Heads of infrastructure and IT operations need to focus their AI efforts carefully if they want to accelerate autonomous operations and support enterprise-scale AI initiatives, Gartner has warned.

In a survey of 272 heads of infrastructure and IT operations (I&O), the company found that 58% believe scaling proof of AI concepts to I&O adoption will be an important priority in the next 12 months.

And, said the firm, there are five AI trends worth adopting if firms want to improve operational efficiency with AI and optimize costs.

“AI is creating one of the most significant shifts the I&O function has ever faced,” said Jonathan Forest, VP analyst at Gartner. “Rather than trying to address every aspect of AI simultaneously, heads of I&O should focus on a small number of strategic investments that deliver meaningful impact.”

The five AI trends worth adopting for infrastructure and IT operations

The first, unsurprisingly, is agentic AI, used for tasks such as incident triage, change validation, and remediation. And while virtually no organizations are currently using AI agents with minimal human involvement in IT operations, Gartner expects half to be doing so by 2030.

“AI agents are driving a shift from reactive, human-in-the-loop operations to proactive, human-on-the-loop operations,” said Forest. “Heads of I&O who fail to restructure their organization around agentic AI risk higher costs, poor operational agility and performance, and ultimately, irrelevance.”

The move to agentic AI, though, is creating challenges in terms of token consumption – which is rocketing and likely to lead to unforeseen costs.

Through 2027, Gartner predicts that enterprises implementing AI agents in IT operations using uncapped variable pricing models tied to token usage will overrun their initial cost expectations by more than 50%, compared with those using predictable subscription and bundled pricing models.

The only answer, said the firm, is a strict “no telemetry, no deployment” policy to eliminate blind spots in AI spending and token usage.

Another major trend worth watching, said Gartner, is the way AI inference is driving the rise of hybrid AI infrastructure, driven mainly by privacy, cost, performance, and data sovereignty requirements.

Gartner predicted that a quarter of enterprises will have merged their network operations center (NOC) and security operations center (SOC) by 2030, up from just one in 20 now. And this, it said, will see total spend on tooling and staff resources falling by a quarter.

Finally, AI deployments across cloud, data center and edge environments are increasing the need for a coordinated approach to managing AI infrastructure, governance, security and operations. As a result, said Gartner, I&O is evolving from a traditional gatekeeper role to the orchestrator of the enterprise AI platform, enabling the business to deploy and scale AI reliably, securely and in a cost-effective way.

And, Gartner warned, over the next four years, organizations without a strong I&O function driving AI infrastructure design, architecture and support will be more than twice as likely to fail in their AI initiatives.

“The high cost of AI infrastructure leaves little room for error, placing I&O at the center of enterprise AI success,” said Forest. “Heads of I&O must elevate their teams’ capabilities, reduce technical debt, and shift from reactive operations to proactively enabling enterprise AI.”