more than 40 percent of these projects would be cancelled by the end of 2027 because of rising costs, unclear business value or inadequate risk controls. Anushree Verma, Senior Director Analyst at Gartner, described most current projects as early experiments driven by hype and often misapplied.

Frequently Asked Questions

How many organisations successfully scale AI across business units?

Only 22 percent of organisations have scaled AI across multiple business units or adopted an AI-first approach. That was the finding of a Gartner survey of 1,303 participants from organisations with annual revenue equivalent to at least 44 million euros.

What distinguishes high performers in their approach to AI?

They continuously measure the ROI of their initiatives, manage AI as a value portfolio and regularly review project performance. They reallocate funds or end initiatives on that basis. This group reports positive returns from 81 percent of its AI initiatives.

How many agentic AI projects does Gartner expect to be cancelled?

Gartner forecasts that more than 40 percent of agentic AI projects will be cancelled by the end of 2027. The analyst firm cites rising costs, unclear business value and inadequate risk controls.

Image source: AI-generated (September 2026)

Translated from the German original using artificial intelligence. The German version is authoritative.