AI Demand Drives 94% Surge in Global Semiconductor Revenue Forecast for 2026

Omdia projects a 94.1% surge in global semiconductor revenue for 2026, driven by relentless AI demand that outpaces supply. Memory ICs will exceed 50% of total revenue, with Computing & Data Storage approaching $1 trillion. However, bottlenecks in high bandwidth memory (HBM) from SK Hynix, Samsung, and Micron, alongside TSMC's fully utilized advanced packaging and leading nodes, will persist until 2027, driving up costs and causing delays for non-AI markets like smartphones and PCs.

Semiconductor revenue forecast surge

• ▪Computing & Data Storage semiconductor revenue is projected to rise by more than 150% year-over-year in 2026 to approach $1 trillion
• ▪Memory ICs are expected to account for more than 50% of total semiconductor revenue in 2026
• ▪Omdia raised its 2026 semiconductor revenue forecast to 94.1% year-over-year growth, driven by DRAM and NAND growth

AI-driven supply constraints

• ▪AI demand has exceeded the semiconductor industry's ability to produce and package chips, with bottlenecks expected to persist until at least 2027
• ▪TSMC's 2nm and 3nm leading node capacities are largely booked by companies including NVIDIA, AMD, Broadcom, and Apple
• ▪Computing requirements for AI models are growing faster than foundry capacity can expand

Memory IC pricing volatility

• ▪Smartphone price increases began in the fourth quarter of 2025 and are rising further in 2026, particularly in the premium tier
• ▪DRAM suppliers are prioritizing high bandwidth memory and other high-margin products, making commodity memory pricing and lead times volatile

High bandwidth memory bottlenecks

• ▪AI accelerators from NVIDIA, AMD, Intel, and Google require high bandwidth memory stacks, intensifying demand
• ▪High bandwidth memory supply is constrained because production is complex and relies on only three suppliers: SK Hynix, Samsung, and Micron

Advanced packaging capacity limits

• ▪Equipment suppliers ASML and Tokyo Electron face manufacturing constraints that limit the expansion of advanced packaging capacity
• ▪Advanced packaging manufacturing lines at TSMC are operating at full utilization, with capacity unable to expand quickly due to long equipment lead times

Non-AI market cost pressures

• ▪Smartphones, PCs, consumer electronics, automotive, and industrial markets face higher component costs and must compete with AI demand for packaging and memory ICs
• ▪Due to AI prioritization, CPUs and SOCs may face delays, higher average selling prices, or remain on older process nodes longer, slowing performance improvements