The global server market is experiencing unprecedented growth. According to the latest report from IDC, the Worldwide Quarterly Server Tracker, in the first quarter of 2025, the market value reached $95.2 billion, marking a year-over-year increase of 134.1%, the largest quarterly growth ever recorded.
Starting from this figure, IDC projects that the market will hit a record value of $366 billion by the end of 2025, representing a 44.6% increase from the previous year.
The Dominance of GPU Servers and the Rise of ARM
This expansion is led by integrated GPU servers, which will grow 46.7% year-over-year and account for nearly 50% of the total market value by 2025. This trend is being especially accelerated by the demand from major cloud providers and hyperscalers, who increasingly require power for generative AI environments and intelligent agents.
x86 servers will remain predominant, with an estimated value of $283.9 billion (+39.9%), while non-x86 servers, particularly ARM architectures, will grow by 63.7%, reaching $82 billion.
The growth of ARM-based servers—thanks to rackscale configurations like NVIDIA’s GB200 (with Grace CPU + Blackwell GPU) and custom chips like Graviton (Amazon), Axion (Google), and Cobalt (Microsoft)—is rapidly increasing its market share, with a 70.0% growth in units, already representing 21.1% of total global shipments.
Why is Demand Growing So Quickly?
According to Kuba Stolarski, Vice President of Research at IDC, this growth is driven by the massive increase in processing needs stemming from new AI models. The evolution from simple chatbots to reasoning models and autonomous agents will require "several orders of magnitude more computing capacity, especially for inference tasks."
While advancements have been made in model efficiency, experts warn that this efficiency does not decrease demand; rather, it enables a larger scale of simultaneous users with more complex needs and demanding workloads.
Regional Projections: The U.S. Leads Growth
IDC’s report also breaks down growth by regions:
- United States: 59.7% growth compared to 2024, accounting for nearly 62% of total global revenue.
- China: 39.5% growth, with over 21% of quarterly revenue.
- Japan and APeJC: Increases of 33.9% and 10.8%, respectively.
- EMEA (Europe, the Middle East, and Africa): modest growth of 7.0%.
- Latin America: stagnation, with only a 0.7% increase.
- Canada: a decline of -9.6% due to an atypically large contract in 2024 that distorts comparisons.
A Market Tripled in Just Three Years
In 2022, the global server market barely reached $130 billion. If projections hold true, by 2025 that volume will have tripled. This structural change is driven by:
- The consolidation of generative AI as a cross-cutting technology.
- The expansion of inference workloads in sectors like healthcare, finance, industry, and utilities.
- The advancement of distributed computing and the need for specialized servers with accelerators (GPU, FPGA, ASIC).
Classification of Servers by Accelerators
IDC also classifies servers based on whether they are accelerated:
- Non-accelerated: without GPU or coprocessors. This includes those where the user adds a graphics card or uses integrated graphics chips on the board.
- With GPU: servers with programmable graphics units that can also handle AI workloads, increasingly demanded for inference and training.
- Other accelerators: include servers with FPGAs (reconfigurable) or ASICs (customized), typically used in specific high-performance or low-latency workloads.
AI Accelerates the Future of Digital Infrastructure
With the development of new architectures, such as specialized chips for AI and hybrid GPU+CPU solutions, the server market is shaping up to be a key pillar of the digital economy. The coming months will be crucial to confirm whether this trend continues and if companies can maintain a balance between scalability, sustainability, and performance.
Source: IDC