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
GeoffreyA - Friday, July 30, 2021 - link
"it makes more sense to use nonsense-numbers backed up by half-truths."Fantastic.
mode_13h - Sunday, August 1, 2021 - link
> it makes more sense to use nonsense-numbers backed up by half-truths.Or just drop any pretense of meaning and use some other scheme, such as Major.Minor version numbers. Or, even just use sequential integers.
mode_13h - Monday, July 26, 2021 - link
> Intel should've just moved to a transistor-density-based node naming.Agreed, but they could either continue using modifiers like SF and ESF or I'd say they should go with letters.
GeoffreyA - Friday, July 30, 2021 - link
Letters.Amandtec - Tuesday, July 27, 2021 - link
You can measure what you want and they will find a way to score better. Transistors are used for different things so they will start putting huge piles of useless high density transistors into the design to bump up the density measure. At the end of the day, they should stop talking about nodes altogether and talk about sing/multi/heat per dollar - of course, they can't do that because nobody would but another Intel chip for a couple years.TheJian - Tuesday, July 27, 2021 - link
Agreed, lets just go buy transistors, or list both. People can compare whatever makes more sense to them...LOL. Agree with OP too, joining the fake club while Intel tried to stay real for years. OH well, if you can't beat them, join them and come up with something new (2025)...LOL. Whatever works. Buying up 3nm was great too, will keep others from gaining far more NET INCOME from getting said 3nm wafers and competing with their own top end stuff. Direct those 3nm chips at as much server/mobile you can (maybe some desktop top end chips, the top3 maybe) or just make a TON of 3nm gpus with whatever is left all year. I hope Pat bought 18mo of 3nm so nobody else gets them. Apple got batch1, messed it up on the 2nd, so hopefully Intel bought 3-4B of them until the cash runs out on it. Make everything AMD thinks is a cash cow, and leave them with stuff on the shelf (and written off a year later...LOL).Anything still left? GO 3nm GPU to hurt everyone. Meaning, apple coming right after 3nm cpu (gpu just waiting for wafers surely to avoid AMD/NV next, developing now in home consoles likely), but longer now, going 4nm..LOL, AMD, NVDA NET hurt here too. I'll be shocked if pat didn't nail down even more than planned when apple stepped out of batch2.
30K wafers were batch1 - How many in batch 2? 100k? Make the largest gpus you can to eat them all (assuming AMD's top server/mobile dies are attacked already with coming chips), and watch as AMD/NV Q reports go down, and all 3 rejig their next parts on 4nm etc...ROFL (apple already did it). 6nm Warhol dead (well, the fake 6nm process that is). You get the point.
We all need 3nm gpus anyway right? Heck most of you would probably take ANYTHING these days...LOL. Intel to the gpu rescue Q1 2023 with 3nm TSMC all year (and 6nm coming up??)?? Sell them DIRECT like a NVDA founders cards, and only ONE per household AND/or visa etc. I think gamers would warm to Intel if they helped us get cards soon. 6nm sounds good to me if you can't get any 7nm yet really. No I refuse to pay 2-3x, heck for that matter I refuse to pay over MSRP. I'll wait, build HTPCs, etc etc. I won't bite until price isn't stupid even though I can afford it.
mode_13h - Wednesday, July 28, 2021 - link
> I refuse to pay 2-3x, heck for that matter I refuse to pay over MSRP.Dude, you *just* posted in another thread that people should stop whining about high component prices and live like peasants to save up whatever it costs to buy the latest GPU!
FWIW, I'm with you. I'm fortunate that I don't need the latest & greatest GPU, so I plan to wait until I can get one at a reasonable price.
Spunjji - Wednesday, July 28, 2021 - link
TheJian doesn't make sense *within* a single post, let alone between them.Spunjji - Wednesday, July 28, 2021 - link
"I hope Pat bought 18mo of 3nm so nobody else gets them."Why? That would fuck things up for everyone.
Spunjji - Wednesday, July 28, 2021 - link
Agreed. It doesn't make much sense to decry the disconnect between the numbers and reality, and then just jump on board with the existing nonsense!