Gartner: Your AI Strategy Now Runs Through the Server Room
- By CDOTrends editors
- November 23, 2026

Is your company ready to let AI agents run IT operations with minimal human involvement? Half of all organizations will by 2030, Gartner predicts. In 2026, that figure is close to zero.
The strength of your infrastructure and operations (I&O) team matters, though. In its top AI trends in infrastructure and IT operations for 2027, released Oct. 26, Gartner said organizations without a strong I&O function will be over twice as likely to fail in their AI initiatives through 2030. That means that I&O, the team that runs servers, networks and cloud services, will be vital for AI strategy success.
In the same report, the analyst firm also expects companies that pay for agents by the token, with no spending cap, to blow their initial cost estimates by more than 50% through 2027.
The rising importance of the I&O function is also putting its leaders under strain. In a Gartner survey of 272 heads of I&O, conducted in April and May 2026, 58% said scaling AI proofs of concept into I&O adoption will be an important priority in the next 12 months.
“AI is creating one of the most significant shifts the I&O function has ever faced,” said Jonathan Forest, a vice president 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.”
Gartner named five interlinked AI trends in I&O as a result of its importance.
Trend 1: Agents take over routine IT work
AI agents are software programs that pursue a goal and carry out multistep tasks on their own. As IT operations grow more complex and move faster, organizations are using agents for incident triage, change validation and remediation.
“AI agents are driving a shift from reactive, human-in-the-loop operations to proactive, human-on-the-loop operations,” Forest said. “Heads of I&O who fail to restructure their organization around agentic AI risk higher costs, poor operational agility and performance, and ultimately, irrelevance.”
What does that shift look like? A person approves each step in a human-in-the-loop model while the agent acts and a person supervises in a human-on-the-loop model. It also moves governance, from approving each action to setting the limits that agents work within.
Trend 2: Agents run up the token bill
An agent completes a task in many steps, and every step uses tokens. Gartner says this drive up token use, especially as many providers charge by tokens, and create new cost headaches for I&O teams.
Through 2027, the firm predicts enterprises that run IT operations agents on uncapped, token-based pricing to exceed their initial cost expectations by more than 50%.
To control costs, the firm suggests a strict “no telemetry, no deployment” policy. If a team can’t track how many tokens an agent uses, the agent doesn’t go live.
While token prices are falling, it may not be enough to close the gap. In August 2026, Gartner predicted that AI inference costs per agentic workflow will still increase more than fivefold through 2028.
Trend 3: Inference spreads across clouds and data centers
Today, many enterprises use existing foundation models through inference instead of training their own. This means their data is often spread across public clouds, private data centers and edge locations. Rising regulatory concerns on privacy and data sovereignty coupled with cost and performance worries means the AI infrastructure will become more hybrid.
“Heads of I&O who fail to restructure their organization around agentic AI risk higher costs, poor operational agility and performance, and ultimately, irrelevance.” — Jonathan Forest @ Gartner
By 2029, Gartner predicts, 80% of large enterprises will run production AI workloads across at least three different infrastructure environments. These cover public clouds, neoclouds (new providers that specialize in AI computing), on-premises data centers and edge locations.
Trend 4: Operations and security converge
The line between operational risk and cybersecurity risk will blur as AI becomes part of the enterprise infrastructure. Shared AI tools, shared telemetry and the growing use of AI agents will push I&O and security teams to work closer together on resilience, risk management and efficiency.
Gartner expects some companies to take a step further by merging their. By 2030, Gartner predicts, 25% of enterprises will merge their network operations centers (NOCs) and their security operations centers (SOCs), up from 5% in 2026. Those that do will also slash total spending on tooling and staff by more than 25%.
Trend 5: I&O becomes the AI platform orchestrator
The four above trends lead to the fifth one: the development of a coordinated approach to infrastructure, governance, security and operations. As a result, Gartner sees the I&O function moving from a gatekeeper role to orchestrating the enterprise AI platform. This will bring I&O closer to the business and help them to deploy and scale AI securely and at a manageable cost.
This fifth trend explains the report’s starkest number on how likely organizations without a strong I&O function will fail in their AI initiatives through 2030.
“The high cost of AI infrastructure leaves little room for error, placing I&O at the center of enterprise AI success,” Forest said. “Heads of I&O must elevate their teams’ capabilities, reduce technical debt and shift from reactive operations to proactively enabling enterprise AI.”
Image credit: iStockphoto/ master1305
Stay ahead with CDO Nexus
Join an exclusive community of CDOs and data leaders
- Access curated trends and thought leadership from IBM and industry experts.
- Participate in private roundtables and webinars to solve regional CDO challenges.
- Connect with a network of like-minded leaders navigating the same data landscape.