DRAM for robots forecast to grow 20-fold by 2030 ... Could robots become a new growth driver for memory?

[Financial News] As humanoid robots emerge as a next-generation source of demand for memory chips, forecasts suggest that demand for DRAM used in devices serving as robots’ brains will grow to 20 times this year’s level by 2030. Accordingly, analysts say robots could become a new growth driver for memory chipmakers such as Samsung Electronics Co., Ltd. and SK hynix, following high-bandwidth memory (HBM) for artificial intelligence (AI) data centers.
According to market research firm Counterpoint Research on the 11th, total DRAM demand across the humanoid robot industry is forecast to rise at least 20-fold, from about 8,300 terabits (Tbit) this year to about 172,000 terabits in 2030.
This is because the memory capacity installed in each robot is also increasing. The average DRAM capacity per robot is forecast to at least double, from 19 gigabytes (GB) this year to 39 GB in 2030.
A robot’s brain is a computing unit that analyzes in real time information gathered by cameras, LiDAR and tactile sensors to determine its movements. Memory plays a key role in processing what a robot sees and feels and turning that information into action.
The shift from the existing hierarchical structure, in which AI models process information step by step, to an “end-to-end” approach that integrates multiple inputs, such as visual and tactile information, and directly determines actions is driving up the amount of memory required.
In addition to the stage of training AI to help robots understand the real world and predict what comes next, the storage and computing capacity needed to run trained models on actual robots is also increasing.
The memory industry is therefore watching the growth potential of humanoid robots. Sanjay Mehrotra, chief executive officer (CEO) of Micron Technology, Inc., forecast that humanoid robots’ memory and storage requirements would be comparable to those of advanced autonomous vehicles during the earnings announcement for the fourth quarter of fiscal 2026 on September 30 (local time).
Mehrotra said, “Level 4 and higher autonomous vehicles typically require more than 200 GB of memory and several terabytes (TB) of storage,” adding, “Humanoid robots will also require similar amounts of memory and storage as autonomous vehicles.”

[email protected] Kim Hak-jae Reporter