Global spending on artificial intelligence (AI) infrastructure will cross the $100 billion mark by 2028, according to a report by IDC. In the first half of 2024, hardware infrastructure for AI accounted for $31.8 billion.
The growth in AI infrastructure is mainly driven by investments in servers. In the first half of 2024, 89 percent of spending was focused on servers, up 37 percent from the same period last year.
Servers with built-in accelerators are the most popular and accounted for 58 percent of total server spending, a category that grew by 63 percent. IDC expects this share to rise further to more than 60 percent by 2028, with an average annual growth rate of 19 percent.
Cloud remains dominant
Cloud-based infrastructures remain dominant, accounting for 65 percent of server spending in 2024. Hyperscalers and cloud providers continue to expand their capabilities, while traditional enterprises lag behind in the adoption of on-premises AI infrastructures.
Storage also plays an important role, especially with the need to manage large data sets for AI training and inference. In the first half of 2024, storage spending increased 36 percent, with 56 percent of the investments being in cloud environments.
Regional distribution and future growth
The United States remains the largest market for AI infrastructure, with nearly half of the spending in the first half of 2024. China (23%), Asia Pacific (16%) and Europe (EMEA, 10%) follow. IDC predicts that the Asia-Pacific region will experience the fastest growth at 20% over the next five years, followed by the U.S. (16%), Europe (EMEA, 13%) and China (11%).
read also
AI infrastructure spending surpasses $100 billion by 2028
By 2028, global spending is expected to reach $107 billion. Cloud-based servers will account for 75 percent of the market, while servers with accelerators will account for 56 percent of total spending.
Challenges around energy consumption
IDC points to growing concerns about the energy consumption of AI infrastructure. The energy consumption of European data centers could triple in five years. The energy demand of AI workloads worldwide today already approaches that of a small country.
There is also the issue of water consumption that needs to stay on the radar. An AI-generated email of a hundred words takes about three bottles of 0.5 liters of water.
A lot of tech giants are being forced to adjust their climate goals. For example, Mirosoft representative Craig Cincotta stated to the Washington Post that the company “will work on data center cooling methods that will completely eliminate water consumption.” Yet the profit motive of AI still often seems to prevail and environmental goals lag behind. Google, for example, saw its greenhouse gas emissions rise 48 percent in just five years because of the uninhibited AI push.