Artificial Intelligence Infrastructure Market Braces for Massive Growth: Spending to Exceed $100 Billion USD by 2028
The global landscape of Artificial Intelligence (AI) infrastructure is on the brink of tremendous expansion. As per the most recent analysis, investment in this arena is anticipated to soar beyond $100 billion USD by the end of 2028. This remarkable shift is substantiated by an increase in expenditure on computing and storage hardware for AI implementations by an impressive 37% year-over-year during the first half of 2024, culminating in a total of $31.8 billion.
AI infrastructure has been enjoying robust growth for nine consecutive half-year periods, fueled principally by an uptick in expenditures on servers specifically for AI applications. Notably, servers represented a staggering 89% of the total investments in AI infrastructure during the initial half of 2024, marking a 37% growth over the corresponding span from the previous year. A substantial 65% of server spending has been attributed to AI infrastructure in cloud and shared environments, driven by the expansion efforts of hyperscale operators, cloud service providers, and digital service enterprises. Conventional businesses, however, have been slower in the integration of on-premises AI capabilities.
Servers equipped with internal accelerators have emerged as the favored infrastructure solution for AI platforms, constituting 58% of the total server AI infrastructure expenditure. This segment surged by 63% in the first half of 2024. Projections indicate that servers with enhanced acceleration capabilities will surpass 60% of the server AI infrastructure spending by 2028, maintaining a compound annual growth rate (CAGR) of 19% over the next five years.
Moreover, storage spending in AI infrastructure has experienced significant impetus due partly to the necessity of handling vast datasets vital for AI model training, along with storage for checkpoints and data repositories needed during inference. This segment noted a 36% year-over-year advancement in the first half of 2024, with a dominant 56% of spending emanating from cloud-based solutions.
On a global scale, the United States is at the forefront of AI infrastructure spending, constituting nearly half of the total expenditure in the first half of 2024. Following closely are the People’s Republic of China (23%), the Asia-Pacific and Japan (16%), and Europe, the Middle East, and Africa (10%). In the next five-year horizon, it is predicted that the Asia-Pacific and Japan will experience the most rapid expansion with a CAGR of 20%, trailed by the USA (16%), EMEA (13%), and PRC (11%). By 2028, estimates foresee AI infrastructure investments expanding to $107 billion, with cloud-based server deployments accounting for 75% of the market’s total and accelerated servers representing approximately 56% of total market spending.
Moving forward, AI adoption is anticipated to continue on its prolific trajectory as hyperscale businesses, cloud providers, private sectors, and governmental bodies worldwide increasingly emphasize AI innovations. Emerging concerns surrounding the energy consumption required for AI infrastructure may catalyze datacenter operators to explore alternative avenues to streamline their architectural designs, aiming to reduce energy demands.
AI-centric systems in the server and storage categories, as defined, encompass applications aligned with AI platform solutions dedicated to the development of AI applications, where most training activities occur. This framework essentially provides a “workloads-based” perspective on the server and storage markets.
