Global spending on AI is forecast to reach $2.59 trillion in 2026, a 47% increase year-over-year, according to new figures from Gartner published today. The headline number is striking but much of this growth is being driven by technology vendors and hyperscalers, not by enterprises. This may well be explained by the path for businesses to get a return on AI investments remaining relatively unclear.
Infrastructure is doing the heavy lifting
An earlier Gartner forecast, published in January, placed global AI spending at $2.52 trillion for 2026, up 44% on the prior year. Even then, analysts could see what was behind the numbers. AI infrastructure was expected to account for over 45% of all AI spending, with investment in AI-optimised servers alone rising 49%. That picture has not fundamentally changed in today's revision. The infrastructure build-out continues, and so does the spending attached to it.
John-David Lovelock, Distinguished VP Analyst at Gartner, described 2026 as an “inflection year” in which enterprises will start to up their spending, while also providing a frank assessment of where most actually stand: "Currently, organisations show limited appetite for using AI to drive disruptive enterprise change. Instead, they favour tactical AI initiatives with incremental improvements in efficiency and productivity."
‘Trough of Disillusionment’
Gartner's January forecast placed AI firmly in the "trough of disillusionment" for 2026, referring to a phase within technology hype cycles where “interest wanes as experiments and implementations fail to deliver”. Companies, at this time, look for improved ROI predictability before considering enterprise-scale adoption. It seems we are yet to reach the “slope of enlightenment” and “plateau of productivity”, which lie ahead.
New research from Sinch offers a ground-level view of why that predictability remains elusive. The firm's global study of 2,527 senior decision-makers found that 73% of enterprises had shut down or rolled back a live AI customer agent following a governance failure. The figure does not mean enterprises are walking away from AI. It does suggest, however, they significantly underestimated what production-grade deployment requires.
The Cost of Doing it Right
Part of what they underestimated is the cost of governance itself. The Sinch data shows that 84% of AI engineering teams are spending at least half their time building and maintaining safety infrastructure rather than improving the customer experience. This "guardrail tax", as Sinch describes it, compounds over time as organisations add agents, channels and use cases.
As regulatory pressure mounts, understanding what the EU AI Act means for CX leaders is no longer optional, and the true compliance costs are beginning to show. Putting this into CX context specifically, Gartner predicted that by 2030, the cost per resolution for generative AI in customer service will exceed $3, higher than many B2C offshore human agents. For now, any assumption that AI is a cheaper alternative to human workers needs to be challenged.
Spending Does Not Equal Returns (Yet)
The overall direction of AI investment is not in question, with Gartner's hype cycle providing useful perspective for what comes next. As we currently find ourselves in the ‘trough of disillusionment’, however, the difference between vendor ambition and enterprise reality is at its broadest. Spend continues to climb, but the path to the ‘plateau of productivity’ runs directly through the governance, data, and organisational challenges that have caused so many deployments to stumble. Understanding why CX AI projects fail to deliver is less a cautionary tale than a practical prerequisite for what comes next.