A global study of 2,500 chief executives has revealed a clear divide at the heart of enterprise AI adoption: confidence in the technology is at an all-time high, yet so is the anxiety that organisations are not moving fast enough to capitalise on it.
The research, published by Cisco, draws on responses from CEOs across 23 countries and marks the second consecutive year the networking giant has surveyed senior leaders on their attitudes toward artificial intelligence.
The picture that emerges is one of accelerating optimism running headlong into stubborn operational reality.
From Scepticism to FOMO
Ninety-one percent of respondents said they are more optimistic about AI’s potential than they were a year ago – a striking figure that suggests widespread scepticism among the C-suite has largely evaporated.
But optimism has brought a new problem with it: fear of missing out.
Sixty-five percent of CEOs now worry they are underinvesting in AI and missing opportunities as a result.
That is up sharply from 53 percent who said the same in last year’s survey – a 12-percentage-point jump in 12 months that reflects how quickly the competitive stakes around AI have risen in boardrooms globally.
More than two thirds (69 percent) now consider AI adoption mandatory for modern business – not a competitive advantage, not a nice-to-have, but a baseline requirement for staying relevant.
The Execution Gap
Despite the bullish sentiment, the research identifies three concrete barriers that are slowing organisations down and creating a widening gap between leadership ambition and ground-level readiness.
The first is infrastructure. More than half of CEOs (53 percent) fear that infrastructure limitations will hold back their AI ambitions, and upgrading infrastructure to handle AI workloads ranks as the single top organisational priority for 2026.
Closely behind it is upskilling teams for AI readiness – a sign that the people and systems problem is being felt simultaneously.
Those concerns are backed up by separate data from Cisco’s 2026 AI Readiness Index, which surveyed more than 8,000 IT leaders worldwide. That research found that fewer than a quarter (22 percent) rate their network as optimal for AI workloads – a sobering gap between what CEOs want to do and what their technical foundations can actually support.
The second barrier is trust and security. As organisations move toward deploying AI agents – autonomous systems that can execute tasks independently – security and control of those systems has emerged as the leading concern among CEOs.
Deploying AI agents to work alongside employees ranks among CEOs’ top three priorities for 2026, yet the AI Readiness Index found that only 31 percent of organisations feel equipped to secure and control such systems. The appetite to deploy and the ability to do so safely are not yet aligned.
The third barrier is data. One in three CEOs (34 percent) cited fragmented, inaccessible, or poor-quality data as the single greatest obstacle to AI progress – more than any other challenge in the survey.
The AI Readiness Index reinforces this, finding that only 19 percent of organisations have fully centralised, AI-accessible data infrastructure in place. Without clean, connected data, even the best AI tooling struggles to deliver meaningful results.
Humans Stay in the Loop – For Now
One of the more nuanced findings concerns the future relationship between AI systems and human workers.
Almost every CEO surveyed expects AI to play a significant role in business operations by 2030, but the overwhelming majority – 72% – envision a future in which AI only supports or executes work under human direction, judgment, or governance.
The reasons are a blend of the practical and the philosophical: keeping AI systems secure, ensuring productive human-AI collaboration, and navigating the ethics of autonomous decision-making. As one CEO put it plainly in the survey:
“Move fast with AI, but anchor decisions in human values – always.”
The Knowledge Gap Is Closing
One area of genuine progress stands out. The share of CEOs who said their own understanding of AI had held them back in boardroom discussions dropped from 74 to 47 percent over the past year. Those who said limited AI knowledge had blocked informed decision-making fell similarly, from 74 to 49 percent. The learning curve, it seems, has been steep but effective.
Regional differences also surfaced. European CEOs, the research found, are less preoccupied with speed and more focused on getting AI adoption right – placing greater emphasis on trust, transparency, and ethical alignment. One European CEO offered a note of caution that reflects the mood: “Avoid rushing AI into decisions without transparency. Trust is harder to regain than efficiency.”
What It All Means
Taken together, the findings sketch a C-suite that has moved decisively past the “will this work?” phase of AI adoption and into something more uncomfortable: the recognition that falling behind is a real risk, but that the infrastructure, data, and security foundations needed to move faster are not yet in place.
The organisations that manage to close all three gaps – network readiness, data centralisation, and AI security – are already pulling ahead, according to Cisco’s index data. For everyone else, the race is on.
