TechInsights and Semiwiki are publishing important details that Intel and TSMC have revealed at the International Conference on Electronic Devices (IEDM) about upcoming 18A (1.8NM class) and N2 (2NM class) process technologies. According to TechInsights, Intel’s 18A could offer higher performance, while TSMC’s N2 could offer higher transistor density.
Analysts at TechInsights believe that TSMC’s N2 offers a high density (HD) standard cell transistor density of 313 mtr/mm^2. SF3p (231 mtr/mm^2). The information should be noted with some things as well as the SRAM cell sizes for 18A, N2, N3, and TSMC’s expectations for N2 and N3, although there is some information.
First, this only concerns HD standard cells. Virtually every modern high-performance processor that relies on cutting-edge nodes, not to mention TSMC’s Finflex and Coyfability. Nanoflex.
Secondly, it is unclear how Intel and TSMC HP and LP standard cells compare. It is logical to assume that N2 has transistor density leads, but it may not be as large as the leads of HD standard cells. Third, in a paper published at the IEDM event, both Intel and TSMC revealed more than their predecessors the performance, power, and transistor density benefits of the next generation 18A and N2 manufacturing processes. Still, there is no way to directly compare these two manufacturing technologies at present.
In terms of performance, TechInsights believes Intel’s 18A will lead the N2 from TSMC and Samsung’s SF2 (formerly known as SF3P). However, TechInsights uses controversial methods to use TSMC’s N16FF and Samsung’s 14NM process technology as baselines, adding performance improvements between nodes from both companies to make predictions. Compare the performance of nodes. This may be useful as a quote, but it may not be entirely accurate.
Meanwhile, Intel specializes in manufacturing high-performance processors, allowing 18A to be adjusted for performance and power efficiency, rather than HD transistor density. After all, the 18A supports Powervia, a backside power delivery network, and may have performance and transistor density advantages over TSMC’s N2, and does not support this feature. However, this does not mean using PowerVia per 18A chip.
Regarding Power, TechInsights analysts assume that N2-based chips consume less power than similar SF2-based ICs, as N2-based chips have typically led to power efficiency in recent years. As for Intel, this is not yet known, but at least 18A offers advantages in this area.
There are a few other things. Intel’s 18A is set to take part in mass production in mid-2025 when Intel begins production of its core Ultra 3 series “Panther Lake” processors. In contrast, TSMC’s N2 is scheduled for mass production in late 2025, and the first products produced on this node will not be available until mid-2026, when mass market products are expected in fall 2026. If SF2 is accurately entering HVM, and only lists “2025”, this makes sense at any time from the first quarter of this year to Q4.
