Intel Provides Update on Internal Foundry Model

[tooLongInEDA]

Analyst note and he is not wrong, but remember, the foundry business is a marathon not a sprint:

Bernstein analyst Stacy Rasgon, who has a market perform rating and per-share price target of $30 on Intel, said that even though the economics of the foundry look "unsurprisingly horrible" right now, Intel fully recognizes it needs to develop a "structurally more efficient cost structure."

"Up until this point, however, we have been unclear where those [cost of goods sold] savings are coming from (or indeed, whether the company even knew)," Rasgon wrote in an investor note, referencing Intel's comments that it will pull $8B to $10B in costs out of the business, including roughly $3B this year.

"However, they did identify at a high level a number of efficiencies (expedites, sampling, test times, architecture etc) with more (utilizations, steppings, etc) on top that they believe can account for $4-5B and in general do not appear to be hugely revenue-dependent (important)."

Rasgon added this would probably allow Intel to improve gross margins by roughly 10 points - though not get the company to its 60% target - but seeing as there is a significant amount of costs to take out, identifying and admitting to the discrepancy between the company and its competitors is "probably a healthy exercise."

The numbers still doesn't add up for me as presented today.

Again, Slide 19 on the $4-5bn savings.

* Charge for samples the same way a foundry would : $0.5-1.0bn pa

If the vast majority of the IFS business is Intel internal, that's +$1.0bn for IFS and close to -$1.0bn for the Intel design groups and only a marginal saving for Intel overall.

* Product Architecture: Market-based pricing allows BUs to more easily identify feature ROI : > $1bn pa

Surely that's a saving made by the design groups and not by IFS ?

The fact that such a statement is possible implies that Intel has been operating some bizarre non market-based pricing/accounting system for decades. Cumulate this over 10 or 20 years and you're talking serious waste. That's a lot of shareholder funds that wnet missing. And these analysts have only just noticed ?

Whatever, given the scale of the savings and Intel's cost base, it's hard to believe this doesn't imply either serious headcount cuts or operating significantly higher revenues at the same headcount.

 

[count]

These "savings" look more like reallocations than anything. There is an assumption here that if the design group and foundry group each have their own P&L and have the freedom to operate independently, the accountability created will lead both groups to cut costs. In practice, that is unlikely to happen.

 

[nghanayem]

The numbers still doesn't add up for me as presented today.

Again, Slide 19 on the $4-5bn savings.

* Charge for samples the same way a foundry would : $0.5-1.0bn pa

If the vast majority of the IFS business is Intel internal, that's +$1.0bn for IFS and close to -$1.0bn for the Intel design groups and only a marginal saving for Intel overall.

All of the savings listed are for intel overall not IFS.

* Product Architecture: Market-based pricing allows BUs to more easily identify feature ROI : > $1bn pa

Surely that's a saving made by the design groups and not by IFS ?

The fact that such a statement is possible implies that Intel has been operating some bizarre non market-based pricing/accounting system for decades. Cumulate this over 10 or 20 years and you're talking serious waste. That's a lot of shareholder funds that wnet missing. And these analysts have only just noticed ?

They literally said how pricing was working. At the end of the quarter they then subtract chip revenue from fab opex/capex. Their whole point was there was no pricing system. Everything just got chucked into the COGS bucket.

 

[hist78]

There is only one stock symbol "INTC" for Intel. That means investors will pay most attention to the profit or loss number for INTC. It doesn't matter there are two, three, five, or eight business units (BU) under the Intel Corporation. At the end of the quarter or end of the year, the only thing matters is the overall Intel's performance. The so called BU accountability can be easily thrown away for the sake of INTC overall financial performance.

It's the left pocket vs right pocket on the same pant.

 

[hist78]

The transcript, video replay, and presentation material for Intel Internal Foundry Model Investor Webinar can be viewed here:

Internal Foundry Model Investor Webinar

www.intc.com

 

[blueone]

There is only one stock symbol "INTC" for Intel. That means investors will pay most attention to the profit or loss number for INTC. It doesn't matter there are two, three, five, or eight business units (BU) under the Intel Corporation. At the end of the quarter or end of the year, the only thing matters is the overall Intel's performance. The so called BU accountability can be easily thrown away for the sake of INTC overall financial performance.

It's the left pocket vs right pocket on the same pant.

Investors do pay attention to results on important business units that report separate P&Ls. For example, Ford split its EV business into a separate P&L, and investors are paying attention. I'm most interested in Intel's IFS business. If IFS doesn't succeed, IMO, Intel is the next IBM. I think a lot of investors feel that way.

 

[hist78]

The numbers still doesn't add up for me as presented today.

Again, Slide 19 on the $4-5bn savings.

* Charge for samples the same way a foundry would : $0.5-1.0bn pa

If the vast majority of the IFS business is Intel internal, that's +$1.0bn for IFS and close to -$1.0bn for the Intel design groups and only a marginal saving for Intel overall.

* Product Architecture: Market-based pricing allows BUs to more easily identify feature ROI : > $1bn pa

Surely that's a saving made by the design groups and not by IFS ?

The fact that such a statement is possible implies that Intel has been operating some bizarre non market-based pricing/accounting system for decades. Cumulate this over 10 or 20 years and you're talking serious waste. That's a lot of shareholder funds that wnet missing. And these analysts have only just noticed ?

Whatever, given the scale of the savings and Intel's cost base, it's hard to believe this doesn't imply either serious headcount cuts or operating significantly higher revenues at the same headcount.

It seems that Intel is changing its cost cutting projection. It was announced in February 2023 that Intel will cut cost of $3 billion for 2023, $8 billion to $10 billion yearly each for 2024 and 2025. The total 2023 to 2025 cost reduction will be $19 billion to $23 billion. But now Intel is talking about $3 billion for 2023 and $8 billion to $10 billion exiting 2025. The total 2023 to 2025 cost reduction projection has been reduced to $11 billion to $13 billion

On Feb 22, 2023, Intel said:

"Delivering $3 billion in cost savings in 2023, on the path to $8 billion to $10 billion in annualized savings by the end of 2025."

Source: https://www.intc.com/news-events/pr...updates-capital-allocation-to-drive-long-term

Then on June 21, 2023, Intel CFO David A. Zinsner said:

"We first discussed our internal foundry model in Q3 of 2022 as part of our multi-year cost efficiency effort, whichincludes reducing costs by $3 billion in 2023 and $8 billion to $10 billion exiting 2025."

Source: https://d1io3yog0oux5.cloudfront.ne...r+Meeting+-+IAO,+21-June-2023+11+30+AM+ET.pdf

 

[delong.height]

There is only one stock symbol "INTC" for Intel. That means investors will pay most attention to the profit or loss number for INTC. It doesn't matter there are two, three, five, or eight business units (BU) under the Intel Corporation. At the end of the quarter or end of the year, the only thing matters is the overall Intel's performance. The so called BU accountability can be easily thrown away for the sake of INTC overall financial performance.

It's the left pocket vs right pocket on the same pant.

are you going to say the same thing on Amazon?

Absolutely not, AWS and amazon.com are 2 very different BUs operating under AMZN umbrella. AWS brings in huge profit that can be put into amazon.com, its delivery network, and the rest of its services. While AWS is extremely profitable, it doesn't mean that AMZN needs to put all of its attention and resources in that single business unit and AMZN has never done that in the past.

Same case applies to Intel where there are only 3 leading edge foundries in the world, making its IFS business extremely valuable if executed well. While BUs do bring in bigger profit margins, given the importance of IFS in this world, both are equally important to Intel. When BUs bring in a bigger profit, that profit will be reinvested in IFS, and vice versa. It's a special advantage that neither Samsung nor TSMC get to enjoy.

 

[hist78]

are you going to say the same thing on Amazon?

Absolutely not, AWS and amazon.com are 2 very different BUs operating under AMZN umbrella. AWS brings in huge profit that can be put into amazon.com, its delivery network, and the rest of its services. While AWS is extremely profitable, it doesn't mean that AMZN needs to put all of its attention and resources in that single business unit and AMZN has never done that in the past.

Same case applies to Intel where there are only 3 leading edge foundries in the world, making its IFS business extremely valuable if executed well. While BUs do bring in bigger profit margins, given the importance of IFS in this world, both are equally important to Intel. When BUs bring in a bigger profit, that profit will be reinvested in IFS, and vice versa. It's a special advantage that neither Samsung nor TSMC get to enjoy.

"When BUs bring in a bigger profit, that profit will be reinvested in IFS, and vice versa. It's a special advantage that neither Samsung nor TSMC get to enjoy."

Do you think it's exactly what I said: "It's the left pocket vs right pocket on the same pant"? Under the same Intel Corporation or INTC stock symbol, they share the same interest to maximize Intel Corporation's profit, no mater how great or not so great of their own BU's P&L.

 

[tooLongInEDA]

It seems that Intel is changing its cost cutting projection. It was announced in February 2023 that Intel will cut cost of $3 billion for 2023, $8 billion to $10 billion yearly each for 2024 and 2025. The total 2023 to 2025 cost reduction will be $19 billion to $23 billion. But now Intel is talking about $3 billion for 2023 and $8 billion to $10 billion exiting 2025. The total 2023 to 2025 cost reduction projection has been reduced to $11 billion to $13 billion

On Feb 22, 2023, Intel said:

"Delivering $3 billion in cost savings in 2023, on the path to $8 billion to $10 billion in annualized savings by the end of 2025."

Source: https://www.intc.com/news-events/pr...updates-capital-allocation-to-drive-long-term

Then on June 21, 2023, Intel CFO David A. Zinsner said:

"We first discussed our internal foundry model in Q3 of 2022 as part of our multi-year cost efficiency effort, whichincludes reducing costs by $3 billion in 2023 and $8 billion to $10 billion exiting 2025."

Source: https://d1io3yog0oux5.cloudfront.net/_56832a5771104731299fde6f3355d13c/intel/db/857/8957/file/CORRECTED+TRANSCRIPT++Intel+Corp.(INTC-US),+Investor+Meeting+-+IAO,+21-June-2023+11+30+AM+ET.pdf

I always find it useful to normalise numbers as a sanity check.

So Intel's 2022 revenues were $63bn and 2021 was $79bn.

The whole purpose of this call was to prepare us for separate P&L accounting for IFS 2.0 and the "rest of Intel". It was never stated what the current revenue breakdown of IFS vs "rest of Intel" would be if the price-based accounting was used. But the example given was $100 revenue for IFS and $100 for the design BU. That seems feasible. That would give 2022 IFS revenues as $32bn.

Now $8bn (the lower saving quoted) is 25% of $32bn. Or if we spread this over the whole of Intel that 12.5%. Obviously that's a bit lower if sales recover.

On the other hand, Intel's been stripping out costs for some time already, so this isn't from a standing start.

In summary, the cost saving as a % of sales seems extremely ambitious and it's not totally clear it's achievable. I'm assuming all the new fab investment and any of the financial engineering stuff (Brookfield foundry deal, partial IPO share sales in subsidiaries) is outside all such calculations and this is just about ongoing operating costs.

 

[VCT]

" When BUs bring in a bigger profit, that profit will be reinvested in IFS"

The reality is no matter it's a big profit or a big loss, Intel have to reinvest in IFS.

 

[MKWVentures]

So either intel forged data on four academic papers across two industry conferences; or a random guy on the internet is incorrect on his analysis. I know which I would bet on being more likely.

Transistor performance for "3nm" class nodes in 2023 and early 2024.

IDEM and the VLSI symposium have given us a good amount of data to pour over for the node landscape from TSMC and intel in 2023/early 2024. Samsung having declined to give any details on the "in production" 3GAE that has "perfect yields" (their words not mine). I thought it might be nice if that...

semiwiki.com

that is a super funny comment. But in reality, Intel 3 will have its first internal product shipping in May-June 2024 if all goes well. Correct? and while Intel may not have forged data, Obviously data presented by Intel at a conference can be optimistic or cherry picked . Correct?

 

[nghanayem]

that is a super funny comment. But in reality, Intel 3 will have its first internal product shipping in May-June 2024 if all goes well. Correct? and while Intel may not have forged data, Obviously data presented by Intel at a conference can be optimistic or cherry picked . Correct?

I guess that depends on your definition. Maybe back when the data was collected the clock speeds they were were hitting at the listed voltages were on top bins rather than what the median die could achieve. I could see the argument that this is cherry picking results, and I could also see the argument that this is just showing the capabilities of the technology. As for being optimistic, I don't think that is physically possible as intel presented the data as hard silicon results (at least if I properly remember the contents of the whitepaper).

 

[Tanj]

Rejecting hot lots may be driving with the rear view mirror. The existing market where customers have no choice but 100 day cycle times is a real pain, but in a seller's market with a mono/oligopoly of fabs there was no choice. In the coming years with a flood of capacity coming on it would be smart to plan the fabs to be able to produce any product on a 60 day cycle that today only a hot lot gets. The Finance response of charging to remove the pain is not listening to the customer signal, and the next few years of overcapacity will require fabs to please the customers more than just by offering a comparable node.

 

[Tanj]

On Feb 22, 2023, Intel said:

"Delivering $3 billion in cost savings in 2023, on the path to $8 billion to $10 billion in annualized savings by the end of 2025."

Then on June 21, 2023, Intel CFO David A. Zinsner said:

"We first discussed our internal foundry model in Q3 of 2022 as part of our multi-year cost efficiency effort, whichincludes reducing costs by $3 billion in 2023 and $8 billion to $10 billion exiting 2025."

No difference?