Intel's Process Roadmap to 2025: with 4nm, 3nm, 20A and 18A?!
by Dr. Ian Cutress on July 26, 2021 5:00 PM ESTSidebar on Intel EUV
In all of these announcements, one thing to highlight is Intel mentioning its relationship with ASML, the sole company that manufactures the EUV machines powering production of leading edge semiconductor manufacturing.
ASML Wilton
ASML is a unique company in that it is the only one that can produce these machines, because the technology behind them is often tied up with its partners and research, but also because all the major silicon manufacturers are heavily invested in ASML. For any other company to compete against ASML would require building a separate network of expertise, a decade of innovation and design, and a lot of capital. None of the major silicon vendors want to disturb this balance and go off on their own, lest it shuts them out of the latest manufacturing technology, and no research fund sees competing against the embedded norm as a viable opportunity. This means that anyone wanting EUV specialist technology has to go to ASML.
In 2012, it was reported that Intel, Samsung, and TSMC all invested in ASML. This was, at the time, to jumpstart EUV development along with migrating from 300mm wafers to 450mm wafers. While we haven’t moved to 450mm wafers yet (and there are doubts we will any time in the next decade), EUV is now here. Intel’s 2012 investment of $2.1 billion gave them a 10% stake in ASML, with Intel stating that it would continue investing up to a 25% stack. Those stakes are now below the 5% reporting threshold, but all three of the major foundry customers are still big owners, especially as ASML’s market cap has risen from $24 Billion in 2012 to $268 Billion in 2021 (surpassing Intel).
As major investors but also ASML’s customers, the race has been on for these foundries to acquire enough EUV machines to meet demand. TSMC reported in August 2020 that it has 50% of all EUV machines manufactured at ASML for its leading edge processes. Intel is a little behind, especially as none of Intel’s products in the market yet use any EUV. EUV will only intercept Intel’s portfolio with its new Intel 4 process, where it will be used extensively, mostly on the BEOL. But Intel still has to order machines when they need them, especially as there are reports that ASML currently has backorders of 50 EUV machines. In 2021, ASML is expected to manufacture around 45-50 machines, and 50-60 in 2022. The exact number of machines Intel has right now, or has ordered from ASML, is unknown. It is expected that each one has a ~$150m price tag, and can take 4-6 months to install.
With all that being said, Intel’s discussion point today is that it will be the lead customer for ASML’s next generation EUV technology known as High-NA EUV. NA in this context relates to the ‘numerical aperture’ of the EUV machine, or to put simply, how wide you can make the EUV beam inside the machine before it hits the wafer. The wider the beam before you hit the wafer, the more intense it can be when it hits the wafer, which increases how accurately the lines are printed. Normally in lithography to get better printed lines, we move from single patterning to double patterning (or quad patterning) to get that effect, which decreases yield. The move to High-NA would mean that the ecosystem can stay on single patterning for longer, which some have quoted as allowing the industry to ‘stay aligned with Moore’s Law longer’.
| ASML's EUV Shipments | |||||||||||||||||||||
| 2015 | 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | |||||||||||||||
| Actual | 2 | 4 | 10 | 3 | 4 | 5 | 6 | 4 | 7 | 7 | 8 | 4 | 7 | 14 | 8 | 7 | 9 | - | - | ||
| Target (Total) | - | - | - | 20 (18) | 30 (26) | 35 (33) | 45-50 | ||||||||||||||
| 2018 and beyond is split per quarter for actual shipped numbers Data taken from ASML's Financial Reports |
|||||||||||||||||||||
Current EUV systems are NA 0.33, while the new systems are NA 0.55. ASML’s latest update suggests that it expects customers to be using High-NA for production in 2025/2026, which means that Intel is likely going to be getting the first machine (ASML NXE:5000 we think) in mid-2024. Exactly how many High-NA machines ASML intends to produce in that time frame is unknown, as if they flood the market, having the first won’t be a big win. However if there is a slow High-NA ramp, it will be up to Intel to capitalize on its advantage.
Facebook
Twitter
RSS
326 Comments
View All Comments
Oxford Guy - Wednesday, August 11, 2021 - link
Says the person who said, in reply to one of my posts ‘we know you’re smart ... use your powers for good’.You typically post whatever sounds reasonable at a given time, no matter how inaccurate it is. I, by contrast, am capable of remembering what has been said — the positions that have been taken.
One cannot simultaneously claim I’m obviously intelligent and that my posts are valid. One cannot also post ‘agreed’ — as you did in another topic whilst pretending that my posts are vapid — unless you’re with that very vapidity.
I also find it droll that you employ the royal ‘we’ here. Are you a member of staff or merely that entitled?
Oxford Guy - Wednesday, August 11, 2021 - link
And... please — for the benefit of this forum...Learn the list of common logical fallacies. Your latest use of ad baculum is only worthy of yet another eyeroll.
mode_13h - Thursday, August 12, 2021 - link
> Says the person who said, in reply to one of my posts ‘we know you’re smart> ... use your powers for good’.
There's no logical inconsistency, there. Most of your posts seem to pull threads off-topic and offer little of value to the original subject. And quite a few are just snarky, cynical trolls.
> You typically post whatever sounds reasonable at a given time
I try to engage my brain and look at the other side of an issue, or at least from a perspective other than my narrow self-interest. And most often, what draws me to the other side of an issue is when someone takes an extreme position or makes absolutist statements that seem unjustified. If there's one thing you could say I consistently oppose, it's oversimplification.
> no matter how inaccurate it is.
Ah, now that's interesting. Accuracy is rooted in fact. And the facts are where you completely fall apart. You consistently fail to support your claims and assertions with good & relevant sources.
With that said, if I post something that's demonstrably inaccurate, then please do us *all* a favor and point it out. I never claimed to know everything or be infallible. I've even learned things from debates and spirited discussions.
> I, by contrast, am capable of remembering what has been said —
> the positions that have been taken.
This might blow your mind, but I have actually changed positions, on a few occasions. Not many, but I'm actually willing to re-evaluate my position, after looking at the arguments on both sides.
Also, I try not to be overly partisan, which is to say that I try not to take a side of an issue purely on the basis of political allegiance or preoccupation with self-consistency. If I think one side is overstating their case or otherwise acting in bad faith, I might come out against their position, even while I might've previously been supportive on another issue.
> One cannot simultaneously claim I’m obviously intelligent and that my posts are valid.
Why not? Intelligence describes the actor, while the writing of posts describes their actions. I can criticize the latter, without invalidating the former. Plenty of smart people do things that are thoughtless, counterproductive, antisocial, or worse. However, at some point, the actions do begin to define the actor.
> One cannot also post ‘agreed’ — as you did in another topic whilst pretending
> that my posts are vapid
I said they're "consistently", not "uniformly" or "without exception". If I thought you were a complete waste of time, then I wouldn't spend so much time replying to you.
> unless you’re with that very vapidity.
Well, I'm not going to claim I've never made a vapid post. I try to say things worth saying, but I'm not infallible. It is just a news comment thread, and I don't worry too much about a post here or there.
> I also find it droll that you employ the royal ‘we’ here.
It wasn't. I was speaking on behalf of myself AND other forum participants. That it was preceded by "I think", signifies it as a speculative statement. Others are welcome to disagree.
ikjadoon - Monday, July 26, 2021 - link
What? Intel has long sandbagged its numbers. Unfortunately, we've all decided to follow the marketing, so yeah, at least Intel is more honest now. But none of it matters until they deliver it. I'm not trusting any marketing announcements from Intel. I want the desktop / laptop CPU in-hand so that there's actual benchmarks.//
https://www.tsmc.com/english/dedicatedFoundry/tech...
Otherwise there's no need for the TSMC marketing dept to magically shrink the fake 16nm node to become a fake 12nm.
>An enhanced version of TSMC's 16nm process was introduced in late 2016 called "12nm".
https://en.wikichip.org/wiki/16_nm_lithography_pro...
mode_13h - Monday, July 26, 2021 - link
> at least Intel is more honest now.Wow, that sure takes some mental gymnastics to see Intel participating in the same disinformation race as "more honest".
You could say they're being more consistent... until TSMC and Samsung decide to rebrand their process nodes to stay ahead of Intel's naming.
All of this argues that it's an exercise in futility to pretend these names actually mean anything. They should just use codenames, or maybe a completely arbitrary schema involving sequential numbering + letters or Greek alphabet characters.
ikjadoon - Tuesday, July 27, 2021 - link
Sure, noted. "More honest" relative to the industry = more consistent. I'm flabbergasted how anyone has any problem with this, when literally no one had a problem TSMC & Samsung have done this for years, lol.OK? Why wouldn't TSMC & Samsung play more bullshit w/ foundry marketing? They started it a while ago, so it's more than expected to continue. Good technology has never needed exaggerations: Intel, TSMC, and Samsung all know that.
lol, this is just the tip of the iceberg of marketing. We don't need "i7" or "Ryzen 3", either. How deep do you want to go?
Node names *absolutely* mean something: it's the progression within a foundry. Almost nobody dual-sources CPUs any more, but everyone wants to play "Fantasy Nodes".
That's the more interesting problem. Why is the peak density leap between 10->7 larger than 7->4? Because, clearly, density is not the *only* metric involved in a node.
mode_13h - Wednesday, July 28, 2021 - link
> literally no one had a problem TSMC & Samsung have done this for years, lol.How do you know? Did you run a survey?
Unlike what Intel is doing, Samsung and TSMC never had a press conference to announce they're going to use more dishonest naming. If they had, you'd probably have seen the same kind of sentiment you're seeing when Intel did just that.
> Good technology has never needed exaggerations
That's not true. Not as long as exaggerations can help you sell a little more. Nvidia exgerates like all damn day, even while they've been sitting comfortably atop the heap.
> it's the progression within a foundry
Right, so the names just need to reflect that. Like I said, they should use sequential numbering for big steps, and then letter suffixes to denote minor iterations.
> density is not the *only* metric involved in a node.
All the more reason to cut ties between their naming and any pretense of density.
wut - Thursday, July 29, 2021 - link
"Right, so the names just need to reflect that. Like I said, they should use sequential numbering for big steps, and then letter suffixes to denote minor iterations"Tell TSMC, Samsung, along with everyone else to do it, at the same time.
(TSMC with its N7+, N5P, and Samsung with its 3GAE...)
If you want to apply some standard, apply it to everyone first. Lest you'd be the one who ends up looking agenda-ladened.
mode_13h - Sunday, August 1, 2021 - link
> Tell TSMC, Samsung, along with everyone else to do it, at the same time.The beauty of it is that Intel can simply opt out of the game, without requiring others to do the same.
> If you want to apply some standard
No, you don't have to replace a false standard with another standard (false or not). The point is just to drop the pretense that the node names really mean anything.
twtech - Tuesday, July 27, 2021 - link
A really bold move would have been to move away from "nm" naming altogether and call it 100D or something for 100 million transistor density.