STMicroelectronics, an European chipmaker, and the National University of Singapore (NUS) launched a research lab to develop energy-efficient semiconductors for edge AI, the processing of AI tasks on or near a device rather than in the cloud.

In separate statements on Monday, STMicroelectronics and the NUS said the lab is part of a four-year strategic research initiative.

The ST-NUS HELIX Corporate Lab, focusing on Hardware for Embodied Low-power Intelligent Xcceleration (HELIX), is supported under Singapore’s Research, Innovation and Enterprise 2025 plan. It is hosted at the NUS’s College of Design and Engineering.

It will focus on “embodied” AI, in which computing is built into physical devices such as robots, humanoids, and drones. The devices must sense, process, and act in real time, serving compact, low-power hardware.

The lab’s research will span the full chip stack, from AI models and system architecture to circuit design and silicon, with an emphasis on memory-centric designs and in-memory computing. A target is to cut the movement of data between chips, a major source of energy use in AI workloads.

The work will draw on ST’s P18 18-nanometre fully depleted silicon-on-insulator (FD-SOI) process and embedded phase-change memory. STMicroelectronics will supply the NUS with a dedicated design platform built on the FD-SOI process, giving researchers an industrial-grade base for developing and testing new AI accelerator designs without building the underlying infrastructure themselves.

STMicroelectronics and NUS said the lab would also strengthen Singapore’s semiconductor research capabilities and train engineers in AI systems and chip design, with students and researchers working on its projects.

Vietnam’s science ministry, Singapore’s NUS Enterprise to collaborate in AI, semiconductor, startup