What happened
Epoch AI analysis reveals memory (HBM) now constitutes the largest share of AI chip component costs, rising from 52% in Q1 2024 to 63% by Q4 2025. This shift occurred as total component spending for AI chips, designed by Nvidia, AMD, Google, and Amazon, expanded from approximately $22 billion in 2024 to $52 billion in 2025. HBM spending alone drove roughly $20 billion of this increase, while advanced packaging and auxiliary components saw their shares decline. Logic die costs remained stable at 13-14%.
Why it matters
Rising HBM costs will reshape AI hardware procurement and design strategies. For hardware architects, the increasing memory premium dictates chip configurations, potentially limiting logic die integration or driving innovation in memory efficiency. Procurement teams face escalating bills of materials, impacting overall system costs and potentially slowing AI infrastructure build-outs. This follows recent reports of AI chip shortages driving flash memory price surges, indicating broader supply chain pressures on critical components. Investors will scrutinise chipmakers' ability to manage these rising input costs and maintain margins.
