Intel's Process Roadmap to 2025: with 4nm, 3nm, 20A and 18A?!
by Dr. Ian Cutress on July 26, 2021 5:00 PM ESTCustomers Customers Customers
As a roadmap announcement today, the focus isn’t so much on the customers but on the technology. Because Intel is moving into a phase where it expects its IFS offerings to compete against the established players, it has to consider its disclosures with respect to both its internal use and any external interest, which is a new concept for the company – at least on this scale compared to its previous foundry efforts.
Intel CEO Pat Gelsinger, in the company’s Q3 financial call last week, was keen to point out that they already have a large hyperscaler customer signed up for their next generation packaging technology, however today there would appear to also be another customer in the mix. Now we assume that Intel’s Foundry Services is talking to 100s of chip companies, big and small, but it doesn’t take much to sign an NDA to start to talk – what will be interesting is when customers start making commitments to using Intel’s facilities, and if any of those are volume orders.
As part of the announcement today, Intel held a little bit back from us, saying that they are saving some of the details specifically for the event that is going on as we publish this piece. All we know is that our draft press release has a big yellow bar that says ‘[customer news]’ on it, right next to Intel’s 20A process node details.
For reference, Intel 20A is a 2024 technology using first generation Gate-All-Around transistors, marketed as RibbonFETs, as well as backside power delivery, marketed as PowerVias. At this time Intel expects to have second/third-generation EMIB available as well as fourth-generation Foveros Direct. So if a customer is already committing to Intel 20A, there’s going to be a lot of potential here.
When the announcement is made, we will update this news article.
To conclude, Intel maintains that these roadmaps will showcase a clear path to process performance leadership* by 2025. It’s a tall order, and the company has to execute better than it has in recent memory - but that’s kind of why the company has rehired a number of former Intel experts and fellows in research, product design, and execution.
*as measured by performance per watt at iso-power
Here's a secondary comparison chart (compared to the one on page one) with all three main foundry offerings listed in each of the main segments that Intel has discussed today.
Facebook
Twitter
RSS
326 Comments
View All Comments
mode_13h - Thursday, July 29, 2021 - link
Okay, makes sense. Thanks for that.name99 - Tuesday, July 27, 2021 - link
ACKNOWLEDGE their faults?That’s why they used that very careful language to say that they are shipping more WAFERS at 10nm than 14nm, but not more CHIPS? Because they still don’t want to admit 10nm yield issues…
Such transparency. Much honesty.
Spunjji - Thursday, July 29, 2021 - link
Yeah, I do wish little tells like that would be pointed out in the main article.WaltC - Wednesday, July 28, 2021 - link
Intel needs products--vaporware will not help them. Intel no longer "runs" the industry, if you haven't noticed.JasonLD - Monday, July 26, 2021 - link
This is exactly the reason why Intel is renaming their nodes like TSMC/Samsung.TheJian - Tuesday, July 27, 2021 - link
You seem completely unaware of the OTHER ways to fight. IE: Buying out the other guy's next wafer requirements to launch 3nm. Intel already killed 6nm Warhol plans (maybe NV too, they're 5nm now). The 3nm 2nd batch already killed Apple 3nm macs next year too.You are ignorant. The simple thing for Intel to do here, is just buy up everything they think AMD/NV could need making DISCRETE gpu anything (server, workstation, desktop all discrete), and make the crap out of them at 3nm and flood the market vs. all others. They won't be competing with their own stuff for the most part as they have no discrete retail cards for ages.
What is 2nm batch 100K wafers? First was 30k. I can't remember the link to the ramp stats they expect (semiwiki or wikichips, eetimes, semiconductor-digest etc - too many to keep track of and many need translators). Anyone?
Spunjji - Thursday, July 29, 2021 - link
You seem to be awfully confident that the non-appearance of products that were only ever rumoured from leaked roadmaps of an indeterminate age is down to Intel. You also seem to think that the abusive, self-defeating behaviour you hypothesise would be smart. The mind boggles.mode_13h - Sunday, August 1, 2021 - link
> You also seem to think that the abusive, self-defeating behaviour you hypothesise would be smart.Yeah, I can't imagine TSMC would sell out its 3 nm production run to a competitor of its own foundry and a competitor to many of its other customers. I mean, talk about self-defeating...
Spunjji - Thursday, July 29, 2021 - link
Not gonna lie, 8-core Tiger Lake is a solid product. There are trade-offs vs the AMD competition, but it's an ambitious design and no doubt about it - albeit apparently one they can't make in sufficient volume for the desktop, even though it operates with desktop boost power levels and desktop I/O capabilities. Alder Lake looks even more interesting architecturally, though it remains to be seen how that works in practice.dubeg - Monday, July 26, 2021 - link
Typo: "We asked Intel is a pre-briefing"