Intel Pursues Moore's Law With Plan To Make 7-Nm Chips This Year

Megalith

24-bit/48kHz
Staff member
Joined
Aug 20, 2006
Messages
13,000
It will still be quite a while until we can get our hands on this technology, but Intel is setting up a pilot plant as early as this year to test and iron out the kinks in manufacturing 7-nm chips. These chips, which will follow the 10-nm process, are expected to comprise exotic materials, which will allow them to be significantly smaller and more power efficient.

Intel is looking at the 7-nm process to alleviate some of the challenges it faces on the 14-nm and 10-nm processors. The company has hinted it would introduce EUV (extreme ultraviolet) tools in the manufacturing process. EUV will help etch finer features on chips, but its implementation has been delayed multiple times. The pilot factory will help validate all those features, and then allow Intel to order equipment for the new factories, McCarron said. Competitors like Globalfoundries and Samsung are getting a head-start on the 7-nm process. Globalfoundries has said it will start making 7-nm chips by 2018, and ARM has released tools for the design of 7-nm chips.
 
Last edited:
Remember foundry roadmaps and product roadmaps are entirely different things.

Global Foundries 7nm is worse than Intels 10nm as compare and may be another 14XM node (Read never gonna happen). TSMCs 7nm PR node is also bigger than Intels 10nm. Samsung got "7nm" way last of the foundries and actually considers to get rid of its foundry business.

In terms of EUV if anyone wonders, its as far away as always so to day.

7nm CPUs looks to be around 2020-2021.

Expect even more node PR renaming in the future as its getting harder and harder.
 
Last edited:
They are just looking looking for a excuse to put R&D into creating an UltraBlack field... So they can finally start messing around with the timeline.
 
Wow. I knew we'd reach 7nm, but I figured it would be further off.

At what lithography size does it become impossible to shrink it anymore, again?
 
Wow. I knew we'd reach 7nm, but I figured it would be further off.

At what lithography size does it become impossible to shrink it anymore, again?

Companies are already struggling with the economics below 28nm. And even big companies like MediaTek are struggling at 10nm.

The question isn't how low we can go, but how many that can afford it. For a "real 10nm" you may list less than 5 companies that can make money on it. The economic limit for now may be 5nm with maybe only a single company.
 
My crystall ball says don't expect 10nm from Intel before Q4 2017, and 7nm before Q2 2019.
 
Companies are already struggling with the economics below 28nm. And even big companies like MediaTek are struggling at 10nm.

The question isn't how low we can go, but how many that can afford it. For a "real 10nm" you may list less than 5 companies that can make money on it. The economic limit for now may be 5nm with maybe only a single company.


I wonder if GaN will change this at all...
 
Back
Top