Humanoid Robot DRAM Demand Is Forecast to Rise 20-Fold by 2030

DRAM demand for humanoid robot computing systems is projected to climb from about 8,300 terabits in 2026 to about 172,000 terabits in 2030, roughly 20-fold. The forecast measures memory capacity, not chip revenue. Average DRAM capacity per robot is projected to reach 39 gigabytes in 2030, up from 19 gigabytes in 2026. The total is expected to grow through both increased capacity in each robot and rising shipments. Humanoid robots process information from cameras, LiDAR and tactile sensors. Multimodal models combine different kinds of input, while “world models” anticipate changes in the surrounding environment; both add to computing memory needs. Battery-powered robots are also expected to favor low-power DRAM, including LPDDR, alongside high-capacity NAND. Micron Chairman, President and CEO Sanjay Mehrotra said in prepared remarks dated Sept.

BeCrypto News shows a short excerpt and links to the publisher. The full report belongs to TokenPost.