Artificial Intelligence 06.09.2026

Gartner: 85 percent want higher AI budgets

4 min read

The functional leaders of large companies want to spend more money on AI in 2026, even though one in nine does not know what their department paid for it in 2025. That is the core finding of a Gartner survey among 1,303 executives. Eleven percent do not know their function’s AI spending for 2025 and therefore have no basis for 2026.

Key takeaways

85 percent of functional leaders want to spend more on AI in 2026 than in 2025. Eleven percent do not know their function’s AI spending for the previous year.
Only 22 percent of organisations have scaled AI across several departments or work in an AI-centred way. Gartner does not report separately how far the remaining 78 percent have got.
In Gartner’s strongest measurement group, 81 percent of initiatives were in the black. The weakest organisations could not even state a return for 29 percent of their initiatives.
In IT, the most popular and the most profitable initiatives diverge. Threat detection and service desk are the most common initiatives; those for cost optimisation most often achieve positive returns.

What is AI scaling? Gartner uses the term when the same data, permissions, interfaces and responsibilities apply across several business units. An assistant in a single department does not yet amount to scaling. An organisation counts as AI-centred when AI sits at the core of its business model.

More money, even though one in nine does not know last year’s cost

From January to April 2026, Gartner surveyed just over 1,300 functional leaders: the heads of procurement, HR, finance, sales, IT and other departments at companies with annual revenue of around 45 million euros or more. The results were published on 1 September. 85 percent want to spend more on AI in 2026 than in the previous year. At the same time, one in nine says they do not know what their function spent on it in 2025.

The sums at stake are not small. According to Gartner, AI accounts for around twelve percent of functional budgets on average. In the announcement, Tina Nunno, Distinguished Vice President at Gartner, warns that missing financial transparency raises the risk while spending grows. Without measurement against business outcomes, funds risk being wasted.

Only 22 percent have organised AI the same way across departments

Only 22 percent of the surveyed organisations have scaled AI across several business units or pursue an AI-centred approach. Gartner counts both groups in one figure. Anyone reading the 22 percent as company-wide deployment overstates where things stand.

Three quarters of respondents name productivity as the most important goal of their AI spending. Around a third of the AI budget sits there. The money therefore flows to where the benefit is hardest to see in the profit and loss account. Working faster is one thing. A line of its own in that account is another.

Popular does not mean profitable

IT shows this most clearly. Just over half of IT leaders pursue initiatives for threat detection and for automating the service desk, both projects that protect and relieve the load. Positive returns, by contrast, are most often reported by respondents running initiatives to optimise IT estates and costs, 40 percent of them. The popular projects secure operations, while the profitable ones save money.

Gartner grades respondents by the quality of their measurement. The strongest group tracks returns continuously, manages AI as a portfolio and ends weak initiatives; it reported positive returns on 81 percent of its initiatives. The weakest group could not even state a return for 29 percent of its initiatives. Gartner ties budget reallocation to existing return figures; those are exactly what this group lacks.

The sample starts above the typical mid-sized company

The revenue threshold of around 45 million euros is far higher than the annual revenue of a typical mid-sized company in the KfW Mittelstandspanel. The finding therefore does not apply directly to smaller firms. If anything, the same problem is more acute in smaller firms: cost visibility is weaker because AI licences often disappear into departmental budgets. There is less room for costly purchasing mistakes.

Gartner leaves open who will own cost visibility in future: the business department that orders the licences; the IT function that runs them; the finance team that pays for them. As long as nobody decides that, approving the next tranche remains what it is today for eleven percent of respondents: a guess.

Frequently asked questions

Is the 22 percent figure purely a scaling rate?

No. Gartner adds organisations with AI across several business units and organisations with an AI-centred approach. Gartner does not break down the 22 percent into organisations that have scaled AI and those taking an AI-centred approach.

Does the survey cover the traditional SME sector?

No. Gartner surveyed organisations whose 2025 annual revenue was clearly above the level in the KfW Mittelstandspanel. The same problem is often more acute in smaller firms.

Why does productivity tie up so much budget?

Because three quarters of respondents name productivity as their most important goal and around 30 percent of the AI budget sits there. The benefit shows up as time saved, rarely as a line of its own in the profit and loss account.