Agentic AI's $290bn projection masks a governance deficit
- The agentic AI market is projected to grow from USD 8.19 billion in 2025 to USD 290.62 billion by 2035 at a 42.88 per cent compound annual growth rate.
- McKinsey data shows 88 per cent of companies use AI in some area, but only 23 per cent have scaled agentic AI and no business function has reached 10 per cent adoption at scale.
- Agents complete roughly 66 per cent of tasks on the OSWorld benchmark, and only 13 per cent of organisations believe they have adequate governance for large-scale agent deployment.
- Gartner projects a single Fortune 500 enterprise could run more than 150,000 agents by 2028, up from fewer than 15 in 2025.
The global agentic AI market is forecast to grow from USD 8.19 billion in 2025 to USD 290.62 billion by 2035, according to Analytics Insight’s report on the Global Agentic AI Market 2026 forecast. The projection implies a compound annual growth rate of 42.88%, driven by enterprise investment and the emergence of multi-agent systems capable of planning, reasoning and executing tasks with limited human oversight.
Such headline figures deserve context. The forecast draws on industry surveys and analyst modelling, not audited financial results. Stanford HAI’s 2026 AI Index puts global corporate AI spending at USD 581.7 billion, up 130 per cent year on year, while a Deloitte survey of 3,235 executives across 24 countries found 74 per cent expect to be running AI agents by 2027. Those numbers indicate genuine organisational intent. They do not, on their own, validate a specific decade-end market size.
The gap between ambition and deployment
The more revealing data sits inside the same report. McKinsey figures cited in the analysis show that 88 per cent of companies use AI in at least one area, yet only 23 per cent have scaled agentic AI across their operations. No single business function has reached 10 per cent adoption at scale. On the OSWorld benchmark, agents complete roughly 66 per cent of tasks successfully, leaving one in three attempts failing. These are not trivial failure rates for systems being positioned as autonomous coordinators of complex workflows.
Gartner projects that a single Fortune 500 enterprise could run more than 150,000 agents by 2028, up from fewer than 15 in 2025. That trajectory, if it materialises, would represent a wholesale shift in how organisations delegate decision-making to software. Yet only 13 per cent of organisations surveyed believe they have adequate governance in place to manage such systems. The mismatch between projected scale and preparedness is the actual story: not whether the market will be large, but whether institutions can absorb it responsibly.
For universities and schools evaluating agentic AI tools, the reliability data matters most. An agent that fails one in three tasks may be acceptable in low-stakes automation; it is a different proposition when mediating student feedback, flagging academic misconduct or managing assessment workflows. Governance frameworks that exist on paper rarely survive contact with tools operating at speeds and scales their designers did not fully test.
What remains unresolved
The report expects Asia Pacific to outpace North America in growth rate, with a projected compound annual growth rate of 47.40 per cent, and hybrid deployment models to expand fastest among deployment types. Those regional and structural details matter for procurement and policy, but they do not answer the harder question: who is accountable when an autonomous agent makes a consequential error, and how quickly can that accountability be enforced?
The projection tells you money is flowing. The adoption and reliability data tell you the infrastructure to spend it wisely is not yet in place.