Global spending on AI-optimized Infrastructure as a Service (IaaS) is projected to increase by 96% by next year, reaching $42 billion. With the spread of agentic AI, the focus of cloud investment is expected to shift towards inference workloads.
According to Gartner, the global AI-optimized IaaS market is on a strong growth trajectory, expected to reach $66 billion by 2027. This year, spending is anticipated to hit $21.5 billion, a 180% increase from the previous year, with a further 96.4% rise to $42.2 billion next year, and a 56.5% increase to $66.1 billion by 2027.
During the same period, total spending on IaaS is expected to grow from $222.1 billion this year to $287.3 billion in 2026 and $359.9 billion in 2027.
Hardip Singh, a senior analyst at Gartner, stated, "The demand for infrastructure to support large language model (LLM) training continues to drive market growth as AI rapidly commercializes across enterprise applications and business processes."
However, the growth is shifting from training to inference. As agentic AI becomes more widespread, the computational intensity of multi-stage autonomous processing has increased, leading to a reallocation of resource consumption towards inference. This year, global spending on inference is expected to surpass training-related spending, reaching $23.3 billion compared to $19 billion for training. The share of inference in total AI-optimized IaaS spending is projected to grow from 55% in 2026 to 59% in 2027.
Singh noted, "As companies transition from model development to large-scale commercial deployment, fine-tuned models and domain-specific models (DSM) are being rapidly integrated into customer-facing and operational systems. These systems require continuous real-time execution rather than one-time training, accelerating cloud consumption and creating sustained demand for AI-optimized infrastructure."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.