This question will be resolved based on official information released by the relevant companies using the companies’ publicly available quarterly reports and mainstream news sources. Question will be resolved based on official information released by the relevant companies (e.g., Samsung, TSMC, Intel). Note that “volume production” in this question does not include “trial” or “low volume" production. For the purposes of this question, "3nm or smaller" refers to the name of the node used by the manufacturer, not any particular physical measurement.

Although the industry is accustomed to TSMC releasing new product every two years (ANANDTECH), TSMC’s volume production of the 3nm chip is not expected to begin until the second half of 2022 (TSMC, Business-Standard). Samsung has plans to roll out their 2nm design in 2025, but volume production of the 3nm chip was delayed from 2021 to 2022 (Electronic Design, DigitimesAsia). Intel’s roadmap calls for them to overtake their competition by releasing a 1.8nm chip by 2024 (AIM).  

Additional Reading: IBM announced in 2021 it created the world’s first 2nm chip (CNN). Although IBM doesn’t mass produce chips, it will license its design to chipmakers, further driving competition. IBM has a historical relationship with INTEL, which could make the 2nm chip design the key for INTEL’s competitive edge (ServerWatch). 

Issued on 03/18/22 08:50pm
The naming of process nodes by semiconductor manufacturing companies is not necessarily tied to a physical gate length or half-pitch. For the purposes of this question, "3nm chip or smaller" refers to the commercial name of the node, rather than an actual physical measurement.