Why AI data centre boom is leaving consumer electronics short of chips

The AI boom has changed memory demand, prioritizing high-bandwidth memory chips for data centers over consumer electronics. Data centers require maximum compute power, memory bandwidth, and storage throughput, whereas consumer devices need low power use, thermal efficiency, and tight integration. As a result, AI servers rely on graphics processing units and high-bandwidth memory chips, while consumer devices use systems-on-a-chip with dynamic random access memory and NAND. The chip market's concentration, with firms like NVIDIA and TSMC dominating, contributes to the supply crunch. Chipmakers are hesitant to add capacity due to high fixed costs and the risk of underutilization. The industry's consolidation has led to a shift towards higher-margin products, further tightening chip supplies for consumer electronics.