Micron is developing edge inference memory for physical AI, with an expected launch in 2028.

According to ChainCatcher, Lee Je, Vice President of Micron Technology, stated at SEDEX 2026 held at COEX in Seoul that the company is developing Edge Inference Memory (EIM) for physical AI. The industry expects this product to enter mass production in the second half of 2028. Micron is exploring the integration of EIM with LPDDR6 to meet the demand for low-power DRAM in applications such as autonomous driving and humanoid robots, and is currently engaging in discussions with key customers. The Shanghai Securities News reported that EIM is positioned between AI chips and memory, stacking LPDDR on top of logic chips to enable high-speed processing of inference data; however, the full computational architecture has not yet been disclosed. Lee Je noted that current robot AI models contain approximately 1 to 2 billion parameters and have modest bandwidth requirements; however, as more advanced vision-language models emerge, LPDDR alone will be insufficient, and HBM will ultimately be adopted for humanoid robots. Micron estimates that future high-performance humanoid robots will require approximately 200 GB of DRAM and 4 TB of storage, compared to current service robots, which typically use around 12 GB of DRAM and 256 GB of storage. Micron’s CEO, Sanjay Mehrotra, stated that the storage capacity required per humanoid robot is roughly ten times that of a typical L2+ passenger vehicle. The storage industry may enter a multi-decade period of sustained demand growth beginning around 2030, driven by the need to simultaneously process data from hundreds of sensors and run large models locally for millisecond-level control. Counterpoint Research estimates that global humanoid robot DRAM demand will rise from approximately 8,300 Tbit this year to about 172,000 Tbit by 2030.