Gelsinger “retires”

[hist78]

Alternatively, they can accelerate a move out of Oregon to Ohio and Arizona to take advantage of wealthier state governments and better quality workforce.

It might be a good idea to evaluate for the long term. But for the next several years Intel doesn't have the resources and time to get distracted. A major relocation right now will only add more complexity to the troubles Intel is facing.

 

[Maximus]

Intel faces a complex web of objectives, including national security demands from the US government, goals for its foundry, and design targets. It's impossible for Intel to achieve all these aims amidst such strong headwinds.

If the US truly wants to maintain advanced logic manufacturing domestically, perhaps declaring these operations are all tax-free?

 

[Maximus]

Have they considering fired the Board of Directors after firing 3 CEOs?

In Morris Chang's recently published autobiography, he recounts inviting Craig Barrett to join TSMC's board in 2008. This story highlights how much Morris Chang values experience when selecting board members.

English translation:
In 2009, TSMC planned to add an independent director. At the end of 2008, Morris Chang met Paul Otellini, then president of Intel, and asked if he could recommend someone.

Chang explained that at the time, TSMC and Intel saw each other as friendly peers, and Intel was still a TSMC customer. Otellini recommended Intel Chairman Craig Barrett, who was retiring at the time.

When Chang contacted Barrett, Barrett expressed interest but mentioned that everything would have to wait until he officially retired. A few days later, Chang received an email from Barrett stating that his entire career had been at Intel. He believed that one day, Intel and TSMC would compete with each other, so he decided to decline Chang's invitation. This was the only time Chang was declined in his many years of appointing independent directors to TSMC's board.

 

[hist78]

Intel faces a complex web of objectives, including national security demands from the US government, goals for its foundry, and design targets. It's impossible for Intel to achieve all these aims amidst such strong headwinds.

If the US truly wants to maintain advanced logic manufacturing domestically, perhaps declaring these operations are all tax-free?

If people don't want to buy the pizza I made, even with zero tax imposed on my pizza restaurant, I don't think I can keep my restaurant doors open for too long.

 

[Xebec]

"Co CEOs" may be a hint that the board wants to split Intel.

Curious if any insiders know if they're literally assigned the same duties as each other or if one is to focus on Foundry while the other focuses on Products.

 

[Jure]

Rumor: Intel contacted former TSMC Chairman Mark Liu

 

[tooLongInEDA]

"Co CEOs" may be a hint that the board wants to split Intel.

Curious if any insiders know if they're literally assigned the same duties as each other or if one is to focus on Foundry while the other focuses on Products.

You are assuming that there is a unified board view. Who knows ? The only thing we can say is that they were sufficiently unified to demand that Pat left. Spotting that you're heading the wrong way (or alternatively was it losing their bottle with IDM 2.0 ?) is no guarantee that they know and are aligned on a better direction.

I think it's more likely that they reckoned the job is too large for any of the execs they had on the bench.

Someone questioned why companies fire CEOs without any definite replacement plan (as appears to be the case here). Speaking from English football (soccer) experience, managers get fired all the time without replacements already lined up. The average tenure for an English Premier League Manager is 2 years and 4 days (and it's trending shorter). Things just get to the stage where the fans (or shareholders) demand "something must be done". Where something = anything.

 

[Xebec]

You are assuming that there is a unified board view. Who knows ? The only thing we can say is that they were sufficiently unified to demand that Pat left. Spotting that you're heading the wrong way (or alternatively was it losing their bottle with IDM 2.0 ?) is no guarantee that they know and are aligned on a better direction.

I think it's more likely that they reckoned the job is too large for any of the execs they had on the bench.

Someone questioned why companies fire CEOs without any definite replacement plan (as appears to be the case here). Speaking from English football (soccer) experience, managers get fired all the time without replacements already lined up. The average tenure for an English Premier League Manager is 2 years and 4 days (and it's trending shorter). Things just get to the stage where the fans (or shareholders) demand "something must be done". Where something = anything.

Fair and great points - the CEO is definitely a disposable position for situations like these, and I shouldn't assume a unified board. The recent reuters hit pieces may be evidence of this.

.. Separate, but related

If the reason was primarily financials, It should have been pretty obvious to people that regaining fab leadership was going to be expensive and have lagging year effects. (18A fab "operational in 2025" = "profitable no earlier than 2027").

I do think Pat legit earned the blame for how the company mishandled Raptor Lake's stabiity issues (felt a lot like how the company initially handled Pentium FDIV), and the Arrow Lake launch is rough. I do think he deserves some credit for geting the server pipeline back to at least mostly competitive, and doing enough to prevent AMD from taking over mobile despite having access to better manufacturing processes for a few years. Intel Arc Battlemage may also rest on Pat's sholders at this point too - good or bad.

Lastly for AI, it seems like Pat took the outdated approach of "Microsoft will drive AI, we'll provide the hardware" where Nvidia just took the ball and ran to the next stadium with it. (I feel like AMD is also similarly lagging in this space - they're just doing a bit better because their server platforms are the hotness that connects to Nvidia accelerators). Yeah, I guess I'm coming around to Pat should have done a lot better in his role..

 

[OneEng]

Wow. Didn't expect this until mid to late 2025 .... and not for the reasons many here might think.

For years, people have been talking about the downfall of Moore's Law. Everyone has assumed that the cause of this downfall would be physics or some other technical problem when in fact Moore's Law is dead because of financial issues.

Pat's 5NI4Y strategy assumed that Moore's Law could indeed be maintained. From an engineer's point of view, this may well have been true. From the CFO's point of view, it is simply not possible for exponentially escalating costs of node shrinks to be offset by sales of any kind.

The reason I expected Pat to stick around until 2025 was that I assumed that my conclusion (which is simple spread sheet math at the end of the day) would not be recognized until 18A successfully (from a technical point of view) launched. It would then become evident that even WITH the successful implementation of 5NI4Y, the company was going to be eternally awash in the red from a financial standpoint.

I believe that the industry in general will be effected by this new reality. For the past .... well forever (since IC's became a thing), the semi industry has primarily relied on having 50% more transistor budget for each new design. Such huge improvements in process drove the huge improvements in performance.

In addition to the massive change in fortunes for Moore's Law, the grip of x86 on it's markets is coming to an end. There is certainly nothing inherent in x86 designs that make them superior to RISC designs and in fact, x86 chips have been mostly RISC for quite some time in order to have ILP (with equal length data and instructions).

Companies that do not understand the financial ramifications of these new truths will most certainly fail. Companies that do understand and make strategies around these new truths will thrive.

It will be interesting to see how Intel pivots now.

 

[kevin01]

I was thinking why now. The only logical conclusions for me was that 1) the board had decided to remove Patrick G in Aug or Sept, but waited till the CHIPS deal got finalized, 2) the board was not happy with the CHIPS deal, because it required INTC to own at least 51% of the fab. And the board felt that PG was not candid to them about this "new" requirement. This became the last straw.

If it was 1), we may see a new CEO emerging relatively soon; but if it was 2), the ceo search could be a lengthy process.

It was sad to see PG retire. And wish INTC do well in the next few years, as PG did seem to have laid a good technical foundation for his successor.

 

[delong.height]

I do feel whoever comes next have a less of a struggle compared to Pat. And maybe it's good to have two co-CEOs managing largely unrelated business that serve each one's own interest.

What Intel encounter in late-2020, the server roadmap is completely uncompetitive, the process node is in a severe disadvantage, the products doesn't perform well, and if they are going to spend more in manufacturing or outsourcing them all to TSMC, it's all going to be touch on future profit generations. Decisions on these problems need a strong opinion leader like Pat to organize and pave the way.

Now I assume most of the works have been done. It's up to the successors to handle the rest. They will have a good process that's coming onboard assuming they are sticking to Pat's plan strategically.
They now have a competitive server lineup and roadmap, they have their feet in GPU and AI. They at least gained experience working with world-class foundry team, and learning how they should use that tactics to deal with the internal IFS.

I do think that the struggle they have now is less of a struggle, it's more on execution. And splitting the power to 2 may be good for the long run.

 

[count]

Lastly for AI, it seems like Pat took the outdated approach of "Microsoft will drive AI, we'll provide the hardware" where Nvidia just took the ball and ran to the next stadium with it. (I feel like AMD is also similarly lagging in this space - they're just doing a bit better because their server platforms are the hotness that connects to Nvidia accelerators). Yeah, I guess I'm coming around to Pat should have done a lot better in his role..

This is a pretty interesting take, and probably accurate. In logic, this is certainly what happened, and they probably expected something similar for AI.

 

[count]

I was thinking why now. The only logical conclusions for me was that 1) the board had decided to remove Patrick G in Aug or Sept, but waited till the CHIPS deal got finalized, 2) the board was not happy with the CHIPS deal, because it required INTC to own at least 51% of the fab. And the board felt that PG was not candid to them about this "new" requirement. This became the last straw.

If it was 1), we may see a new CEO emerging relatively soon; but if it was 2), the ceo search could be a lengthy process.

It was sad to see PG retire. And wish INTC do well in the next few years, as PG did seem to have laid a good technical foundation for his successor.

This is sort of my thinking as well. The "not moving fast enough in AI" explanation makes sense at the surface level, but does not explain the timing.

 

[bilau]

Wow. Didn't expect this until mid to late 2025 .... and not for the reasons many here might think.

For years, people have been talking about the downfall of Moore's Law. Everyone has assumed that the cause of this downfall would be physics or some other technical problem when in fact Moore's Law is dead because of financial issues.

Pat's 5NI4Y strategy assumed that Moore's Law could indeed be maintained. From an engineer's point of view, this may well have been true. From the CFO's point of view, it is simply not possible for exponentially escalating costs of node shrinks to be offset by sales of any kind.

The reason I expected Pat to stick around until 2025 was that I assumed that my conclusion (which is simple spread sheet math at the end of the day) would not be recognized until 18A successfully (from a technical point of view) launched. It would then become evident that even WITH the successful implementation of 5NI4Y, the company was going to be eternally awash in the red from a financial standpoint.

I believe that the industry in general will be effected by this new reality. For the past .... well forever (since IC's became a thing), the semi industry has primarily relied on having 50% more transistor budget for each new design. Such huge improvements in process drove the huge improvements in performance.

In addition to the massive change in fortunes for Moore's Law, the grip of x86 on it's markets is coming to an end. There is certainly nothing inherent in x86 designs that make them superior to RISC designs and in fact, x86 chips have been mostly RISC for quite some time in order to have ILP (with equal length data and instructions).

Companies that do not understand the financial ramifications of these new truths will most certainly fail. Companies that do understand and make strategies around these new truths will thrive.

It will be interesting to see how Intel pivots now.

So, if

it is simply not possible for exponentially escalating costs of node shrinks to be offset by sales of any kind.

then TSMC is doomed?

 

[MKWVentures]

Can Intel focus all its attention and resources on Oregon, Arizona, and Ireland? Even if additional fabs are needed, can they just build the new ones at those three sites? IMO, Intel doesn't have the luxury to spread out further to Ohio, Israel, and Germany.

That would be a good plan..... IF demand picks up for foundry, then expand further (But they made all these wild commitments)

First impression I get from people (sometimes these are wrong) Is that the Intel board wants to spin off IFS faster. I do know there are proposals that have been on the table for years.... and they are still discussing these. This is the only way to have clarity going forward without Pat IMO. A subsidiary doesn't fix the stock price anytime soon.... It needs to be off the books.

I think the ex IBM guys at intel have experience when giving up and ending the IDM plan. How it works ownership wise could be anything.

Products or Foundry ... the "let's do both!!!" crowd is getting smaller (which is good)

 

[kawaiwong910176]

Is it possibly because of the unwell progress of 18A? But recent discussion about 18A saying the process is "too good". https://semiwiki.com/forum/index.php?threads/intel-18a-too-good-but-design-lags.21568/
The timing is tricky, after Intel finalized the CHIPS money with DOC, Pat exited. I'm wondering if the board is pissed because of the terms, like they cannot completely spin off Intel Foundry by the end of the decade.

 

[delong.height]

I do feel whoever comes next have a less of a struggle compared to Pat. And maybe it's good to have two co-CEOs managing largely unrelated business that serve each one's own interest.

What Intel encounter in late-2020, the server roadmap is completely uncompetitive, the process node is in a severe disadvantage, the products doesn't perform well, and if they are going to spend more in manufacturing or outsourcing them all to TSMC, it's all going to be touch on future profit generations. Decisions on these problems need a strong opinion leader like Pat to organize and pave the way.

Now I assume most of the works have been done. It's up to the successors to handle the rest. They will have a good process that's coming onboard assuming they are sticking to Pat's plan strategically.
They now have a competitive server lineup and roadmap, they have their feet in GPU and AI. They at least gained experience working with world-class foundry team, and learning how they should use that tactics to deal with the internal IFS.

I do think that the struggle they have now is less of a struggle, it's more on execution. And splitting the power to 2 may be good for the long run.

And pretty much, I think maybe co-CEO is going to to live up from now on. Can they find someone like Pat but maybe better in managing client and investor expectation. I don't think that such a perfect candidate exist.

And successful companies do have co-CEO from time to time, like Netflix and Salesforce. Both were able to grow during that period of time.

 

[delong.height]

Is it possibly because of the unwell progress of 18A? But recent discussion about 18A saying the process is "too good". https://semiwiki.com/forum/index.php?threads/intel-18a-too-good-but-design-lags.21568/
The timing is tricky, after Intel finalized the CHIPS money with DOC, Pat exited. I'm wondering if the board is pissed because of the terms, like they cannot completely spin off Intel Foundry by the end of the decade.

I think that Pat left similar to how Lip-bu Tan were back a few months ago, but maybe different this time around.

Maybe Pat left because he have too much affiliate with Biden administration, and it's not good in the eyes of the board. And they want someone who is pro trump to get more favored government policy from next administration.

 

[MKWVentures]

And pretty much, I think maybe co-CEO is going to to live up from now on. Can they find someone like Pat but maybe better in managing client and investor expectation. I don't think that such a perfect candidate exist.

And successful companies do have co-CEO from time to time, like Netflix and Salesforce. Both were able to grow during that period of time.

The board has a plan that will be implemented. The only limiter was the CEO. Now that issue has been solved

 

[delong.height]

The board has a plan that will be implemented. The only limiter was the CEO. Now that issue has been solved

I do think that the plan is clear now, they are seeing good signal in 18A and maybe future generations of product. Pat is good as a vision, but he's probably the worst in managing client and investor expectation and confidence. The board need someone who can sell the story. Hence Pat is out.