It is rumored that Intel’s 18A process yield rate reaches 80% and may cooperate with Huida to develop AI CPU

Technology online media "Wccftech" reported on the 15th that analysts from Korean financial company GF Securities pointed out that Intel and Huida may jointly build artificial intelligence (AI) central processing units (CPUs), and Intel's 18A process yield has significantly improved to 80%.

The report said that for Intel's foundry business, the 18A and 14A processes are key technologies, and Intel hopes to use this technology to gain a firm foothold in the global chip foundry field. In the current AI era, the industry is booming, but almost all orders for high-end AI chips go to TSMC. According to revenue, in the second quarter of this year, TSMC accounted for 72% of the global market, consolidating its leading position in global chip foundry.

In this context, Intel boldly claimed that its 14A and 18A improved version 18A-P process technology has made great progress. 14A is Intel's next-generation technology platform, equivalent to TSMC's 1.4nm platform.

GF Securities analyst Jeff Pu reported that the current yield rate of Intel's Panther Lake chip 18A process has reached 80%. His report in July said that in the second quarter of this year, the yield rate of the 18A process was 65%.

Jeff Pu also pointed out that Intel's next-generation 14A process series is expected to produce 6,000 wafers per month in 2027 and 24,000 wafers per month in 2028. This number is in sharp contrast to TSMC’s 1.4nm process series. TSMC reported that due to the accelerated construction progress of the wafer fab, it is expected to reach mass production in 2027.

The report also said that Intel may cooperate with Huida in the field of personal computing. Jeff Pu even claimed that Rosa Feynman will launch a version equipped with Intel CPU. Previous reports on Huida's next-generation graphics processor (GPU) said that the Rosa CPU may rely on TSMC's 2nm or 1.6nm process technology.