Intel reducing staff by 15% and dividend suspended

[Xebec]

I

I think Intel also said the market is tough LOL. the reply was "your competitors seem to be doing great" ...

To be fair, for Intel’s core business, the market is pretty tough right now.

x86 sales are slowing again, Apple Mac is resurgent, and Mobility is still growing. The GPU business is tough - even AMD is dropping prices constantly this year due to a combo of Nvidia strength and market apathy for products. (AMD is also seeing a drop in semi-custom business as console sales waver). Automotive business is flat or down which both AMD and Intel feed into a bit.

The server side continues to see competition from all directions. High interest rates reduce capital availability and risk appetites which makes it harder for a ‘new’ foundry to attract customers that have other proven options.

There’s a trade war (cold war?) with China that is also reducing Intel (and peer)’s opportunities for selling into that large market. The EU consumer is also softening a bit due to high energy prices. Several of Intel’s fabs sit not far from an active war zone (and maybe two active war zones) in Israel.

This would be tough even if Intel were executing well.

 

[MKWVentures]

To be fair, for Intel’s core business, the market is pretty tough right now.

x86 sales are slowing again, Apple Mac is resurgent, and Mobility is still growing. The GPU business is tough - even AMD is dropping prices constantly this year due to a combo of Nvidia strength and market apathy for products. (AMD is also seeing a drop in semi-custom business as console sales waver). Automotive business is flat or down which both AMD and Intel feed into a bit.

The server side continues to see competition from all directions. High interest rates reduce capital availability and risk appetites which makes it harder for a ‘new’ foundry to attract customers that have other proven options.

There’s a trade war (cold war?) with China that is also reducing Intel (and peer)’s opportunities for selling into that large market. The EU consumer is also softening a bit due to high energy prices. Several of Intel’s fabs sit not far from an active war zone (and maybe two active war zones) in Israel.

This would be tough even if Intel were executing well.

Surely you are joking. The market is not tough (it will get tough soon):

1) Datacenter CPU market is growing (slowly). There is tons of competition. All the competitor are doing fine.
2) Datacenter GPU market is growing wildly still. All Intel products here (years and years of work) are failing
3) PC CPU Market is recovering (Intel is doing OK here).
4) all competitors are growing.... DC Spending on Compute is through the roof (this will change in the future)
6) Intel CHOSE to focus on foundry and wafer fabs, something they are not good at.... Now the excuse is Intel 4/3 process/buildout LOL
7) Intel is losing A LOT of money, they are not claiming it by not including billions in the non-gaap. please read to the 10Q.
8) Actually Intel is Executing OK (other than Raptor Lake PR). If Lunar lake, Granite rapids come out, that will be pretty good for Intel.


Intel is not growing during a wild AI/DC build out. Winter is coming in 1-2 years. The capital spending and asset partnerships may put Intel at a risk they have never seen before.

I will have a blog that shows what Intel will do to get out of this mess both from engineering and accounting point of view. Hint it was in my scenarios published on Semi wiki. Intel has a VERY GOOD strategy and chance to be successful once they chose to focus on what they actually do well.

 

[blueone]

Surely you are joking. The market is not tough (it will get tough soon):

1) Datacenter CPU market is growing (slowly). There is tons of competition. All the competitor are doing fine.
2) Datacenter GPU market is growing wildly still. All Intel products here (years and years of work) are failing
3) PC CPU Market is recovering (Intel is doing OK here).
4) all competitors are growing.... DC Spending on Compute is through the roof (this will change in the future)
6) Intel CHOSE to focus on foundry and wafer fabs, something they are not good at.... Now the excuse is Intel 4/3 process/buildout LOL
7) Intel is losing A LOT of money, they are not claiming it by not including billions in the non-gaap. please read to the 10Q.
8) Actually Intel is Executing OK (other than Raptor Lake PR). If Lunar lake, Granite rapids come out, that will be pretty good for Intel.


Intel is not growing during a wild AI/DC build out. Winter is coming in 1-2 years. The capital spending and asset partnerships may put Intel at a risk they have never seen before.

I will have a blog that shows what Intel will do to get out of this mess both from engineering and accounting point of view. Hint it was in my scenarios published on Semi wiki. Intel has a VERY GOOD strategy and chance to be successful once they chose to focus on what they actually do well.

Good post, but I disagree with the last line. Intel's overall product strategy still seems uncompetitive. Apple and Qualcomm are out-innovating them in client CPUs (arguably AMD is too), and in server CPUs they're behind AMD for enterprise and, IMO, Ampere for cloud servers. Cost-wise both Intel and AMD look uncompetitive with the cloud company internal CPUs (Graviton, et al). I find it fascinating that Ampere's top line staff is entirely ex-Intel, which I think says a lot about Intel data center product engineering challenges.

I really hope the fabs and IFS execute very well, for the sake of my friends who still work there. I know it's not a popular opinion here, but I think the fabs are the future, so Intel had better figure them out.

 

[Xebec]

Surely you are joking. The market is not tough (it will get tough soon):

1) Datacenter CPU market is growing (slowly). There is tons of competition. All the competitor are doing fine.
2) Datacenter GPU market is growing wildly still. All Intel products here (years and years of work) are failing
3) PC CPU Market is recovering (Intel is doing OK here).
4) all competitors are growing.... DC Spending on Compute is through the roof (this will change in the future)
6) Intel CHOSE to focus on foundry and wafer fabs, something they are not good at.... Now the excuse is Intel 4/3 process/buildout LOL
7) Intel is losing A LOT of money, they are not claiming it by not including billions in the non-gaap. please read to the 10Q.
8) Actually Intel is Executing OK (other than Raptor Lake PR). If Lunar lake, Granite rapids come out, that will be pretty good for Intel.


Intel is not growing during a wild AI/DC build out. Winter is coming in 1-2 years. The capital spending and asset partnerships may put Intel at a risk they have never seen before.

I will have a blog that shows what Intel will do to get out of this mess both from engineering and accounting point of view. Hint it was in my scenarios published on Semi wiki. Intel has a VERY GOOD strategy and chance to be successful once they chose to focus on what they actually do well.

I am a lot more pessimistic on the overall market than most on this forum right now .

The PC CPU and GPU market is seeing seriously declining ASPs from all vendors for a while, with Nvidia recently announcing production cuts on GPUs to prop up pricing again. This is still (in 2024) the core of Intel’s business.. (and DRAM and NAND seem to be trending upward which will make increasing other component’s ASPs a challenge over the next 6-12 months).

I definitely agree there are some areas Intel should be doing better (if they had actually good execution). I just think some areas of the market that Intel is facing are tough already.

 

[KevinK]

GPU market is seeing seriously declining ASPs from all vendors for a while, with Nvidia recently announcing production cuts on GPUs to prop up pricing again.

I'm not buying this one... The kind of PC graphics GPUs you're talking about WRT to NVIDIA are small potatoes compared to the real GenAI GPU business. They wouldn't waste their time doing that for revenue enhancement - purely as part of the switchover to a higher performance replacement.

Nvidia's reportedly scaling back RTX 40 series production by as much as 50% in preparation for the Blackwell RTX 50 launch

Either that or the dearth of new GPUs means it's just a nice time to hike prices. Again.

www.pcgamer.com

 

[hist78]

He said:

"We are targeting a headcount reduction of greater than 15% by the end of 2025 with the majority of this action completed by the end of this year."

Why is everybody saying 15%?

I think Intel and Pat Gelsinger himself haven't settled on what exactly the number of employees who will be eliminated from the Intel payroll. I saw 15%, over 15%, 15,000 people, or 1,6000/17,000 people statements showing up in Intel internal announcement and Pat Gelsinger's interviews.

But they don't add up. Intel had 130,700 employees as of March 30th, 2024. 15% of 130,700 will be 19,605!

It's whopping 4,605 persons more than the 15,000 headcount cut Intel had said.

Note: On Intel own website, it said Intel has 131,000 employees worldwide.

 

[SaberG]

Maybe, Intel should fully operate like AMD and quit the idea of competing with TSMC in foundry field .

 

[Xebec]

I'm not buying this one... The kind of PC graphics GPUs you're talking about WRT to NVIDIA are small potatoes compared to the real GenAI GPU business. They wouldn't waste their time doing that for revenue enhancement - purely as part of the switchover to a higher performance replacement.

Nvidia's reportedly scaling back RTX 40 series production by as much as 50% in preparation for the Blackwell RTX 50 launch

Either that or the dearth of new GPUs means it's just a nice time to hike prices. Again.

www.pcgamer.com

I was referring to the consumer market, of course. I’ll be more careful with wording in the future here. (Do people normally lump in AI datacenter sales with “PC GPU” ?).

The PC graphics card market is still important for some companies - AMD sells to both console gamer markets and the PC. The GPU vendors are switching wafers to GenAI *and* still seeing a downturn in ASPs for PC graphics cards. Gaming GPUs still keep a decent chunk of fab capacity busy, etc.

Re: Nvidia gaming graphics pricing - They launched 4080 SUPER at 20% less than 4080, while being faster, and you’ve been able to get some of their cards below MSRP this year which is unusual.

I fully agree though they’re better off using wafers to sell AI hallucination .

 

[siliconbruh999]

Maybe, Intel should fully operate like AMD and quit the idea of competing with TSMC in foundry field .

Not gonna happen and too late they are way too deep in foundry and it is the right bet imo

 

[Xebec]

Not gonna happen and too late they are way too deep in foundry and it is the right bet imo

We also really don't want a monopoly on advanced processes.

 

[osnium]

We also really don't want a monopoly on advanced processes.

There isn’t a monopoly. Samsung is there as well.

 

[tooLongInEDA]

I think Intel and Pat Gelsinger himself haven't settled on what exactly the number of employees who will be eliminated from the Intel payroll. I saw 15%, over 15%, 15,000 people, or 1,6000/17,000 people statements showing up in Intel internal announcement and Pat Gelsinger's interviews.

But they don't add up. Intel had 130,700 employees as of March 30th, 2024. 15% of 130,700 will be 19,605!

It's whopping 4,605 persons more than the 15,000 headcount cut Intel had said.

Note: On Intel own website, it said Intel has 131,000 employees worldwide.

Indeed. You could be forgiven for looking at the headcount history over the past 5 years and coming to the conclusion that this company is completely out of control.

Pat Gelsinger takes over as CEO in February 2021. He know he has the biggest turnaround job in semiconductor history on his hands. A net 10K employees are added in 2021 and a further 10K in 2022. A net 7K then leave in 2023 (statista.com). There were no significant corporate acquisitions or disposals during this period.

If we were confident that the new hires were more productive than the people that left, perhaps this might make some sense. But it's quite a stretch to believe that. And rather easier to believe the reverse.

Add in the severance costs for Intel (and from Pat's statements this isn't going to be cheap).

What are they thinking ?

You've got to suspect that Intel also has a very serious problem with its sales forecasting over the past few years. Yes, it's tougher now they have to play in more competitive, fragmented and dynamic markets than they are used to. But they should be much better than this.

This is certainly not a tightly run ship.

My gut feeling -as a complete outsider - is that the problems at Intel run many levels deep within the management if these sort of things are happening.

 

[Maximus]

Intel slide from Gamers Nexus’ video
No more free fruits and beverages

 

[samwilde]

Intel is in deep shit. This is an accumulation of issues which cannot be blamed on Gelsinger.

Intel is behind in fabs, process technology, hardware architecture, did not head off ARM on mobile, then ARM on the cloud business, did not go into AI compute in time, etc. They spent money on stock buybacks which should have been spent on modern fabs. This is a similar deal of institutional rot to that experienced at Boeing.

I think Gelsinger is the ideal leader to try to get Intel through this rough patch. He is a technical person with long experience of working at Intel. But to solve all those issues will take at least 5 years. It is just how industry cycles work. The thing is, Gelsinger had a highly successful career as CEO at VMware. Economically for him it would have been better to stay there. Gelsinger only came back to Intel because it is basically where he was trained. He came to work at Intel during high school and left the company as a PhD. He designed the Intel 80486 CPU. So he treasures the company in a way you can bet other CEOs would not.

The reason Pat Gelsinger left Intel is that former CEO Craig Barrett pitted him against Paul Otellini to see who would be the next CEO after he retired. Gelsinger was put in charge of the Itanium division, which was floundering, and Otellini in charge of the laptop chip division. While Gelsinger was there the Itanium 2 CPU came out. But by then it was too late and few people wanted Itanium. Otellini was there when the Pentium M/Centrino came out. So Otellini was made CEO of Intel and Gelsinger left.

It takes two years or more to build a fab. Two years to develop a process. Two years to develop a processor architecture. You won't be designing that processor architecture without having first having a proven process.

The question is can Intel do it or not? I think the US government won't let Intel go down since it is the last chance the US has of remaining on the leading edge. If Intel dies then all the advanced chip manufacturing will be done in Asia.

 

[blueone]

View attachment 2157
Intel slide from Gamers Nexus’ video
No more free fruits and beverages

Some of these cuts are really silly.

Reducing sabbaticals? This won't really save anything appreciable. Not allowing vacation time to be appended to sabbaticals might have merit, but the basic duration cut is silly for a big part of Intel culture.

Eliminating the air shuttles again. When many development teams are distributed in the US, I think a small savings for a big overall cost in time, and reducing productivity due to less face time.

Okay, fitness coaching is a little much. Cancelling fruits and beverages... seriously?

I don't know what right-sizing dining programs mean, but it sounds like more employees will be going off-campus for lunch. That will probably extend lunch times.

 

[blueone]

Intel is in deep shit. This is an accumulation of issues which cannot be blamed on Gelsinger.

Intel is in deep shit, but Gelsinger is part of the problem in product design, not the solution.

Intel is behind in fabs, process technology, hardware architecture, did not head off ARM on mobile, then ARM on the cloud business, did not go into AI compute in time, etc. They spent money on stock buybacks which should have been spent on modern fabs. This is a similar deal of institutional rot to that experienced at Boeing.

I think Gelsinger is the ideal leader to try to get Intel through this rough patch. He is a technical person with long experience of working at Intel. But to solve all those issues will take at least 5 years. It is just how industry cycles work. The thing is, Gelsinger had a highly successful career as CEO at VMware. Economically for him it would have been better to stay there. Gelsinger only came back to Intel because it is basically where he was trained. He came to work at Intel during high school and left the company as a PhD. He designed the Intel 80486 CPU. So he treasures the company in a way you can bet other CEOs would not.

Pat does not have a traditional PhD. He has an honorary doctor of letters from Jessup University, which he did not attend. He does have a Stanford MSEE. He was chief architect of the 486 CPU, but that's not the same as "designing it".

The reason Pat Gelsinger left Intel is that former CEO Craig Barrett pitted him against Paul Otelini to see who would be the next CEO after he retired. Gelsinger was put in charge of the Itanium division, which was floundering, and Otellini in charge of the laptop chip division. While Gelsinger was there the Itanium 2 CPU came out. But by then it was too late and few people wanted Itanium. Otellini was there when the Pentium M/Centrino came out. So Otellini was made CEO of Intel and Gelsinger left.

This is not an accurate history.

It takes two years or more to build a fab. Two years to develop a process. Two years to develop a processor architecture. You won't be designing that processor architecture without having first having a proven process.

It takes more than two years to develop a new processor architecture. Of course, Pat still appears to be the x86 lover he always was, so I doubt he would start from scratch anyway.

The question is can Intel do it or not? I think the US government won't let Intel go down since it is the last chance the US has of remaining on the leading edge. If Intel dies then all the advanced chip manufacturing will be done in Asia.

It is not clear to me how the USG could save Intel if it really flounders.

 

[samwilde]

Intel is in deep shit, but Gelsinger is part of the problem in product design, not the solution.
...
It takes more than two years to develop a new processor architecture. Of course, Pat still appears to be the x86 lover he always was, so I doubt he would start from scratch anyway.

Gelsinger has been CEO at Intel since 2021. It is not enough time to fix the process gap with TSMC of like two generations and do a new core architecture. Which was my point.
All the products you are seeing come out by Intel today still had their design started under previous CEOs.

It is not clear to me how the USG could save Intel if it really flounders.

They can inject enough capital in Intel to make sure they can build their fabs. Which is half the battle.

Intel milked their existing ArFi fabs as much as they could. But now it is over. They hit a brick wall and massive investments are required.

 

[KevinK]

The PC graphics card market is still important for some companies -

Maybe important, but not that substantial compared to other revenue sources and growth. Gaming now represents less than 10% of NVIDIAs overall revenue and less than 1/8 of AMD’s. And it’s not really a growth business at either of those companies - more of an incremental revenue opportunity piggybacking off AI designs, the reverse of what it was 10 years ago.

 

[Xebec]

There isn’t a monopoly. Samsung is there as well.

Lol.

 

[osnium]

Gelsinger has been CEO at Intel since 2021. It is not enough time to fix the process gap with TSMC of like two generations and do a new core architecture. Which was my point.
All the products you are seeing come out by Intel today still had their design started under previous CEOs.


They can inject enough capital in Intel to make sure they can build their fabs. Which is half the battle.

Intel milked their existing ArFi fabs as much as they could. But now it is over. They hit a brick wall and massive investments are required.

At this stage, do you think the US government should cancel the grants given to Intel and the money to TSMC in exchange for expanding AZ even more? For example, the US govt allowed ASML to acquire SVG, the last remaining US based litho company in the early 2000s. Why not realize market reality and go all in on TSMC? If Intel shuts down the fabs, TSMC can hire the unemployed Intel engineers and technicians for 50-75 cents on the dollar so it would make US based manufacturing more cost effective as well.