The global Artificial Intelligence (AI) infrastructure market is experiencing exponential growth, with spending projected to surpass $200 billion by 2028.
Organizations have significantly increased their investments in computing and storage hardware to support AI deployments, with a 97 percent growth in the first half of 2024 alone, reaching $47.4 billion. This expansion underscores the increasing reliance on robust AI infrastructure to support the rapid evolution of artificial intelligence applications, according to the International Data Corporation (IDC) Worldwide Semiannual Artificial Intelligence Infrastructure Tracker.
The sustained double-digit growth in the AI infrastructure market since 2019 has been largely driven by investments in servers optimized for AI workloads.
In the first half of 2024, server spending dominated the market, accounting for 95 percent of total AI infrastructure expenditures and experiencing an impressive 105 percent growth compared to the same period in the previous year. The dominance of server spending highlights the pivotal role of powerful computational resources in AI model training and inference.
A significant portion of this investment — 72 percent — was allocated to AI infrastructure deployed in cloud and shared environments, as hyperscalers, cloud service providers (CSPs), and digital service providers continue to enhance their infrastructure capabilities to meet the rising demand. In contrast, traditional enterprises have been relatively slow in adopting on-premises AI infrastructure, often citing cost and scalability concerns as key barriers.
A major shift in AI infrastructure preferences has been the growing adoption of servers with embedded accelerators, which accounted for 70 percent of total AI server spending in the first half of 2024. This segment grew at an extraordinary rate of 178 percent in just six months, reflecting the increasing demand for high-performance computing solutions that can efficiently process AI workloads.
IDC forecasts that by 2028, servers with embedded accelerators will represent over 75 percent of total server AI infrastructure spending, with an estimated compound annual growth rate (CAGR) of 42 percent over five years. Accelerated computing is becoming a crucial component of AI infrastructure, as enterprises and cloud providers strive to enhance processing power and efficiency in handling complex AI models.
Storage infrastructure has also witnessed steady growth, fueled by the necessity to handle vast datasets essential for AI model training, inference, and continuous learning. In the first half of 2024, spending on AI storage grew by 18 percent, with cloud deployments accounting for 40 percent of this expenditure. The need for scalable, high-performance storage solutions is becoming increasingly critical, especially as AI applications demand more sophisticated data management strategies to optimize performance and reduce latency.
Geographically, the United States remains the dominant force in AI infrastructure spending, accounting for 59 percent of total investments in the first half of 2024. China follows with a 20 percent share, while the Asia-Pacific region excluding Japan (APJ) and Europe, the Middle East, and Africa (EMEA) account for 13 percent and 7 percent, respectively. However, China is expected to experience the fastest growth in AI infrastructure spending, with a projected CAGR of 35 percent over the next five years. The United States is close behind, with an expected growth rate of 34 percent, followed by APJ at 21.3 percent and EMEA at 20.9 percent.
By 2028, global AI infrastructure spending is forecasted to reach $223 billion, with cloud-based server deployments comprising 82 percent of the market and accelerated servers making up approximately 74 percent of total expenditures.
As AI adoption continues to expand across industries, hyperscalers, cloud service providers, private enterprises, and governments are prioritizing investments in AI infrastructure to support next-generation applications. However, alongside this rapid growth, concerns regarding energy consumption and sustainability are becoming increasingly relevant.
AI workloads require significant power, and data centers are actively exploring architectural optimizations and alternative energy solutions to minimize environmental impact while maintaining high-performance computing capabilities.
According to Lidice Fernandez, Group Vice President of Worldwide Enterprise Infrastructure Trackers at IDC, optimizing AI infrastructure for energy efficiency will be a critical challenge in the coming years as the demand for AI-driven applications continues to surge.
Baburajan Kizhakedath
