News Intel to reveal its roadmap beyond the 18A (1.8nm) process node in a few weeks — the company will share its post-5N4Y plans during February event

[Admin]

Intel vows to disclose its post-5N4Y roadmap at its IFS Direct Connect event in February.

Intel to reveal its roadmap beyond the 18A (1.8nm) process node in a few weeks — the company will share its post-5N4Y plans during February event : Read more

 

[bit_user]

What to expect from Intel beyond 18A is up in the air ...

I'd wager some form of CFET transistors will factor into their announcements:

https://www.tomshardware.com/news/intel-and-tsmc-to-report-on-next-gen-cfet-transistor-progress
​

For anyone who hasn't already picked up on this, the semiconductor industry is broadly following a roadmap set forth by IMEC. If you want the big picture view of where things are headed, look to what they're saying:

https://www.tomshardware.com/news/imec-reveals-sub-1nm-transistor-roadmap-3d-stacked-cmos-20-plans
​

 

[Tom Sunday]

The INTEL stock has shown some very good results or comeback lately! Internationally: Their Magdeburg plants (2) in East Germany (formerly the DDR) has garnered them as of late much increased subsidies or seed monies than originally expected. Intel is now set to receive almost US$11 billion instead of the €6.8 billion (US$7.2 billion) in government aid. Then later this year for our enthusiasts here we will have their 15th generation of CPU’s (Arrow Lake) breaking loose. It’s always good to see an all American company doing well!

 

[bit_user]

Internationally: Their Magdeburg plants (2) in East Germany ... It’s always good to see an all American company doing well!

Or, in other words, not quite "all American". Very much "multinational", I'd say.

 

[ottonis]

Intel keeps keeping the press busy with plenty of plans and visions, e.g. the 1.8 nm process and so on and so forth, but the market is yet to see commercially available Intel chips fabbed on a smaller than a SuperFin 10 nm process.
Intel's first 10nm process was first introduced with the CannonLake CPUs (2018) and later refined as SuperFin 10 nm and renamed to "Intel 7".

 

[bit_user]

Intel keeps keeping the press busy with plenty of plans and visions, e.g. the 1.8 nm process and so on and so forth, but the market is yet to see commercially available Intel chips fabbed on a smaller than a SuperFin 10 nm process.

Meteor Lake is made on what they used to call their 7 nm node. It's EUV, as opposed to "Intel 7", which is still DUV + multipatterning.

later refined as SuperFin 10 nm and renamed to "Intel 7".

What they renamed to "Intel 7" is the node they called "10 nm Enhanced Super Fin". Tiger Lake was made on the original 10 nm SF, and that node wasn't renamed.

 

[Giroro]

Is Intel itself actually reffereing to 18A as "1.8 nm class"? Or is this just a case of taking the bait and running with the assumptions built into their cynical naming scheme?

 

[thestryker]

At this point I wish the pretense of "nm" with regards to process nodes would just be dropped. Smaller number = better core node is basically what they've all shifted to. There are also more modifications to nodes than ever before muddying the waters even further.

I look forward to seeing what Ann Kelleher has to say as she is always interesting. The last interview I saw there wasn't a whole lot about the more advanced nodes she could say so this should deliver.

 

[JamesJones44]

Is Intel itself actually reffereing to 18A as "1.8 nm class"? Or is this just a case of taking the bait and running with the assumptions built into their cynical naming scheme?

All of them are "cynical naming scheme"s. TSMCs 4nm isn't 4nm for example. Everything you see from Samsung, TSMC, Intel, etc. is marketing names, not actual transistor size. The advertised nm measurement equaling the actual size of a transistor went obsolete decades ago.

 

[bit_user]

All of them are "cynical naming scheme"s. TSMCs 4nm isn't 4nm for example. Everything you see from Samsung, TSMC, Intel, etc. is marketing names, not actual transistor size.

I think Intel used to claim that its node was the fin pitch of the Fin FET transistor - the smallest feature they contained. What's interesting about that is the fins weren't made that size by lithography, but by a chemical etching technique which eroded them down to that size.

 

[JamesJones44]

I think Intel used to claim that its node was the fin pitch of the Fin FET transistor - the smallest feature they contained. What's interesting about that is the fins weren't made that size by lithography, but by a chemical etching technique which eroded them down to that size.

I believe they did as well. I've read at one time Intel's advertised size (45 nm for example) was actually larger than their smallest feature, but they wanted to pad there numbers just in case they ran into issues with the next node and could say they did a small shrink even though it was actually the same node process. I don't know how true those rumors are, but I thought it was interesting that Intel at one time my have actually be fudging the numbers in the other direction and claimed to be larger than they were.

 

[bit_user]

I thought it was interesting that Intel at one time my have actually be fudging the numbers in the other direction and claimed to be larger than they were.

Back then, Intel had unrivaled process supremacy. So, there was no real harm in keeping a little "gas in the tank".