Intel CEO Pat Gelsinger Retires

[starsmine]

Summary

Pat Gelsinger steps down from the CEO role as well as his position on the Board of Directors. There is no immediate permanent replacement

 

Quotes

Quote

Intel has named two senior leaders, David Zinsner and Michelle (MJ) Johnston Holthaus, as interim co-chief executive officers while the board of directors conducts a search for a new CEO. Zinsner is executive vice president and chief financial officer, and Holthaus has been appointed to the newly created position of CEO of Intel Products, a group that encompasses the company’s Client Computing Group (CCG), Data Center and AI Group (DCAI) and Network and Edge Group (NEX). Frank Yeary, independent chair of the board of Intel, will become interim executive chair during the period of transition. Intel Foundry leadership structure remains unchanged.

The board has formed a search committee and will work diligently and expeditiously to find a permanent successor to Gelsinger.

 

My thoughts

No, the fabs are not being spun off. They cant, and even if they could, they cant. 14, 10 and 7nm and 20A do not use Cadence or Synopsys
Pat came in to right the ship of the fab side, and its sad to see him leaving right before 18A launches so he wont get credit for that. 

In terms of products, I  think the only wrong choice made during his tenure that was the investment in large dGPUs
The AI miss is unfortunate

Leaving the board of directors feels odd, but without more information, one can only draw baseless speculation as to why. 
 

Sources

https://www.intc.com/news-events/press-releases/detail/1719/intel-announces-retirement-of-ceo-pat-gelsinger

[Haswellx86]

The messiah himself as left. He has done his work and paved a new way for Intel... hopefully that is.

 

7 minutes ago, starsmine said:

sad to see him leaving right before 18A launches so he wont get credit for that. 

Or should we say, he won't get criticized for it ?  We don't know, just wait for Panther Lake.

 

I don't know why he would leave at such a critical moment for the company, but I guess that has nothing to do with his own life. Not that Intel will plummet in a day, but he was the CEO.

 

7 minutes ago, starsmine said:

In terms of products, I  think the only wrong choice made during his tenure that was made was the investment in large dGPUs

They reportedly lost about 3.5 billion dollars by now. The only R&D part would be much lower. It's just that it wasn't that successful that they lost so much. We anyways needed better integrated graphics, and now they finally compete against AMD. I believe this is just an investment for a better future. But it was unfortunate for them to make this decision between its crisis, but back then I think they were doing good.

[Shimejii]

i mean its not hard to speculate why. He sucked as Intels CEO, he made bad decisions after bad decisions to the investors/board members, and they told him retire or you will be fired. 

 

Intels 13th and 14th gen debacle, the struggle to get their GPU division to be functional, all around it just made intel look like a disorganized and a mess of a company that cannot be trusted.

 

From some early reports he was debating whether to sell the FAB business off, and was told No. He kept pushing until the US GOV told them if you try to sell it off, you will not be receiving the aid from the CHIPS ACT. So the investors/board was probably not to happy about another bad look.

 

Overall this has been an unmitigated disaster for intel, they had hoped he could help return them to form, but instead it was more of the same and an even worse dive. Intel is not completely out of the game yet, but they are starting to get left behind in ALL aspects, from consumer CPUS, to Server CPUS, to Super computer contracts, to having to go to their biggest competitor TSMC to make their chips because they lack the ability to with high enough yields and performance. They just keep taking losses and have not had a single standout good thing happen under the leadership. Every release has been met with some sort of catastrophic issue.

[starsmine]

6 minutes ago, Shimejii said:

i mean its not hard to speculate why. He sucked as Intels CEO, he made bad decisions after bad decisions to the investors/board members, and they told him retire or you will be fired. 

 

Intels 13th and 14th gen debacle, the struggle to get their GPU division to be functional, all around it just made intel look like a disorganized and a mess of a company that cannot be trusted.

Saying he sucked as intel CEO is crazy given how these products are developed
Raptor lake was already the road map when he came in.  The debacle with it is such a detail detail that its hard to ever say its on anyone in the c suite

He came in and fixed the fab issue

[CWS-Starbuck]

At least now that Pat is retiring, Riley will get to spend more time with his self-proclaimed "Second Dad"

[StDragon]

Their dGPUs wasn't what sunk them. In fact, it's probably their only consumer product that isn't viewed negatively.

[Lurick]

2 hours ago, starsmine said:

Saying he sucked as intel CEO is crazy given how these products are developed
Raptor lake was already the road map when he came in.  The debacle with it is such a detail detail that its hard to ever say its on anyone in the c suite

He came in and fixed the fab issue

But CPUs only take 3 to 6 months to fab and design, didn't you know he was the sole person behind everything!!!!!

/s

 

Seriously, if people think he is in any way responsible for 13th and 14th gen having issues they need to go back to the drawing board. Now he could be blamed for the response and handling of the whole debacle though, that I won't disagree on.

[starsmine]

34 minutes ago, StDragon said:

Their dGPUs wasn't what sunk them. In fact, it's probably their only consumer product that isn't viewed negatively.

And they aren't sunk

but they are not even using it as an off shoot of their AI branch so the R&D cant even be ammoratized that way. Gaudi is based off an entirely different acquisition from 2019 (Habana), hence why gaudi 1 was on TSMC and they have not transferred to intel yet. AMD makes money in dGPUs by doing semi custom work based off the IP, not the dGPUs themselves (and activly unifying the arch back to improve their AI ecosystems). Nvidia makes money on dGPUs because enterprise and AI, not the dGPU itself. I still dont know what the plan is for intel's dGPUs, that will be a money pit of no return for more then a decade.

[Lurick]

https://www.cnbc.com/2024/12/02/intel-ceo-pat-gelsinger-is-out.html

It would appear he "stepped down" after a heated disagreement with the board, aka was likely to be ousted anyway.

[leadeater]

4 hours ago, Shimejii said:

From some early reports he was debating whether to sell the FAB business off, and was told No. He kept pushing until the US GOV told them if you try to sell it off, you will not be receiving the aid from the CHIPS ACT. So the investors/board was probably not to happy about another bad look.

No he was not, there was literally zero evidence has was actually proposing or even thinking of proposing that. This was 1000% investor speculation to put pressure on the company to do something in their interest to make a quick buck. Selling the Intel fabs goes against everything Pat ever stood for in his entire career at Intel.

 

Never mistake or attribute a news article about something as the thoughts or the intentions of a company or a person by existence of the article alone.

[starsmine]

If there was to be a conspiracy around the selling of fabs, it would be that Pat asked for that to be a clause to be inserted the CHIPS act money so he can tell investors that his hands are tied to not sell the fabs and to fuck off. Not that the government told Pat not to.

[cj09beira]

8 hours ago, Lurick said:

But CPUs only take 3 to 6 months to fab and design, didn't you know he was the sole person behind everything!!!!!

/s

 

Seriously, if people think he is in any way responsible for 13th and 14th gen having issues they need to go back to the drawing board. Now he could be blamed for the response and handling of the whole debacle though, that I won't disagree on.

he joined in early 2021, early enough that these latest launches are all on him.

[starsmine]

3 minutes ago, cj09beira said:

he joined in early 2021, early enough that these latest launches are all on him.

Whats wrong with the latest launches?
Meteor/Lunar/Arrow are all solid.

[tagKnife]

9 hours ago, Lurick said:

https://www.cnbc.com/2024/12/02/intel-ceo-pat-gelsinger-is-out.html

It would appear he "stepped down" after a heated disagreement with the board, aka was likely to be ousted anyway.

This ^

 

Intels board already having 2 people lined up for CEO (David Zinsner and Michelle Johnston Holthaus named interim Co-CEO). They were going to fire Pat regardless, they would have given him an ultimatum, step down and get paid, or be forcefully removed and forfite his shares.

Also a CEO being forcefully removed doesnt look good on their resume if they want to go for another job.

 

Its a shame, I liked Pats plan, but the execution was a desaster.

[Lebon14]

4 hours ago, starsmine said:

Whats wrong with the latest launches?
Meteor/Lunar/Arrow are all solid.

BRUH.

 

Meteor and Lunar Lake were all affected by a degradation bug that Intel tried to swipe away under the carpet and throw the blame onto the motherboard vendors. That killed many otherwise good CPUs.

 

Arrow Lake is just, mostly, really really underperforming and too expensive. You know... when your new CPU gets beaten by your own CPU from 2+ years ago... Oh and, not like it's Intel's fault but on Win11 24H2, performance's even worse than on 23H2.

 

Last good Intel gen was 12th gen. That was actual solid overall and it's still a good buy imo.

 

Not saying AMD's AM5 launch was any better with all the false/misleading performance claims, also claiming 8000MHz+ RAM working on all motherboards while most of them fail to even post, etc.........

[starsmine]

1 hour ago, Lebon14 said:

BRUH.

 

Meteor and Lunar Lake were all affected by a degradation bug that Intel tried to swipe away under the carpet and throw the blame onto the motherboard vendors. That killed many otherwise good CPUs.

Are you confused, or am I? What reports have there been for meteor or lunar? They are not Raptor Lake 

1 hour ago, Lebon14 said:

Arrow Lake is just, mostly, really really underperforming and too expensive. You know... when your new CPU gets beaten by your own CPU from 2+ years ago... Oh and, not like it's Intel's fault but on Win11 24H2, performance's even worse than on 23H2.

For gaming... on a brand new foundational architecture. 

1 hour ago, Lebon14 said:

Last good Intel gen was 12th gen. That was actual solid overall and it's still a good buy imo.

 

Not saying AMD's AM5 launch was any better with all the false/misleading performance claims, also claiming 8000MHz+ RAM working on all motherboards while most of them fail to even post, etc.........

No one ever claimed 8000MT/s would work on AM5 AT launch. 

[Stahlmann]

Intel hasn't done well since he became CEO. Whether it's really his fault is debatable. But as CEO, he's still responsible. He doesn't design new CPUs himself, but he's part of the team that decides how and when they come out. Compared to the perception of the brand 5 years ago, the Intel of today is a joke.

[Salted Spinach]

For 2025, AMD will be releasing new GPU lineup aimed at gaining market share with good price/perf in the mid range. This is same space intended for Battlemage.

 

On the mobile front, Strix Halo is probably going to blot out the sun for Intel's APUs

[porina]

23 minutes ago, Salted Spinach said:

On the mobile front, Strix Halo is probably going to blot out the sun for Intel's APUs

Different market areas. Intel mobile CPUs will be going against Ryzen AI 300, when both are used without dGPUs. Strix halo will be going against dGPUs, which I guess will be mainly NV.

[igormp]

11 hours ago, starsmine said:

Whats wrong with the latest launches?
Meteor/Lunar/Arrow are all solid.

MTL was meh, LNL was pretty good, still a little bit behind in efficiency but nonetheless good. However, only came out with lower end parts and it's a one-off trick since it won't have any successor, which is really sad since a slightly upgraded LNL CPU (maybe an extra 4 or 8 E-cores with 192 or even 256-bit memory bus) would've been amazing and an interesting Strix Point competitor.

ARL was mid. Gave up on their major current foothold (games), while not beating AMD in productivity (but getting really darn close). If priced just a bit cheaper, it'd have been amazing, but being a CPU for a socket that will only last for this gen only makes its value even worse.

 

Let's see how Nova and Panther will pan out.

1 hour ago, Salted Spinach said:

On the mobile front, Strix Halo is probably going to blot out the sun for Intel's APUs

Complementing what was said above, Strix Halo is a 100W+ MCM chip, Intel doesn't have any direct competition in that front.

[Haswellx86]

Yeah, he was making very aggressive financial decisions. But what else was he supposed to do? Ditch the foundries and use TSMC? Well the government wouldn't let that happen. So even if the government wanted Intel to do their thing and rise up, then why did this Patrick guy get fired? Yes, fired. I thought he was resigning by himself, which feels so weird and odd in such a crucial timing, but in reality, he was forced to resign which makes much more sense... or rather not.

 

Yeah I get investors sometimes don't get anything and they just need the company to perform. And how Intel has acted recently, it has been a mess. But didn't Pat clarify that the whole point of Intel failing is because of their fabs, and that he has brought tons of innovation in them, and heavily betting the company on 18A? Do the investors now not believe or trust Intel anymore? Is 18A already being seen as a flaw?

 

Tbh, I still don't exactly get where Intel has been losing all this money. That Raptor Lake crashing issue doesn't put a dent. Intel is still very dominant in the OEM space. Sure, the techies could be buying Ryzen, but what percentage are they of the whole population? Plus we people still do buy Intel. And for Arc, well that could be some fraction of their losses. It is reported that Intel lost $3.5 billion in Arc. But that's the money lost, not aimed to invest initially (right, or am I wrong?). Pat probably didn't think that it could turn out this expensive, but still it is not the major factor where Intel is losing their money. And about their lose in the AI race, I have no idea why people compare it to Nvidia. Nvidia makes high performance GPUs, who themselves drove the AI innovation, whereas Intel is only dominant in CPUs, which are not the most efficient means to drive AI. So why not compare Intel, with say, AMD? From my knowledge, I think the only thing AMD has in AI is their AI accelerators like the MI300X, and maybe their workstation GPUs. So Intel also also has their Gaudi AI accelerator. I don't know how competitive that is, but that seems like even AMD doesn't have much role in AI yet, so why is Intel getting so much backlash? AMD is only ahead in portfolio by their GPU division, where Intel has just put a step in.

 

So how do they lose $16.6 Billion in just a quarter? It's all because of their fabs? Well before even Pat start investing heavily in their fabs (Intel 7 was already out, right?), it wasn't even that Intel is on an ancient technology, no? Yes, they didn't have the latest EUV tech, but even with their older tech they could make process nodes, well not efficient, but performant, and not horribly inefficient. I don't see how Intel 7 affected them. OEM sells would never be affected, techies, again in small number, might choose Ryzen because of their efficiency, and I guess in the enterprise space, that's where it would affect the most? I have no idea about the enterprise space and the profits of Intel in there.

So all the money lost, has only been in innovating the fabs so that they catch up in process node? This much money? Well of course yeah, it's expensive, but are fabs the only reason of Intel's huge losses? If say, Intel never invested so much in their process node, would their foundry have been doomed?

 

Pat honestly was such a good CEO for Intel. No I mean, besides all the innovation he has brought, he is an engineer, and was one of the engineers who originally designed the 486. What could be better than having a bright engineer who is also capable of being a CEO? And so now what, they going to get some one from a marketing background (the new interms are already from that)? We have seen many times in the past and of different companies that tech companies benefit with CEOs from engineering background, and not business.

[Monkey Dust]

It's part of the CEO deal, earn the epically big bucks, but have to take the fall (usually with a huge payout) when the shit hits the fan. Even though his short tenure means many of the missteps weren't his.

 

Intel seem to be in a cost-cutting spiral of doom, it's like they have been bought by private equity already. Though, I assume they are in the same 'strategically important' category as Boeing, and wouldn't be allowed to actually fail.

[Zodiark1593]

2 hours ago, Monkey Dust said:

It's part of the CEO deal, earn the epically big bucks, but have to take the fall (usually with a huge payout) when the shit hits the fan. Even though his short tenure means many of the missteps weren't his.

 

Intel seem to be in a cost-cutting spiral of doom, it's like they have been bought by private equity already. Though, I assume they are in the same 'strategically important' category as Boeing, and wouldn't be allowed to actually fail.

Considering that there aren't a lot of alternatives for OEMs to move to (AMD isn't especially good at getting their silicon to OEM hands, nor seem interested in the low-end, and ARM is not ready for prime-time in PC-space), I suspect Intel will probably be fine for the immediate future.

[SpaceGhostC2C]

15 hours ago, Haswellx86 said:

And about their lose in the AI race, I have no idea why people compare it to Nvidia. Nvidia makes high performance GPUs, who themselves drove the AI innovation, whereas Intel is only dominant in CPUs, which are not the most efficient means to drive AI. So why not compare Intel, with say, AMD?

Because you only win a race by overtaking the leader. If you manage to wrestle the 9th position from a rival in the final lap, you've still lost the race. There's no consolation price for failing in good company.

Nvidia jump-started everyone else in this race, long before his appointment. But the board surely expected to see some plan for a catchup. This isn't some random market, it's the buzzword market of the moment, so investors may (rightly or wrongly) perceive it as crucial for the future of the company. And (should they be right) a company like Intel doesn't survive on making small dents in something as crucial. It's also true that results take time to materialize, but the board will have an opinion on whether the company was steered into the right path on that front. Was Intel seeding its Zen-like moment for AI? Or was it Radeoning it around? 

 

15 hours ago, Haswellx86 said:

OEM sells would never be affected, techies, again in small number, might choose Ryzen because of their efficiency, and I guess in the enterprise space, that's where it would affect the most? I have no idea about the enterprise space and the profits of Intel in there.

Ryzen doesn't matter, it's really Epyc what can threaten profits (as opposed to revenue, although also revenue )

 

15 hours ago, Haswellx86 said:

Pat honestly was such a good CEO for Intel. No I mean, besides all the innovation he has brought, he is an engineer, and was one of the engineers who originally designed the 486. What could be better than having a bright engineer who is also capable of being a CEO?

Well, the board must have agreed with the last part when he was appointed; not sure they agree now, though.

Sometimes bright engineers are most useful at engineering roles. Although I know many among the more engineering-inclined folks in this forum tend to underestimate the bean-counting, people-handling content of a CEO role.

 

15 hours ago, Haswellx86 said:

We have seen many times in the past and of different companies that tech companies benefit with CEOs from engineering background, and not business.

I don't know. Intel's investors may have found their profits inversely proportional to Linus' enthusiasm for the company in the last couple of leadership changes.

[starsmine]

1 hour ago, SpaceGhostC2C said:

Because you only win a race by overtaking the leader. If you manage to wrestle the 9th position from a rival in the final lap, you've still lost the race. There's no consolation price for failing in good company.

Nvidia jump-started everyone else in this race, long before his appointment. But the board surely expected to see some plan for a catchup. This isn't some random market, it's the buzzword market of the moment, so investors may (rightly or wrongly) perceive it as crucial for the future of the company. And (should they be right) a company like Intel doesn't survive on making small dents in something as crucial. It's also true that results take time to materialize, but the board will have an opinion on whether the company was steered into the right path on that front. Was Intel seeding its Zen-like moment for AI? Or was it Radeoning it around? 

To give a date, Intel bought Habana labs, aka, Gaudi, in 2019. as in BEFORE Pat came back to intel as CEO. Gaudi 3 concept of a spec was being written at that time. and intel fabs were... well the reason why Pat was hired.

1 hour ago, SpaceGhostC2C said:

Ryzen doesn't matter, it's really Epyc what can threaten profits (as opposed to revenue, although also revenue )

 

Well, the board must have agreed with the last part when he was appointed; not sure they agree now, though.

Sometimes bright engineers are most useful at engineering roles. Although I know many among the more engineering-inclined folks in this forum tend to underestimate the bean-counting, people-handling content of a CEO role.

I honestly dont know what the board expected. From my understanding there are a lot of engineers with burnt bridges with intel from that time. he was hired in 2021, put in the plan for 5N4Y, that plan is now ending, with 18A, a node that intel can finally sell because its not using intel only tools, and is a market leader so people want in. I know people say wait for panther lake, Panther lake is a product, it does not tell you about the node. 

I know explicitly of a few companies that WANTED to buy into intel 4/3 (confidential chips, as in why CHIPS act is important), but it being the FIRST node intel could sell (ish), the tools were literally immature vs what they were used to with gloflo, so they bailed hard.
18A is the only one synopsys and cadence are really supporting.