TSMC’s cutting-edge 2nm (N2) node is slated to undergo mass production sometime in 2025. As is the tradition with the Taiwanese foundry, Apple will be the first client of the process but foundry rival Intel will be joining in this time around. According to sources from Taiwan, Intel and TSMC have already started talks to allocate a chunk of early 2nm capacity to the former.
Intel plans to begin mass production of its 2nm (20A) node in the first half of 2024, followed by a release later that year, effectively giving it a process advantage of over 13-15 months. From what we know, the chipmaker will be using this advanced capacity to manufacture its client (Core) and server (Xeon) processors. Meanwhile, the contemporary Arc GPUs will be produced using TSMC’s 3nm and 2nm process nodes.
Prior to that, the 14th Gen Meteor Lake processors will be fabbed on Intel’s 4nm node. This will be a tiled architecture with the iGPU die outsourced to TSMC’s 3nm. 2nm Arrow Lake will use the same process node for its iGPU chiplet the following year. It’s unclear whether Intel will adopt the N2 process to manufacture iGPU dies of succeeding processors.