According to a Digitimes report, Intel is slated to become TSMC’s second-largest customer right next to Apple in the coming years. Initially, rival AMD was supposed to take this position the present year. However, considering the size and spending power of Team Blue, it wouldn’t be surprising if it actually seizes the opportunity right from under the red rug.
Intel isn’t planning to abandon or spin off its in-house fabs though. According to the report, the agreement has already been concluded and it will be the largest deal between the two companies to date. Intel will leverage the foundry’s 3nm node to produce its 13th Gen Core-class processors, with mass production slated for the second half of 2022.
Although the new Intel CEO claimed that Intel will manufacture most of its 2023 CPUs in-house, sources claim that the deal with TSMC had already been concluded by the time Gelsinger was brought in.
The main goal behind outsourcing their core products to TSMC is to spread the risk over a wider range of production facilities and avoid a repeat of the 10nm scenario. This will allow Intel to research and develop advanced processes without having to ramp up yields for relatively immature nodes as soon as they are fit for production. While the post states that the cooperation between Intel and TSMC will continue with the 2nm process as well, that’s going a bit too far.
That decision will fall to Pat Gelsinger and depending on how Intel’s own foundries do in the coming years. However, there’s a very good possibility that certain families will be outsourced to ease the strain off the in-house fabs.
What does this mean for AMD? This simply implies that the competition for higher capacity at the Taiwanese Foundry will become tougher and tougher. Although this won’t have an effect in the near future as all of AMD’s products leverage the 7nm node, after which they will be migrated to the 5nm EUV process. Post that, the competition will become much tougher as both the chipmakers will be using the same process technology, and the end result will be decided by just the respective micro-architectures and packaging technologies. It’ll also give Samsung a chance to take back some of its lost shares in the pure-play foundry market.