Having docked at 14nm for too long, Intel has fallen significantly behind TSMC and Samsung in nominal process. At a recent conference at the Damo TMT, Intel CFO George Davis admitted that 10nm would not be as productive as 14nm and 22nm. His words have two meanings, one is with the 14nm parallel state, 10nm product itself is limited, and the other is 10nm will only improve 1 to 2 generations (10nm plus, 10nm, 10nm, ” in order to quickly launch 7nm.
Davis confirmed that the first generation of the improved 10nm-plus process Tiger Lake processor will be available by the end of this year.
He also revealed that he expects to produce 7nm products by the end of 2021, then quickly switch to 5nm and regain leadership in the process.
This “leadership” needs to be explained, as both TSMC and Samsung are at least 3nm when Intel 5nm is present. But TSMC and Samsung are only nominal processes, and transistors are more dense than Intel’ generation. In other words, Intel is simply back to the “no-squeeze toothpaste” state before 14nm.
Even so, Intel pays a small price.
Davis to the outside world sounded the alarm, 10nm mass production, 7nm speed, 5nm investment, these are the need for money, considering that this part of the technology is mainly concentrated in 2020-2021, will inevitably affect Intel gross margin.