Lip-Bu Tan Update on 18A and 14A

[tomatoma]

They may, indeed, make that choice. However, I would caution against setting expectations based on Intel's past behavior. Intel has had two CEO's that weren't steeped in the Intel culture.

The first was Bob Swan who, if rumors are to be believed, was looking to divest Intel of its fabs. Heresy to the old guard who bleed blue. He was also making significant changes to the culture that were designed to move Intel away from the twisted version of the Grovian culture that had grown up over time. Just my opinion, but I think Andy Grove was rolling over in his grave when he saw what the culture he had fostered had become. A culture where data was ignored and people were afraid to question decisions wasn't what Grove advocated for.

The second is Lip-Bu Tan. Like Swan, he is an outsider and less interested in doing things the Intel way. From all I hear, he is doing far more than giving lip service to the idea of listening to the customer. It will be a long journey (I've seen statements that he committed to 10 years in the CEO role), but sees a very different company at the end of the road than his predecessor. Gelsinger seemed to want to raise the old Intel from the ashes, but Lip-Bu Tan seems like he wants to fundamentally change the company. I believe his approach is far more pragmatic and the changes he is driving in Intel culture will result in a very different company from the Intel of the past.

Even though the methods of realization are different, I think it can be said that LBT agreed with Gelsinger's ideas to some extent when they didn't give up on the fab...

 

[lispwright]

So do you think Intel wafer cost is higher or lower than TSMC's, or at least be more competitive than TSMC?

I'm not sure it matters a lot for my thesis, as long as the difference is not "huge." See dkr1986's reply for that sort of analysis.

But I'd like to know why Intel chose N3B for Lunar and Arrow Lake. As far as I can tell, for whales who get press, Apple is the only other customer, using it for the M3 line and A17 but not following ones.

Isn't it an "orphan" node, with limited HVM experience? It's not PDK etc. compatible with the other N3 nodes, which are classed as being part of the N5 family. And from the claims that N2 is projected to be more used than the N3 families, I assume that's because TSMC mastered GAA with it, and it's more like what N3B should have been.

For Intel was that a valid decision at a point in time, was it "fair," vs. being picked because it was the highest price TSMC node to make Intel Foundry look better? For I'm generally betting on this of your two Intel failure theory options, things I've observed for decades:

Or do we believe Intel’s problems have nothing to do with the IDM business model? Instead, is the real issue that over the past 20 years Intel consistently hired poor engineers, ineffective managers, weak boards of directors, and bad CEOs who made bad decisions?

Or some of both, but with IDM weakness being dominated by bad management. Including IDM being much more vulnerable to that since you have to be good at two difficult things at once. And highest end logic fabs are the hardest.

However you assign the blame, clear weaknesses of Intel were illustrated by the 10 nm debacle. Would a design only Intel have stuck with a failing other company foundry node for years, vs. the history of e.g. all of Samsung's whale customers moving to TSMC?

And staying there except maybe some of Musk's business, which I interpret as a bet Samsung might just pull it off, if so valuable with TSMC being so booked? Which is also getting orders of similar size, "2 billion dollars" I read being thrown around for both.

 

[MKWVentures]

Exactly as you say I suppose. It will be under "Income attributable to non-controlling interest" and will be deducted before the "Net income attributable Intel investor" is calculated. Net income, Earnings per Share and Operating & Free Cash Flow metrics will be lower. Also I would not assume that since Brookfield owns 49% of the Fab & tools, they would be entitled to 49% of the profit. It will be much lower in my opinion as Intel operates the fab & owns the IP being produced in the fab. It is like leasing a co-owned property. I don't know if that is 20% or 30% but definitely cannot be 49% (if it is then the CFO is a fool).

We need to keep in mind Intel Product is already paying high wafer prices to TSMC (including 50% TSMC gross margin). So high wafer price for a leading-edge node is already accounted for in recent times for Intel Products team. It is a matter of ramping down Arrow lake+ Lunar Lake volume and increasing Panther Lake volume to offset. Considering 18A based products are much better than Arrow Lake & lunar lake (yes, all things indicate PTL is a better product than the ARL & LNL so far), ASP increase for 18A based products can be incremental. So, I think as 2026 progresses, Intel Product margins should stay flat. The threshold is Intel 18A wafers starting to eat into Intel 7 wafers volume (as opposed to N3), then product margin will start taking some hit. is the 18A based product ASP increase enough to offset 18A wafer ASP increase is a question, I don't have answer for yet. I have not accounted to demand destruction due to high memory prices.

On the foundry side, the 18A ramp cost would eat into margins but 18A wafer price would help margins. If yield is decent, the more they use foundry, the more the economics of foundry look good. If the cost of 18A wafer is higher than 18A wafer price due to low yield, then nothing changes and foundry bleeds money. But even these changes over time as yield improves. I only see margin improvement for Intel Foundry from here after some startup costs during initial ramp. The bigger problem is idle older node fabs that not producing anything or being underutilized.

Lets see what the intel report out shows in this quarter and next quarter. Not sure all of the impacts and what changes Intel makes. I am very interested on the Brookfield impact.

some thoughts:
1) 18A wafer cost will not be equal to TSMC N2 or N3. The Wafer price charged to Product group should be same as N2. So we should be able to see impact on product and IFS margins (this impact was clear on Intel 4, which lowered IFS margins).
2) 18A cannot be cost effective in 2026, Low volumes, outs per tool challenges, Low yield. Even at full ramp in Fab 52 by Q1 2027, that is still relatively low volume,
3) I am not sure how panther lake will ramp. Arrow lake and Meteor lake and lunar lake are incompatible designs with different ballouts. I am not sure how OEMs are redesigning. I think it is possible that Lunar and Arrow ramp throughout 2026. Hopefully they tell us but a good question is "do you expect to sell 40M Panther lake units in 2026???"... they should easily if everything is working well. And Nova lake is the desktop replacement end of year. and we have Arrow lake refresh coming out.
4) On the memory side: what is the impact of Intel paying for memory on Lunar lake vs ODM buying it on Panther lake. I dont know but it should be significant one way or the other given DDR5 prices doubled.

Hopefully we get some inputs from DZ and LBT

 

[hist78]

Or some of both, but with IDM weakness being dominated by bad management.

After two decades of setbacks and failures tied to Intel’s management missteps, it’s reasonable to ask whether these failures are truly the fault of those Intel executives or whether they are an inherent outcome of the IDM business model itself.

I believe that as long as the IDM model remains a fixed doctrine that cannot be meaningfully changed, Intel is likely to continue making strategic and operational mistakes. It’s not because its executives lack ability or good intentions, but because the underlying IDM model leads them or forces them into those disasters.

 

[hist78]

3) I am not sure how panther lake will ramp. Arrow lake and Meteor lake and lunar lake are incompatible designs with different ballouts. I am not sure how OEMs are redesigning. I think it is possible that Lunar and Arrow ramp throughout 2026. Hopefully they tell us but a good question is "do you expect to sell 40M Panther lake units in 2026???"... they should easily if everything is working well. And Nova lake is the desktop replacement end of year. and we have Arrow lake coming out.

My question is: if those PC sales are mostly replacements for older PCs, what will be the implications? Intel may say that AI PCs will increase the average selling price and consequently increase Intel’s revenue and profit. But if the overall PC market cannot grow significantly (as in the past 10 years shown below), there is very limited potential for Intel to increase its revenue and profit meaningfully. Even worse, there are several well positioned competitors sharing the same pie, which will put a lot of pressure on Intel not to raise prices too much.

 

[hist78]

4) On the memory side: what is the impact of Intel paying for memory on Lunar lake vs ODM buying it on Panther lake. I dont know but it should be significant one way or the other given DDR5 prices doubled.

This is a strange situation is that Intel claimed the bundled memory makes the Lunar Lake too expensive for OEM customers. I don't believe that as big as Intel can't get better memory price than many PC makers and offer a greater value (or savings) proposition to Intel OEM customers.

Could it be that Intel’s advanced packaging (Foveros 3D stacking), which Lunar Lake chiplets are based on, is not cost competitive and therefore eliminates the savings from the integrated memory on Lunar Lake products? Or something else?

 

[tomatoma]

After two decades of setbacks and failures tied to Intel’s management missteps, it’s reasonable to ask whether these failures are truly the fault of those Intel executives or whether they are an inherent outcome of the IDM business model itself.

I believe that as long as the IDM model remains a fixed doctrine that cannot be meaningfully changed, Intel is likely to continue making strategic and operational mistakes. It’s not because its executives lack ability or good intentions, but because the underlying IDM model leads them or forces them into those disasters.

So Intel has no value
IDM is a dirty, old-fashioned way of doing things.

 

[blueone]

This is a strange situation is that Intel claimed the bundled memory makes the Lunar Lake too expensive for OEM customers. I don't believe that as big as Intel can't get better memory price than many PC makers and offer a greater value (or savings) proposition to Intel OEM customers.

Price is not a factor. The OEMs want to sell memory themselves, to allow them to market varying memory sizes, giving them a wider pricing range than SoCs with integrated memory. Also, they want to avoid Intel's markup on DRAM, since the OEMs buy so much DRAM it is very likely they get a good enough price on DIMMs that Intel's cost advantage from integration is not that interesting to them.

Could it be that Intel’s advanced packaging (Foveros 3D stacking), which Lunar Lake chiplets are based on, is not cost competitive and therefore eliminates the savings from the integrated memory on Lunar Lake products? Or something else?

In-package DRAM chips decrease latency considerably, and reduces system power and board complexity (by a lot), but PC OEMs do not like Intel controlling their configurations. Integrated memory also makes laptops based on Lunar Lake have only physical differentiation from competitors, which tends to increase the production cost to differentiate.

 

[siliconbruh999]

Lets see what the intel report out shows in this quarter and next quarter. Not sure all of the impacts and what changes Intel makes. I am very interested on the Brookfield impact.

some thoughts:
1) 18A wafer cost will not be equal to TSMC N2 or N3. The Wafer price charged to Product group should be same as N2. So we should be able to see impact on product and IFS margins (this impact was clear on Intel 4, which lowered IFS margins).

Definitely there ain't no way but it can be good enough to make a a decent profit in their Books unlike Intel 7 which is more important imo.

2) 18A cannot be cost effective in 2026, Low volumes, outs per tool challenges, Low yield. Even at full ramp in Fab 52 by Q1 2027, that is still relatively low volume,
3) I am not sure how panther lake will ramp. Arrow lake and Meteor lake and lunar lake are incompatible designs with different ballouts. I am not sure how OEMs are redesigning. I think it is possible that Lunar and Arrow ramp throughout 2026. Hopefully they tell us but a good question is "do you expect to sell 40M Panther lake units in 2026???"... they should easily if everything is working well. And Nova lake is the desktop replacement end of year. and we have Arrow lake refresh coming out.

Well MTL/ARL are pin compatible and than PTL/NVL is pin to pin compatible LNL is different all together but Panther Lake has a selling point of the best iGPU in it's class that is going to drive sales

 

[MKWVentures]

@siliconbruh999 Said : Well MTL/ARL are pin compatible and than PTL/NVL is pin to pin compatible LNL is different all together but Panther Lake has a selling point of the best iGPU in it's class that is going to drive sales


This is a good point and test. Apparently, at least some skus, have the best performance. So we will see whether having the best performance is dominant effect. I have not gotten feedback yet on whether the OEMs are excited about highest performance processor. I tend to thing they will want low prices given the DRAM issues. AND DRAM could end up being the constraint so TAM could be limited. We will know in about 6 months

This is a strange situation is that Intel claimed the bundled memory makes the Lunar Lake too expensive for OEM customers. I don't believe that as big as Intel can't get better memory price than many PC makers and offer a greater value (or savings) proposition to Intel OEM customers.

I heard Intel say it makes gross margin lower for lunar lake. but in general, I would think EMS/ODMS/OEMs would be happy to not have to buy DRAM and support the MB cost of DRAM support.

This is another interesting market study that I believe will have some unexpectected business impacts ... The price swing of DRAM is Far more than the MB and foveros impact of DRAM. Simple test is whether Lunar lake sale drop and give way to Panther lake based on DRAM pricing and DRAM availability. Do PC OEMs release minimalist DRAM configurations and flexibility (which lunar lake doesnt support)??? I am running 3 windows Notebooks with 8, 16, and 32GB. most of the time, I cannot tell the difference (I even have AI apps running on Alder lake 8GB)

At the end of the day, All that matters from a business point of view is that Intel Revenue and Income starts to increase like every other compute company. Lets see what happens. Thursday resets the bar

 

[siliconbruh999]

This is a good point and test. Apparently, at least some skus, have the best performance. So we will see whether having the best performance is dominant effect. I have not gotten feedback yet on whether the OEMs are excited about highest performance processor. I tend to thing they will want low prices given the DRAM issues. AND DRAM could end up being the constraint so TAM could be limited. We will know in about 6 months

My name got changed with hist78 is it some admin power that i don't have

 

[hist78]

My name got changed with hist78 is it some admin power that i don't have

I was worried that my brain couldn’t remember what I said before.

 

[pepgpu]

High-NA EUV is garbage

Can you please expand on this comment?

 

[MKWVentures]

My name got changed with hist78 is it some admin power that i don't have

@hist78 Apparently I am not smart enough to know how to reply to two comments on same message..... my bad sorry for cut and paste issue!

 

[hist78]

I heard Intel say it makes gross margin lower for lunar lake. but in general, I would think EMS/ODMS/OEMs would be happy to not have to buy DRAM and support the MB cost of DRAM support.

Any advanced packaging technology will be an added cost. If Intel can use its buying power to integrate cheaper (compared to what OEMs pay) memory into Lunar Lake products and create more room for OEM customers such as HP, Dell, Lenovo, Asus, Acer, etc., to make more profit, those customers will buy Lunar Lake products. Why not! They are already in a market (in terms of specs, technology, protocols, standards, etc.) that is tightly controlled by Intel. If Intel’s Lunar Lake can bring more profit margin to them, they will buy. Especially considering their thin net profit margins (hovering around 2% to 5%), any help is welcome.

The problem is: can Intel do that? Further, are notebook PCs the right target for such an advanced packaging arrangement?

1. Intel’s manufacturing efficiency and cost structure may make Lunar Lake less attractive to consider. The OEMs’ slim net profit margins do not have room for it. Intel’s advanced packaging may be great in performance, but it may not be a good fit for such a low margin PC market.

2. Intel is an IDM. Intel needs to pay for and acquire the Lunar Lake memory in advance, then go through the lengthy packaging, testing, and sales/distribution/supply chain/accounts receivable process. With Intel’s current financial stress, it’s an expensive journey.

TSMC follows the foundry/fabless business model. Apple, Qualcomm, Nvidia, AMD, and Broadcom choose the right product to use a particular TSMC advanced packaging technology for the right target market. In that process, all the memory is acquired by the customers, and TSMC pays nothing. Those expensive memory components are owned by the customers and are not counted as TSMC’s inventory. The financial implications between TSMC and Intel are totally different.

 

[MKWVentures]

TSMC follows the foundry/fabless business model. Apple, Qualcomm, Nvidia, AMD, and Broadcom choose the right product to use a particular TSMC advanced packaging technology for the right target market. In that process, all the memory is acquired by the customers, and TSMC pays nothing. Those expensive memory components are owned by the customers and are not counted as TSMC’s inventory. The financial implications between TSMC and Intel are totally different.

Intel is not integrating DRAM on future chips and TSMC isnt part of this discussion IMO. the question is whether ODMs would like Lunar Lake with integrated DRAM or Panther Lake without it. I am not sure how this turns out.

Whether Intel can be profitable on 18A or panther lake remains to be seen in 2026 and 2027.

 

[siliconbruh999]

@MKWVentures Looks like we can get a glimpse of what is going to sell apparently the flagship Panther Lake is sold out the X7 variant both in 14/16inch variant.

Dell XPS 14-inch Premium Lightweight Laptop with Intel Core Ultra | Dell USA

Shop the XPS 14 ultra-thin & light all day performance laptop featuring Intel Core Ultra Processor & Intel Arc graphics.

www.dell.com

 

[tomatoma]

Can you please expand on this comment?

It's expensive

 

[blueone]

It's expensive

Yet sold out at the max configuration.

 

[siliconbruh999]

Yet sold out at the max configuration.

I think he meant high na is expensive not sold out for XPS