Intel’s 18A semiconductor process marks a technological milestone as the first 2nm-class node developed and manufactured in the US, with substantial improvements in performance and density compared to previous processes. Key innovations include the RibbonFET transistor design and PowerVia backside power delivery system that optimize power efficiency and signal integrity. Additionally, Intel utilizes Foveros 3D stacking technology to combine multiple chiplets, enhancing system flexibility and scalability.
The 18A node underpins the Intel Core Ultra series processors and next-generation Intel Arc GPUs, delivering AI acceleration with up to 180 TOPS and improved CPU and graphics performance by over 50%. These products will support diversified applications ranging from consumer PCs and gaming to edge computing and robotics through a scalable multi-chiplet architecture and new AI software tools. Mass production began in 2024 in Intel’s advanced Fab 52 facility in Arizona, emphasizing domestic innovation and manufacturing.
Intel also previewed the Xeon 6+ server processor (Clearwater Forest) built on the 18A process, targeting hyperscale data centers with enhanced IPC and energy efficiency. This server chip offers up to 288 Efficiency cores and overall improvements supporting next-generation intelligent services. Intel’s commitment to integrating cutting-edge manufacturing with research and design in the US reinforces its strategic innovation direction and industry leadership.
👉 Pročitaj original: CIO Magazine
