Gartner has issued at least four dated AI forecasts that reach into 2026 through 2030: 40 percent of enterprise applications with task-specific AI agents by the end of 2026, more than 40 percent of agentic AI projects canceled by the end of 2027, 15 percent of day-to-day work decisions made autonomously by 2028, and more than USD 3.5 trillion of AI infrastructure spending exposed to impairment by the end of 2030. As of October 27, 2026, none of those deadlines has passed, so none can be scored yet. Each is an analyst forecast with a stated reason, not a measured result.
Key takeaways
- The agentic AI cancellation forecast was issued on June 25, 2025, and its deadline is the end of 2027.
- The forecast that 40 percent of enterprise apps will include task-specific agents was issued on August 26, 2025, with a deadline at the end of 2026.
- The infrastructure impairment warning dated October 14, 2026 describes spending that could be at risk, not confirmed losses.
- Gartner's own survey figure behind the agentic forecast came from a webinar poll of 3,412 attendees, which is not a representative sample.
The forecasts in one table
The table lists each prediction as Gartner worded it in its public release, the issue date, the deadline, and what a later check would need. The last column is an illustrative reading aid.
| Forecast | Issued | Deadline | What a check would need |
|---|---|---|---|
| 40 percent of enterprise applications include task-specific AI agents, up from under 5 percent | August 26, 2025 | End of 2026 | A count of enterprise apps with agents, and a definition of "task-specific" |
| More than 40 percent of agentic AI projects canceled | June 25, 2025 | End of 2027 | A list of projects started and stopped, and a definition of "agentic" |
| At least 15 percent of day-to-day work decisions made autonomously, up from 0 percent in 2024 | June 25, 2025 | 2028 | Workplace data on which decisions agents made without a person |
| More than USD 3.5 trillion of planned AI infrastructure spending exposed to impairment | October 14, 2026 | End of 2030 | Company write-downs and revenue figures against the stated shortfall |
The June 2025 agentic AI forecast
Gartner's release of June 25, 2025 attributed the expected cancellations to escalating costs, unclear business value, and inadequate risk controls. Senior director analyst Anushree Verma said most agentic AI projects at that point were early-stage experiments or proofs of concept. The release also described "agent washing," in which vendors relabel assistants, robotic process automation, and chatbots as agentic, and estimated that only about 130 of the thousands of agentic AI vendors are genuine.
The survey behind the investment picture was a January 2025 poll of 3,412 webinar attendees. It found that 19 percent had made significant investments in agentic AI, 42 percent conservative investments, and 8 percent none, with 31 percent waiting or unsure. A poll of people who chose to attend a webinar says little about all organizations.
The August 2025 forecast for enterprise apps
The August 26, 2025 release predicted that 40 percent of enterprise applications would include task-specific AI agents by the end of 2026, up from under 5 percent at the time. Coverage of the release also reported a five-stage path: assistants embedded in apps in 2025, task-specific agents in 2026, collaborative agents within apps in 2027, cross-application agent ecosystems in 2028, and by 2029 a workforce in which about half of staff are trained to build or oversee agents. The release credits the same analyst, Anushree Verma, and cites a possible share of about 30 percent of enterprise application software revenue by 2035, a figure far outside any checkable window.
The October 2026 infrastructure warning
The newest release, dated October 14, 2026, comes from senior director analyst Will Sommer. Gartner calculates that covering announced capital spending, financing costs, and a 12 percent return would take USD 8.5 trillion in cumulative revenue through 2030, while it projects no more than USD 3.6 trillion in enterprise and consumer AI revenue tied to those data centers. It expects AI providers to spend nearly USD 1 trillion on infrastructure in 2026 and about USD 6 trillion more from 2027 through 2030, and it expects more than 80 percent of neocloud providers, the GPU-rental firms below the large cloud companies, to stop operating independently by 2030.
The figure is exposure, meaning spending that could be impaired if the revenue does not appear. Gartner says large cloud providers are cushioned by diversified revenue.
Reading a Gartner prediction
- Note the verb. "Will" statements and "could be exposed" statements carry different strength.
- Note the baseline. A rise "from under 5 percent" depends on how the starting point was counted.
- Note the definition. Terms such as agentic, task-specific, and autonomous are Gartner's, and the public release may not give the full wording.
- Note the date. A prediction made in 2025 reflects 2025 data.
- Look for a published follow-up. No Gartner outcome report for these four forecasts was found in the sources reviewed.
FAQ
Have any of these Gartner AI forecasts come true?
None can be judged yet. The earliest deadline is the end of 2026, and no result data had been published by October 27, 2026.
Does the USD 3.5 trillion figure mean investors will lose that amount?
No. It is Gartner's estimate of spending that could face impairment under its revenue assumptions. The release does not state a loss forecast of that size.
What is agent washing?
Gartner uses the term for vendors that rebrand existing products as agentic without substantial agentic capability.
Where can the original wording be read?
Gartner posts press releases in its newsroom. The full research notes behind the releases are available to clients.
Related reading
- How to read the Gartner Hype Cycle for AI
- Which 2025 AI predictions can be checked
- Trustworthy AI trends and challenges
- Key topics in artificial intelligence
More in AI Tools and Industry.
Last verified: October 2026
Sources
- Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027, Gartner press release, June 25, 2025
- Gartner Predicts 40% of Enterprise Apps Will Feature Task-Specific AI Agents by 2026, Up from Less Than 5% in 2025, Gartner press release, August 26, 2025
- Gartner Says Over USD 3.5 Trillion of AI Infrastructure Spending Could Be Exposed to Impairment by the End of 2030, Gartner press release, October 14, 2026
- Gartner predicts sharp rise in AI agents within enterprise applications, Consumer Goods Technology, September 11, 2025