The spending on Artificial Intelligence (AI) infrastructure market is projected to hit $758 billion by 2029, according to IDC. Organizations worldwide ramped up investments in compute and storage hardware for AI deployments by 166 percent year-over-year in Q2 2025, reaching $82 billion, IDC’s Worldwide Quarterly Artificial Intelligence Infrastructure Tracker said.
The surge reflects the growing importance of advanced infrastructure to support generative AI, machine learning, and large language model (LLM) workloads. The majority of this investment was driven by servers for AI workloads, which represented 98 percent of total AI-centric infrastructure spending in the second quarter. Server spending rose 173.2 percent from the same period last year, reflecting a massive shift toward high-performance, accelerator-based systems.
Servers equipped with embedded accelerators — mainly GPUs and specialized AI chips — dominated, accounting for 91.8 percent of total server AI spending and growing at 207.3 percent. IDC expects accelerated servers to represent over 95 percent of all AI infrastructure spending by 2029, expanding at a 42 percent five-year CAGR.
The momentum is driven by major hyperscalers, cloud service providers, and digital platforms, which collectively accounted for 86.7 percent of total AI infrastructure spending in Q2 2025. Infrastructure deployed in cloud and shared environments made up 84.1 percent of total AI-related investments, as enterprises increasingly adopt cloud-based models for scalability and flexibility.
For instance, Amazon has planned to invest over US$100 billion in 2025 on AI and cloud infrastructure, especially via its AWS business.
Microsoft announced around US$80 billion in fiscal 2025 for data-centres designed to train and deploy AI models.
Alphabet (Google) is planning approximately US$75 billion in capital expenditures in 2025, with much of it earmarked for AI infrastructure (servers, data-centres, networking).
Meta Platforms is expecting capex of around US$60‒65 billion in 2025 for large-scale AI infrastructure build-outs (data centres, AI model training).
IDC analysts noted that demand for GPU-powered servers in the U.S. continues to exceed expectations, replacing earlier forecasts of a slowdown in late 2025 and early 2026.
“There is a distinct possibility that more AI-related investment will be announced in the coming years that will add to and extend the current mass deployment phase of accelerated servers well into 2026 and even beyond,” said Lidice Fernandez, Group Vice President, Worldwide Enterprise Infrastructure Trackers at IDC.
Beyond compute, AI-driven storage spending is also rising steadily — up 20.5 percent year-over-year in Q2 2025 — as organizations manage vast datasets needed for AI model training, checkpoints, and inference data repositories. Nearly half (48 percent) of all AI storage investment came from cloud deployments.
Geographically, the United States leads the global AI infrastructure market, accounting for 76 percent of total spending in Q2 2025. It is followed by China (11.6 percent), Asia-Pacific excluding Japan (6.9 percent), and EMEA (4.7 percent). Over the next five years, China is expected to grow fastest at 41.5 percent CAGR, with the U.S. close behind at 40.5 percent, as AI infrastructure expansion accelerates globally.
By 2029, IDC expects AI infrastructure spending to reach $758 billion, with accelerated servers capturing 94.3 percent of the total market — a clear signal that the next phase of global digital transformation will be powered by AI-centric compute systems.
Rajani Baburajan
