IBM Unveils World's First 'Sub-1nm' AI Semiconductor Technology

IBM has introduced the world's first sub-1nm semiconductor technology, showcasing a new architecture targeting the next generation of AI chips. The company announced a breakthrough that enables approximately 100 billion transistors to be integrated onto a fingernail-sized silicon chip, achieving twice the transistor density of its 2nm technology from 2021. This advancement significantly enhances AI computing performance while reducing energy consumption. IBM's new technology, termed 'sub-1nm chip technology,' represents a 0.7nm (7 angstroms) performance node, though it does not denote physical circuit dimensions. The core innovation is IBM's NanoStack transistor architecture, which stacks transistors in 3D layers to increase component density. This builds on its 2021 nanosheet transistor technology, enabling up to 50% higher computing performance or 70% improved power efficiency compared to 2nm chips. The technology was presented at the recent IEEE VLSI Symposium in Kyoto, with memory performance also enhanced by 40% in SRAM integration for AI applications.