AI Infrastructure Projects Often Fail to Deliver Expected Returns

Gartner research finds over 20% of AI initiatives in IT operations fall short of expectations

Apr. 9, 2026 at 4:20pm

Got story updates? Submit your updates here. ›

A highly detailed, glowing 3D illustration of a complex circuit board or server infrastructure, with neon cyan and magenta lights illuminating the various components, conceptually representing the intricate nature of modern AI systems and data centers.As organizations increasingly turn to AI to streamline operations, new research suggests many projects fail to deliver the expected returns, highlighting the need for realistic expectations and careful planning.NYC Today

Recent research from Gartner suggests that AI infrastructure and operations (I&O) projects frequently fail to live up to their promised benefits, with one in five such initiatives resulting in failure. The research director at Gartner cites unrealistic expectations about AI's capabilities as a key factor, as organizations assume the technology will immediately automate complex tasks, cut costs, or fix long-standing issues. In reality, challenges like skill gaps, poor data quality, and limited data access often undermine the success of these AI projects.

Why it matters

As more organizations look to incorporate AI into their workflows, this research highlights the importance of setting realistic expectations and properly scoping AI initiatives. Overconfidence in AI's abilities can lead to costly failures, making it critical for companies to carefully evaluate the technology's fit within their operations and infrastructure before investing significant resources.

The details

Gartner's survey of 782 I&O managers found that 57% had experienced at least one failed attempt to implement AI in their work. The most common issues arose from AI-led workflow management and automated security threat remediation, with 38% of managers citing skill gaps, data quality problems, and limited data access as the primary causes of failure. However, the research also found that 53% of managers reported success in applying more mature generative AI (GenAI) to IT service management and cloud operations.

  • Gartner surveyed 782 I&O managers at the end of 2025.

The players

Melanie Freeze

Research director at Gartner who led the study on AI infrastructure project failures.

Got photos? Submit your photos here. ›

What they’re saying

“The 20 percent failure rate is largely driven by AI initiatives that are either overly ambitious or poorly scoped. AI that doesn't fit into the organization's operations simply can't deliver [a return on investment].”

— Melanie Freeze, Research director, Gartner

The takeaway

This research underscores the need for organizations to approach AI integration with a clear-eyed view of the technology's capabilities and limitations. By setting realistic expectations and carefully aligning AI projects with their operational needs, companies can improve their chances of realizing the promised benefits of AI and avoiding costly failures.