Threads
- Intel’s 18A chip manufacturing process is scheduled for mass production in the latter half of 2025.
- In tests, an Arm core built with the 18A process showed a 15% performance increase over the Intel 3 process at the same power level.
- Manufacturing the 18A chips will require 44% fewer masks due to the use of low-NA EUV lithography.
Intel highlighted the key advantages of its most recent manufacturing process, Intel 18A, which is set to enter mass production in the second half of 2025, over the current Intel 3 at the VLSI Symposium in Japan. Most importantly, the first chips produced using this approach are expected to enter mass production as early as the second half of 2025.
We predict that these chips will be ready for purchase by the end of this year or early in 2026.
For the company, the Intel 18A production process represents a major milestone. This node integrates the PowerVia power delivery network on the backside with new Gate-All-Around transistors. As a result, the metal stack architecture is entirely different.
It has improved performance and simplified production by modifying the interconnect pitches on essential layers and relaxing the spacing on the top layer by passing power via the die’s backside.
Intel released standardised power, performance, and area testing conducted on a subblock using an Arm core as a reference during the symposium. Compared to the Intel 3, this Arm core, which was produced utilising the Intel 18A node, showed about 15% better performance at the same power consumption.
Clock speeds might increase by up to 25% without requiring more power when the core was restricted to 1.1 volts. On the other hand, performance may rise by 18%, or power consumption could drop by nearly 40% at 0.75 volts.
Taking advantage of the symposium, Intel presented other significant enhancements related to the Intel 18A node. It has considerable reductions in cell height. Performance-tuned cells are 180 nanometres tall, while high-density designs are 160 nanometres, both smaller than their predecessors.
The front-facing metal layers have been lowered from 12 to 19 nm in the Intel 3 process to 11 to 16 nm in the Intel 18A, with three extra rear metal layers added for PowerVia compatibility.
The M1 to M10 layers’ pitches have been lowered from 60 nanometres to 32 nanometres, and the top layers’ pitches have also been lowered. The M0 to M4 layers are exposed to low-NA EUV, which simplifies the manufacturing process and reduces the number of masks required by 44%.
Additionally, compared to Intel 3, the SRAM density has improved by 30% with Intel 18A.
Later this year, Intel Panther Lake laptop processors will be the first to use Intel 18A chips. The forthcoming Intel Xeon 7 Clearwater Forest server CPUs, which are built entirely of low-power cores (E-Core), will follow these.
Thank you! Please share your positive feedback. 🔋
How could we improve this post? Please Help us. 😔
Join Our Community
