What powers the AI boom behind the scenes? A new study predicts semiconductor revenue from AI data centres could skyrocket by 2028, just two years from now.
The annual revenue from semiconductors deployed in global artificial intelligence (AI) data centres could exceed US$1.2 trillion by 2028, according to a new study by the Semiconductor Industry Association (SIA) and Deloitte. The report, Powering AI: The Semiconductor Ecosystem at the Foundation of Data Centers, highlighted the growing economic significance of chips in AI infrastructure behind this.
Published on June 1, 2026, the report also concluded that semiconductors account for more than 95 per cent of the content value of a modern AI server rack and represent over half of the capital expenditure associated with building and operating AI data centres.
According to SIA, AI infrastructure relies on a broad range of semiconductor technologies, including advanced logic, memory, analogue and foundational chips. Researchers examined the composition of a state-of-the-art AI server rack and found that a single rack contains more than 4,500 packaged chips.
It is also estimated that governments and industry worldwide will invest more than US$4 trillion in AI data centre infrastructure by 2028, with as much as US$2.8 trillion directed towards semiconductors. The projected US$1.2 trillion annual semiconductor revenue from AI data centres would represent nearly tenfold growth over the previous four years and exceed total global semiconductor sales recorded in 2025 across all end markets.
Key components identified in AI server racks include AI accelerators, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), central processing units (CPUs) and data processing units (DPUs), networking chips, high-bandwidth memory, DRAM, SRAM, NAND flash memory, power management devices, controllers, sensors, and transceivers.
AI accelerators account for the largest share of server rack value, contributing 74 per cent, with logic chips making up 70 per cent of the semiconductor content within that category. CPUs represent 8 per cent of rack value, while volatile memory and networking/interconnect components each account for 5 per cent.
Meanwhile, power and cooling components contribute 3 per cent, DPUs 2 per cent, and non-volatile storage 1 per cent. Across the rack, logic and memory technologies dominate semiconductor content, together representing more than 85 per cent of total semiconductor value.
Researchers note that continued innovation across the semiconductor supply chain will be essential as global competition in AI intensifies and demand for advanced computing infrastructure expands.
