Intel is currently stuck on 14nm lithography because they wanted to go too far with 10nm. Bob Swan, CEO of the company indicated a few months ago that improving the density by x2.7 times with respect to 14nm had still been excessive. This has made producing 10nm chips complicated and dependent on a lithography that does not give more of itself.
All of this at a time when AMD is making a steady pace with its Zen architecture processors. Making the leap in the Ryzen 3000 to TSMC's 7nm lithography has allowed them to take a leap in performance. This together with the adjusted prices is causing a sharp increase in sales.
[amazon box="B005404P9I"]Intel talks about its future lithographs
Intel's problem with 10nm is that the number of valid chips per wafer would still be low. A productivity problem that means that the company should sell its processors at least twice the current price. Logically, if they are now expensive and fall in sales, doubling prices would be crazy.
The company expects to expand processor production by 10nm next year based on improved lithography (10nm +). After months of silence the company has put black on white regarding its lithographs and when they should be released.
Despite all the current problems, the company indicates that in 2021 the 7nm lithography will be ready. It will be at that moment when they will return to the Tic-Tac process. By 2023 they expect to launch 5nm, by 2025 they expect to launch 3nm, by 2027 they expect 2nm and by 2019 they expect 1.4nm. Intel also highlights that they have invested a lot of money to make up lost ground in research and development. They have also increased investment in production plants this year and will continue to do so in 2020.
According to Intel, its 7nm will be equivalent to TSMC's 5nm and will also use extreme ultraviolet light (EUV). In addition, the company indicates that by 2021 they will be able to use advanced processes to manufacture FPGAs, ASICs and processors for laptops.
New lithography strategy
Intel has commented something quite curious about lithographs. The company indicates that they are working on porting advanced nodes to old lithographs. For us to understand it, it would be working to port the 10nm node to the 14nm node in order to take advantage of the latter's production capacity. Remember that Intel currently has more demand than production capacity in 14nm.
The company indicates that these special nodes will be named in a special way. Commercially the 7nm node manufactured under the 10nm node will be called 10nm +++. The 5nm for the 7nm node will be called 7nm ++ and so on. Yes, it is something complex to understand and a rather curious strategy, no doubt.
In addition, the company indicates that they trust a lot in Moore's Law. Hence, they present this very aggressive plan in terms of lithographs, returning to the Tic-Tac system that I abandoned with 14nm. For this they have invested billions of dollars in R & D & i.
Remember that Bob Swan, CEO of Intel a few days ago indicated a change in the company's strategy. It stands out that they will forget 90% of the market share of domestic processors for go for 30% of the silicon market. It indicates that the strategy they had devised was disastrous and hence the current problems of the company.
Source: Anandtech
