Global AI-Optimized IaaS Spending to Hit $42 Billion in 2026: Gartner Report

Worldwide AI-optimized infrastructure as a service (IaaS) spending is on track to experience massive growth, projected to surge by 96 percent to reach $42 billion in 2026. According to a new forecast by business and technology insights firm Gartner, Inc., this momentum is expected to carry into 2027, with the market forecast to reach a staggering $66 billion. Hardeep Singh, Senior Principal Research Analyst at Gartner, explained that this exponential growth is primarily driven by the continuous demand for infrastructure required to support large language model (LLM) training, alongside the rapid operationalization of AI across enterprise applications and workflows.

A Shift in Cloud Infrastructure Investments
The broader IaaS market is also showing strong momentum. Total IaaS spending, which stood at $222.1 billion in 2025, is projected to grow to $287.3 billion in 2026 and reach nearly $360 billion by 2027. However, the AI-optimized segment is growing at a significantly faster rate (surging 180 percent in 2025 and 96.4 percent in 2026), highlighting its increasing importance and footprint in overall IT budgets.

Inference Workloads to Surpass Training by 2026
A major shift is underway in how AI infrastructure is consumed. The rise of “agentic AI”—which requires multistep, autonomous execution—is amplifying compute intensity. As a result, inference is rapidly becoming the dominant consumption model, positioning AI-optimized IaaS as a critical enabler of long-term enterprise AI strategies. By 2026, global spending on inference workloads is projected to reach $23.3 billion, officially surpassing the $19 billion spent on AI training. Inference is forecast to account for 55 percent of all AI-optimized IaaS spending in 2026, growing further to 59 percent by 2027. This growing share is expected to fundamentally reshape how enterprises prioritize their cloud investments.

“As organizations shift from model development to production-scale deployment, fine-tuned and domain-specific models (DSMs) are increasingly integrated into customer-facing and operational systems, requiring continuous, real-time execution rather than periodic training,” Singh stated. “This shift is accelerating cloud consumption patterns and creating sustained demand for AI-optimized infrastructure.” Industry leaders and CIOs will further explore these shifting infrastructure trends and the operationalization of AI at the upcoming Gartner IT Symposium/Xpo events, scheduled globally throughout late 2026 in locations including the Gold Coast, Orlando, Yokohama, Barcelona, and Kochi.