Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/intel-earnings-announcement-gloomy-2023.17375/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Intel earnings announcement - gloomy 2023

The transcript is worth reading. Here are some quick cuts from Pat on the manufacturing/foundry side.

"On Intel 20A and Intel 18A, the first nodes to benefit from RibbonFETs and PowerVia, internal test chips and those of a major potential foundry customer have taped out with the silicon running in the fab. We continue to be on track to regain transistor performance and power performance leadership by 2025."

"Additionally, we continue to make progress on Intel 18A, and I've already shared the engineering release of PDK0.5 with our lead customers and expect to have the final production release in the next few weeks."

"We will, one, deliver on five nodes in four years, achieving process performance parity in 2024 and unquestioned leadership by 2025 with Intel 18A.... and four, expand our IFS customer base to include large design wins on Intel 16, Intel 3 and 18A this year."

"IFS achieved record quarterly revenue of $319 million, up 87% sequentially and 30% year-over-year on increased automotive shipments."

As a foundry person I am quite pleased. If Intel's PDKs are in the same class as TSMC this is great news. If Intel PDKs are in the same class as Samsung this is not great news. From what I have heard the Intel 18A PDK is good and there will be tape outs at the end of this year / early next year.

Pat did say they are still working on the Tower Semi acquisition but I have serious doubt it will happen in Q1 as planned or if it will happen at all. US relations with China are getting worse so I don't see it happening. No fault of intel of course. I still think it would be a good thing but not critical to the success of IFS.

 
Interesting how as Intels tech issue are being ironed out their financials are falling apart. Complete opposite of the last 5-8 years.
 
Gelsinger doesn't have a lot of credibility with his unquestioned process leadership claim by 2025, not after his remark that AMD will never again be in their windshield. The bravado is getting old.
 
The transcript is worth reading. Here are some quick cuts from Pat on the manufacturing/foundry side.

"On Intel 20A and Intel 18A, the first nodes to benefit from RibbonFETs and PowerVia, internal test chips and those of a major potential foundry customer have taped out with the silicon running in the fab. We continue to be on track to regain transistor performance and power performance leadership by 2025."

"Additionally, we continue to make progress on Intel 18A, and I've already shared the engineering release of PDK0.5 with our lead customers and expect to have the final production release in the next few weeks."

"We will, one, deliver on five nodes in four years, achieving process performance parity in 2024 and unquestioned leadership by 2025 with Intel 18A.... and four, expand our IFS customer base to include large design wins on Intel 16, Intel 3 and 18A this year."

"IFS achieved record quarterly revenue of $319 million, up 87% sequentially and 30% year-over-year on increased automotive shipments."

As a foundry person I am quite pleased. If Intel's PDKs are in the same class as TSMC this is great news. If Intel PDKs are in the same class as Samsung this is not great news. From what I have heard the Intel 18A PDK is good and there will be tape outs at the end of this year / early next year.

Pat did say they are still working on the Tower Semi acquisition but I have serious doubt it will happen in Q1 as planned or if it will happen at all. US relations with China are getting worse so I don't see it happening. No fault of intel of course. I still think it would be a good thing but not critical to the success of IFS.

The mention of snagging a cloud provider for intel 3 is surprising. Given how long the design cycle is, and MediaTek saying that they were getting their i16 chips in 2024; I figured that i3 would be almost unused by external IFS customers (unlike the 20/18A family which could be a real winner). It is also sad to see that NEX seems to be at the end of the line since it seemed to be a good performer. A shame, but as they say better to amputate the limb than lose the whole body.

As for the 18A PDKs looking good, any specifics on how they look relative to the competition? Or more so that they are progressing well enough without any major snags?
 
Last edited:
Intel's new node mass production could mean low yield and low margin in the first 12-18 months while TSMC's new node becomes very profitable at once.

After Pat claimed SR was not delayed again in the last earning release, his credibility decrease to me.

Most likely Intel will deliver new nodes with very small qty and claim big achievement.
 
So much is made by folks both in Intel and outside about 18A being unquestionably the best node when it comes out. Is there any credence to that. I find it hard to believe TSMC won’t have an equal or better process ready to defend against 18A
 
  • Like
Reactions: VCT
So much is made by folks both in Intel and outside about 18A being unquestionably the best node when it comes out. Is there any credence to that. I find it hard to believe TSMC won’t have an equal or better process ready to defend against 18A
i4 is about at parity with N3E. All one needs to do is look at intel and TSMC's roadmaps going forward to see that IF everyone does what they say they will do, then intel will without a doubt have a PPW advantage and at the very least density parity with N2. When they add BSPD to an N2P or N1.8 they will probably have a density lead over 18A, but at that point we are talking 2026/27. Whatever is after the N2 and 20/18A family is where the real race will be. 1st generation HNSs are being conservative on density (N2/3GAE also seems to only be 3 sheets which will hurt their PPA).

Obviously intel needs to hit it's roadmap milestones for this to be the case, and they have not earned the confidence that TSMC has earned. Given their parallel node development and how long intel has had to mule over GAA during their extended time in the finFET era, it is not totally unbelievable that intel already has a good body of research to speed along their HNS development efforts.

Both fortunately and unfortunately nothing is certain, as GAA is the great equalizer. Time will tell who will have the better start in this new era, and who will end this chapter of the transistor's never ending advancement in the technological lead. I excitedly await the results, but one thing I can say with certainty is that regardless of who has the best process TSMC can still hold onto their claim of being "the most trusted foundry".
 
Last edited:
The mention of snagging a cloud provider for intel 3 is surprising. Given how long the design cycle is, and MediaTek saying that they were getting their i16 chips in 2024; I figured that i3 would be almost unused by external IFS customers (unlike the 20/18A family which could be a real winner)
Do you know what company this is?

Whats interesting is that is Intel is working on RISC-V SoCs with SiFive on Intel 4. Surprising given its not really a full node.
 
Do you know what company this is?

Whats interesting is that is Intel is working on RISC-V SoCs with SiFive on Intel 4. Surprising given its not really a full node.
No and if I did I wouldn’t talk about it until it went public.

Another interesting thing that was mentioned is that ifs has inked deals worth over 4B. Obviously this is spread out across multiple quarters, and it probably isn’t revenue we will see until 2024; but IMO it is a real good start for a division that is less than 2 years old. The closest thing to compare to would be DCAI, which for this quarter did 4.3B in sales (and this is with like 75% of the world’s total server market). I am certainly excited to see the specifics of the new IFS deals later this year, because at this point I think it is clear that this is not another tenith-hearted attempt like intel custom foundry was.
 
No and if I did I wouldn’t talk about it until it went public.

Another interesting thing that was mentioned is that ifs has inked deals worth over 4B. Obviously this is spread out across multiple quarters, and it probably isn’t revenue we will see until 2024; but IMO it is a real good start for a division that is less than 2 years old. The closest thing to compare to would be DCAI, which for this quarter did 4.3B in sales (and this is with like 75% of the world’s total server market). I am certainly excited to see the specifics of the new IFS deals later this year, because at this point I think it is clear that this is not another tenith-hearted attempt like intel custom foundry was.
I still remain highly pessimistic about IFS chances of success. Pat has positioned the IFS initiative as directly gunning for TSMC's process leadership and thus its clients. We all know how capital intensive the foundry space is. Does anything really find this realistic considering Intels cratering financials? Last year was horrific, last quarter apocalyptic and Q1 guidance is comparable to 2010 levels!!! How could they possibly hope to win a price war by undercutting TSMC when their margins are speeding towards the center of the earth at mach 2 while burning down current assets. Intel has had to resort to extraordinary measures by effectively mortgaging 50% of its future revenue on certain fabs through entering into partnerships with Brookfield. Pat indicated on their call they intend to do more such deals. How in the world is this going to lead to a sustainable margin profile? If they really want to gun for TSMC they will have to provide substantially better pricing, which I'm sure TSMC's management will respond to. TSMC has way more ammo in this hypothetical war and can simply turn the screws and outlast Intel. In short it sounds good in theory if you don't game it out much but I just cant find a scenario where Intel wins a meaningful war against TSMC. I know many will argue that they aren't trying to beat TSMC and are gunning for 2nd or 3rd place but that is not the way Pat and management have framed it. Why would Pat harp on "unparralled process leadership" offered at IFS if he wasn't gunning for first. What is the point of having the best process if you aren't trying to be first? It doesn't make any sense. The R&D resources of the eco system i.e Apple, AMD, NVDIA, MEDIATEK etc seem unassailable. Never mind TSMC's own financial position.
 
In short it sounds good in theory if you don't game it out much but I just cant find a scenario where Intel wins a meaningful war against TSMC. I know many will argue that they aren't trying to beat TSMC and are gunning for 2nd or 3rd place but that is not the way Pat and management have framed it. Why would Pat harp on "unparralled process leadership" offered at IFS if he wasn't gunning for first. What is the point of having the best process if you aren't trying to be first? It doesn't make any sense. The R&D resources of the eco system i.e Apple, AMD, NVDIA, MEDIATEK etc seem unassailable. Never mind TSMC's own financial position.
Because first is literally impossible and Intel surely knows it. TSMC is a pureplay foundry and the best one at it (service and process wise). Part of TSMC's recently growing margin has been their relatively recent process leadership. Just like how TSMC can't magically have the capacity if intel wanted to go fabless, intel can't magically have the capacity to supply 100% of the big 10 fabless firms in addition to their own demands. Also TSMC won't thrash it's margins for years to kill IFS. This isn't total war, it's business. TSMC's margin makes shareholders fat and happy and enables their colossal re-investment. To ruin it would piss off shareholders and potentially sabotage TSMC's future as an innovation powerhouse.

Margin wise let's say TSMC lowers their margin to a "measly" 45 or 50%. This still leaves plenty of room to undercut. If intel has PPW leadership, and your customer service is better than Samsung, then I see no reason why intel can't catch many customers with a 35% margin (some firms might even pay the same price for a small foundry service hit if it lets them have better products than their competitors). The made in the west, and best not TSMC alternative might even allow for higher than 35% margins. Obviously this is nowhere near intel's old 60+% margin but it is probably enough to subsidize their node development like the memory business does for Samsung foundry. IFS also serves as a nice way to monetize the business that intel and AMD's CPU bushiness will eventually lose to systems/software companies. This alternate revenue stream can also prop up the company if the design side goes through a rough patch, and further monetize tools/fabs that are too old to continue using.

At the end of the day design will still be the profit leader but there is plenty of room for foundry to be a big side business that can monetize intel's investments and add a fair amount of extra revenue that can make the company more sustainable.
 
Last edited:
Because first is literally impossible and Intel surely knows it. TSMC is a pureplay foundry and the best one at it (service and process wise). Part of TSMC's recently growing margin has been their relatively recent process leadership. Just like how TSMC can't magically have the capacity if intel wanted to go fabless, intel can't magically have the capacity to supply 100% of the big 10 fabless firms in addition to their own demands. Also TSMC won't thrash it's margins for years to kill IFS. This isn't total war, it's business. TSMC's margin makes shareholders fat and happy and enables their colossal re-investment. To ruin it would piss off shareholders and potentially sabotage TSMC's future as an innovation powerhouse.

Margin wise let's say TSMC lowers their margin to a "measly" 45 or 50%. This still leaves plenty of room to undercut. If intel has PPW leadership, and your customer service is better than Samsung, then I see no reason why intel can't catch many customers with a 35% margin (some firms might even pay the same price for a small foundry service hit if it lets them have better products than their competitors). The made in the west, and best not TSMC alternative might even allow for higher than 35% margins. Obviously this is nowhere near intel's old 60+% margin but it is probably enough to subsidize their node development like the memory business does for Samsung foundry. IFS also serves as a nice way to monetize the business that intel and AMD's CPU bushiness will eventually lose to systems/software companies. This alternate revenue stream can also prop up the company if the design side goes through a rough patch, and further monetize tools/fabs that are too old to continue using.

At the end of the day design will still be the profit leader but there is plenty of room for foundry to be a big side business that can monetize intel's investments and add a fair amount of extra revenue that can make the company more sustainable.
All reasonable points. TSMC shareholders should encourage the strangling of IFS in the cradle IMO. Sacrificing short term margins to kill IFS before it can become a threat makes sense to me.
 
Intel is in a very difficult position here. Building a fab without committed volumes is a very risky strategy, especially when existing fabs are underutilized. I personally don't believe there is a V-shaped recovery in semi demand around the corner. When the cloud providers are all saying they are going to extend the useful life of their systems from 4 to 5 years, that means there is going to be a sustained 25% reduction is demand from that sector. Without a slowdown in tech hiring, one would also expect a slowdown in PC sales, especially to corporations. Intel is seeing this massive demand drop and we aren't even in a recession yet. Imagine how bad it could get if there is a wider economic slowdown.
 
A lot of sceptical people here (including myself).
I think it'd be interesting to do a poll, "Who will have the process lead in 2025?"

The problem is that the process lead alone is not enough and sometimes it can be misleading or leading to attention be misplaced. There are many factors and capabilities need to be in place such as yield, IP, ecosystem, supply chain, timing, implementations, execution, cost, price, and capacity, etc.
 
I find myself thinking INTC is a buy at the current levels of pessimism and piling on Pat.

The bull case: Intel can do 18A in Hillsboro. They have the fab sites and some shells to build HVM for this node. 5 nodes in 4 years means in 2027 Intel will be in a process leadership position (4 years from 2023).

Bear case: Too little too late, 4 years is an eternity, and in the meantime cash flow will erode and competitors will respond. There is demand for 18A today, but firm demand? Will demand still be here in 2027? There is a monumental fab re-equipping task, and a monumental yield learning task, that needs to be complete in 2-3 years, to be in production in 4. Competitors are not standing still and don't need to do monumental, only average, to beat Intel to 18A.
 
Back
Top