Demand for DRAM for robots will increase 20 times by 2030... Samsung and SK emerge as new food sources

Counterpoint Research analysis… DRAM load per robot: 19GB → 39GB

(Seoul = Yonhap News) Reporter Kang Tae-woo = It is predicted that the demand for DRAM used in humanoid robots will increase approximately 20 times by 2030.

Following high-bandwidth memory (HBM) for artificial intelligence (AI) data centers, robots are emerging as a new growth engine for the memory semiconductor industry such as Samsung Electronics and SK Hynix.

According to market research firm Counterpoint Research on the 11th, the overall DRAM demand in the humanoid robot industry is expected to increase more than 20-fold from about 8,300 Tbit (Tbit) this year to about 172,000 Tbit in 2030.

This is because as the spread of robots expands, the memory capacity mounted on one robot also increases.

In fact, the average DRAM capacity mounted on one humanoid robot is expected to more than double from 19 gigabytes (GB) this year to 39 GB in 2030.

Humanoid robots determine their movements by analyzing information collected by cameras and various sensors in real time. Recently, as AI models have become more sophisticated, the memory capacity required to process them is also increasing.

The industry believes that robots will become a new source of demand for Samsung Electronics and SK Hynix. As power efficiency is important for battery-powered robots, demand for low-power DRAM (LPDDR) and high-capacity NAND flash is expected to increase.

Global memory companies are also paying attention to the growth potential of the robot market.

In an earnings announcement on the 30th of last month, Micron CEO Sanjay Mehrotra predicted, "Level 4 or higher self-driving cars are usually equipped with more than 200 GB of memory and several terabytes (TB) of storage," and predicted, "Humanoid robots will also require similar levels of memory and storage."

Counterpoint Research also predicted that as robot AI models become more sophisticated, demand for computational semiconductors, including NVIDIA's graphics processing units (GPUs), will also increase.