Intel 18A

Intel 18A. It is an advanced semiconductor manufacturing process node developed by Intel as part of its ambitious IDM 2.0 strategy and process roadmap. Scheduled for manufacturing readiness in late 2024, it represents the culmination of the company's "five nodes in four years" plan aimed at regaining process leadership. The node incorporates two major transistor architecture innovations: the RibbonFET gate-all-around design and a backside power delivery network called PowerVia.

Overview

The development of this process is a cornerstone of Pat Gelsinger's strategy to restore Intel's competitive position against rivals like TSMC and Samsung Electronics. It is positioned as the final step in the company's aggressive catch-up plan before pursuing further leadership. The node is central to the operations of Intel Foundry Services, aiming to attract external clients such as Microsoft and the United States Department of Defense. Key research and development for its components has been conducted at facilities like the Gordon Moore Park and the D1X development fab in Hillsboro, Oregon.

Technology and innovations

The process introduces the RibbonFET, Intel's implementation of a gate-all-around architecture, which replaces the older FinFET design to improve electrostatic control and drive current. Concurrently, it debuts PowerVia, a backside power delivery technology that separates power routing from signal interconnects to reduce voltage drop and improve transistor performance. These innovations were enabled by advanced EUV lithography tools from ASML and new materials developed through partnerships with companies like Applied Materials. The integration of these features is designed to offer significant improvements in performance per watt, a critical metric for next-generation CPUs, AI accelerators, and 5G infrastructure.

Development and production timeline

Announced in 2021 as part of a revised roadmap, the node followed the preceding Intel 20A process. Intel achieved first test chip tape-out in early 2023, a milestone verified on its internal Core Ultra client processors. Major manufacturing readiness is scheduled for the second half of 2024, with volume production expected in 2025. Initial production will occur at the company's leading-edge fabs, including facilities in Hillsboro, Oregon and the new Ocotillo campus in Arizona. The timeline is closely tied to product schedules for future Xeon server processors and Arrow Lake client architectures.

Comparison with competing processes

When introduced, this process is expected to compete directly with similar-generation nodes from other leading foundries, such as TSMC's N2 and Samsung's SF2 processes. Industry analysts from firms like TechInsights and International Business Strategies will scrutinize its transistor density, performance, and power efficiency against these rivals. A key differentiator is its early adoption of backside power delivery, a feature competitors like TSMC plan to introduce later with their Backside Power Rail technology. Its success is critical for Intel Foundry Services in competing for contracts against the dominant TSMC in markets spanning from Apple to Nvidia.

Products and applications

The first products utilizing this node are anticipated to be future generations of Intel's client and server processors, potentially including the Clearwater Forest architecture for the Xeon lineup. It is also destined for high-performance computing applications, such as parts for the Aurora supercomputer and custom chips for the United States Department of Energy. Through Intel Foundry Services, the technology will be available to external partners, with early customers like Microsoft reportedly designing chips for its Azure cloud platforms. The process is also targeted for critical applications in the automotive, communications, and defense sectors. Category:Semiconductor fabrication plants Category:Intel microprocessors Category:Integrated circuit manufacturing