AI infrastructure drives data center chip growth

The global data center chip market reached US$ 16.51 billion in 2025 and is expected to reach US$ 50.91 billion by 2035, growing with a CAGR of 11.92% during the forecast period 2026-2035. Growth is being driven by generative AI, large language models, high-performance computing and hyperscale cloud buildouts, with AI servers using 4-8 GPUs or accelerators and rack-scale systems using up to 72 AI accelerators.

Graphics processing units remained the leading chip type with 46.0% market share in 2025, supported by broad use in AI training, inference and cloud workloads. North America led with 38.54% share in 2025, while the IT and telecommunications segment is expected to grow at a CAGR of 13.8% during 2026-2035 as 5G, edge data centers and AI network virtualization expand.

NVIDIA holds a leading position through its AI GPU portfolio, CUDA software ecosystem, Blackwell GPUs, NVLink and DGX platforms. AMD, Intel, Broadcom and Marvell are strengthening positions across accelerators, server CPUs, networking chips and custom silicon, while AWS, Google, Microsoft, Meta and ByteDance are investing in proprietary AI processors. Constraints around HBM supply, advanced packaging capacity, export controls, power use and cooling remain key market pressures.