At the moment, Intel’s foundries are lagging behind both Samsung and TSMC in terms of process node technology. The chipmaker has just started moving the bulk of its processors to the 10nm node which was originally slated to arrive some five years back. TSMC and Samsung, on the other hand, have been using their equivalent 7nm node for several years now. TSMC has even started the mass production and deployment of its 5nm process with EUV lithography in the form of Apple’s M1 and A14 processors.
Intel’s answer to TSMC’s 5nm node is its 7nm technology which is slated to launch sometime in 2023 with Meteor Lake. The company has already invested over $20 billion to construct advanced fabs in Arizona and Ireland, with more investments planned in the future (thanks to aid from the Biden administration). The company recently declared that it had completed the tape-in of the Meteor Lake architecture. On the server-side, the 5th Gen Xeon-SP codenamed Granite Rapids (the successor to Sapphire Rapids) will be the first to leverage the 7nm process.
Following Meteor Lake and Granite Rapids on 7nm, the next step is going to be an even more advanced node with a peak transistor density of 400MTr/mm2 (as per analysts). This process will be called Intel’s 5nm node, being 4x denser than its 10nm node and nearly on par with TSMC’s 2nm node which will have a transistor density of 500MTr/mm2 (only 20% higher).
However, going by estimates, TSMC’s 2nm node with GAA technology should launch before Intel’s 5nm process which has still not been officially announced by the US chipmaker.