Research from TrendForce indicates that ongoing investments in AI infrastructure by North American cloud service providers are projected to propel global AI server shipments to soar by more than 28% year-on-year in 2026. During the same period, total global server shipments are anticipated to experience a 12.8% year-on-year increase. From 2024 to 2025, the server market will predominantly concentrate on training advanced large language models. Beginning in the second half of 2025, cloud service providers' (CSPs) profit strategies will pivot towards inference services. Concurrently, AI inference workloads will start to be deployed on general-purpose servers.

It is predicted that the total capital expenditure of the top five North American cloud service providers will jump by 40% year-on-year in 2026. Google and Microsoft are likely to spearhead the expansion in general-purpose server procurement. The bulk of the growth in AI server market shipments will stem from North American cloud service providers, among other entities. GPUs will continue to hold sway, with systems based on NVIDIA GB300 driving a significant portion of the growth.

Meanwhile, the market share of AI servers built on ASIC chips is expected to reach 27.8%. Their shipment growth may even outpace that of GPU-based systems. Google stands out as a frontrunner in investing in the ASIC chip sector, with its TPU chips being marketed more extensively to external customers.