News Intel CEO Pat Gelsinger retires, effective immediately — also steps down from BOD, two co-CEOs step in

[atomicWAR]

Sad days. Intel was in trouble before this and it feels like they ripped the rudder right off their ship by losing Pat. I didn't agree with all his decsions, beast lake cancellation for example...assuming rumors are true. But I do think Pat was likely the best chance Intel had to right the ship as a CEO with an engineering background. Now I worry Intel will sail along aimlessly selling itself off piece by piece to satisfy shareholders until they run a ground and truly become irrelavent. I hope I am wrong as we need good competition in the CPU space.

 

[spongiemaster]

The AI bus is careening down a hill! 😉 Poor sales of 'Copilot PC's' have both MS and Qualcomm wondering what to do next.

AI is at a very early stage, and whilst there are some neat tricks on phones and some of the chatbots, the 'AI' thing is only slowly gathering pace for general consumers. There just isn't that much interest in it.

In the last 4 quarters, Nvidia has generated over $60 billion in profits. One year. That's not monopoly money from the soaring stock value that could disappear tomorrow, that's actual money generated from product sales. Over that same time frame, AMD has generated $1.8 billion in profits. It doesn't matter where AI goes from here. AMD missed the bus.

 

[palladin9479]

Intel was leading for a VERY long time, but in a marathon, TSMC got focused by Apple and worked on minor NODE improvements every year and they eventually won the race.

Now that TSMC is using the philosophy instilled by Apple, they are leading by ALOT.

Ehh ... Apple has absolutely nothing to do with this ... at all..

A long time ago Intel eschewed trying to pursue EUV on it's own due to the massive costs involved and instead kept pursuing it's own process technology. Eventually there were only two companies with viable EUV technology, Silicon Valley Group (SVG) and ASML. ASML later bought SVG and in 2018 was able to successfully create a product that could use EUV to create semiconductors. TSMC entered into an agreement with ASML to buy their devices a couple hundred million each.

That is how TSMC got the ability to produce the fastest chips in the world. They built their process technology off ASML's EUV lithography technology. While Intel was stuck for years trying, and failing, to make their own lithography technology make smaller circuits, TSMC (and the Samsung) leap frogged them using ASML.

 

[spongiemaster]

Athlon 64, Thunderbird and the whole Zen That is almost 10 years of the last 25.

Let's not rewrite history. Zen didn't surpass Intel until Zen3.

As for Athlon. Here is the opening paragraph to Anandtech's initial Athlon 64 review:

Seemingly overnight AMD went from about to fall off of the performance charts to being competitive with Intel's latest and greatest. But there's much more to this situation than proclaiming a winner and leaving it at that; AMD has lost a considerable amount of credibility, and the Athlon 64 (and FX) of today will not bring AMD back to the heydays of the Athlon.

This was from 2003, while Athlon was initially launched in 1999. So AMD's time at the top was very brief and doesn't include the Athlon 64 era.

 

[atomicWAR]

Ehh ... Apple has absolutely nothing to do with this ... at all..

Directly that is true...indirectly I am not sure I agree. Yes Intel fumbled EUV royally but TSMC and Samsung used these small socs for cells/tablets to fine tune their new process nodes. Small chips were easier to get new nodes working at peak capacity. In turn making it more profitable for companies like AMD and Nvidia to eventually adopt said nodes do to yields increasing for larger chips more quickly than they would have had companies like Apple not been such big customers and earlier adopters of these smaller nodes.

At the very least Apple helped TMSC get ahead and in turn quickened the pace of Intel's race to irrelavency. But at the end of the day between Intel's poor manufacuring choices, resting on their laurels and AMD's big come back...Intel made their bed and are forced to lay in it. So directly you're not wrong about Apple but the indirect also played a pivitol role in Intel's down fall.

 

[LaminarFlow]

For my entire graduate school years, Intel was the top job choice for my group and a lot of fellow classmates/collaborators. I still have classmates that have worked at Intel straight out of grad school for more than a decade now. Last time I chat with them, the consensus was their jobs were not in danger, but they almost operate in "semi-retirement" mindset. The company as a whole is direction-less. Many, especially younger and more ambitious engineers have left.

Pretty sad to see how hard has the one-time giant fallen.

 

[-Fran-]

At least he'll have more time to make religious tweets/X's/posts (?) from now on.

Leaving my cynical side for a moment, this is a bad look for anyone hoping for a somewhat-pain-free recovery. They put a bean counter, then Intel is being sold for parts, even wth the US Govt saying they can't get rid of their fabs and all that.

Hard times and uncertainty for Intel employees, I'm afraid. Good luck to all, specially for the people in the graphics division.

Regards.

 

[dalamite]

I said minus a few skus....

point is without Intel, you will no longer afford AMD cpus.

But if you are an AMD investor, you'll be able to afford those AMD CPUs after you sell some of AMD stock that shut up.

 

[palladin9479]

Directly that is true...indirectly I am not sure I agree. Yes Intel fumbled EUV royally but TSMC and Samsung used these small socs for cells/tablets to fine tune their new process nodes. Small chips were easier to get new nodes working at peak capacity. In turn making it more profitable for companies like AMD and Nvidia to eventually adopt said nodes do to yields increasing for larger chips more quickly than they would have had companies like Apple not been such big customers and earlier adopters of these smaller nodes.

At the very least Apple helped TMSC get ahead and in turn quickened the pace of Intel's race to irrelavency. But at the end of the day between Intel's poor manufacuring choices, resting on their laurels and AMD's big come back...Intel made their bed and are forced to lay in it. So directly you're not wrong about Apple but the indirect also played a pivitol role in Intel's down fall.

Apple did nothing, absolutely nothing. Remove them completely from the equation.

The reason AMD chips started doing so much better then Intel's is they ditched GloFo and licensed TSMC's 7nm process to make Zen 2. Apple's SOC products were only attractive because they paid TSMC a massive pile of cash to be the first people to have access to their new production process. Take a hard look at the timings, this all happened right after ASML announced their new EUV backed process technology enabling the creating of sub "10nm" chips at high efficiency rates.

Intel is hurting because they refused to pursue this technology over a decade ago and just iterated on what they already had developed (and had patents for). ASML and TSMC were willing to spend the large amounts of time and money to develop an entirely new chip fabrication technology, Intel was not. This was a gamble that paid off well for TSMC / ASML.

Like seriously, why do iPeople think Apple is involved with this. That's like saying Microsoft is responsible for Intel's problems because people use Windows for gaming and games need GPU's and nVidia use's TSMC instead of Intel. Some six degree's to Keven Bacon stuff.

 

[dalamite]

In the past, AMD was the underdog and people were rooting for them. Now that they came ahead, people are rooting for Intel to rise from the ashes.
Can't win either way.

I'm a bit different from those people. I want Intel to die because they deserve it. And selfishly, I want my AMD stock to go up.

 

[palladin9479]

For those wondering the relationship between Intel, Samsung, GloFo, TSMC and ASML. The first four research and develop the techniques to create semiconductors while the last only produces obscenely expensive chip printers. No matter how amazing your processor design is, you are going to be limited by the size of your transistors with how much thermal energy you generate and how expensive it is going to be to produce. The higher your chip printing resolution, the smaller your transistors can be and the more efficient your printing process.

[News] Price for ASML’s Hyper-NA EUV Rumored to Double, Causing TSMC, Samsung and Intel to Hesitate | TrendForce News

Semiconductor equipment giant ASML plans to launch Hyper-NA Extreme Ultraviolet (EUV) machines by 2030, signaling the advent of the Angstrom era for s...

www.trendforce.com

 

[ezst036]

Directly that is true...indirectly I am not sure I agree. Yes Intel fumbled EUV royally but TSMC and Samsung used these small socs for cells/tablets to fine tune their new process nodes. Small chips were easier to get new nodes working at peak capacity. In turn making it more profitable for companies like AMD and Nvidia to eventually adopt said nodes do to yields increasing for larger chips more quickly than they would have had companies like Apple not been such big customers and earlier adopters of these smaller nodes.

At the very least Apple helped TMSC get ahead and in turn quickened the pace of Intel's race to irrelavency. But at the end of the day between Intel's poor manufacuring choices, resting on their laurels and AMD's big come back...Intel made their bed and are forced to lay in it. So directly you're not wrong about Apple but the indirect also played a pivitol role in Intel's down fall.

You're creating a pretzel shaped formation trying to insert Apple here.

Well over half the phone market, probably 60-75% of it is Android phones. My only thought is on the CPUs that are contained in those said phones and tabs.

TSMC has had access to small-node phone CPUs long before Apple came along and even after as well. You would have to imagine that all phones on the planet are Apple phones for this to work and that all small socs are isocs. It just isn't the case.

What you're saying in regard to small socs and ease of getting to new nodes makes some sense, I don't know the fine details to elaborate.

 

[tamalero]

Seems more like a fall guy.
He came in and intel was a hot mess. They seem to be trying but it takes a long time to turn around not only design issues but the decades long manufacture issues he was handed.

Not saying he is faultless for intels current status. But 4 years is not a long time for the hand he was dealt.

Still. I do not believe he will suffer one single bit. Most fall in CEOS get their parachute and retire.

If they were in charge, they would sell off every asset Intel has for maximum profits, and you'd have nothing left but a hollowed out carcass in a few years. - Asset stripping.

Technically, that is what US companies have been doing in the past 20 years.
See Boeing and Disney, cost cutting to the bone and inflating their value via stock buy backs while their products flounder.

 

[tamalero]

Co-CEO's? LoL. There is a reason there aren't co-Captains on a ship.

That is an in cheek for company splitting.. Fundry on one side and Chip making on the other.

Or just "maintaining the boat barely afloat"until they get another possible fall guy or another attempt to fix the company.

Competition is definitively good for all of us. Hopefully AMD recognizes that, and doesn't abuse their position. But, being a "publicly owned company", I wouldn't count on it.

That said, the 9950 came out at $150 less than the 5950?

Plus, technologically speaking, comparing the price of today's CPUs to those of Athlon or earlier years really is comparing apples and oranges. Companies have to make their profit too, as much as we'd like everything to be cheap. (In most cases, you do get what you pay for...)

And there's this nasty thing called inflation.

I'm still angry at the threadripper price hike with the 3000-4000 series XD

 

[Priscus]

Reported in British press that he failed to achieve the needed turn around in spite of huge injection of Biden's cash injection! Is that the anti inflation package, or has US been specifically giving state aid to Intel? Not wanting this to be political, but outside of US, we don't get the details of the back story.

 

[Thunder64]

Let's not rewrite history. Zen didn't surpass Intel until Zen3.

As for Athlon. Here is the opening paragraph to Anandtech's initial Athlon 64 review:



This was from 2003, while Athlon was initially launched in 1999. So AMD's time at the top was very brief and doesn't include the Athlon 64 era.

You must be joking. Other than a breif period of P4C Northwood P4's, Athlon's dominated from about 1999-2006.

 

[bill001g]

Reported in British press that he failed to achieve the needed turn around in spite of huge injection of Biden's cash injection! Is that the anti inflation package, or has US been specifically giving state aid to Intel? Not wanting this to be political, but outside of US, we don't get the details of the back story.

It was a separate thing called chips and science act. It was one of the very few things passed that had some support from both parties. Of course like all legislation there is a bunch of garbage they hid that is only kinda related. The so called anti inflation package was mostly environmental regulations.

 

[redgarl]

Let's not rewrite history. Zen didn't surpass Intel until Zen3.

Zen 1 was trashing Intel in application, just not into games.

AMD was offering 8 cores when the Client sector max offering was 4 cores.

 

[thestryker]

ASML and TSMC were willing to spend the large amounts of time and money to develop an entirely new chip fabrication technology, Intel was not.

Intel actually invested significantly in ASML to help EUV development as did several other companies. The industry as a whole picked a winner and propped up ASML's EUV development due to how capital intensive it was.

The problem is that Intel management was stuck in a money above all mindset and if 10nm had hit the timeline targets would have saved them short term money. They were always going to need EUV machines and it likely would have been better to be acquiring them early even if 10nm hit targets, but when you're run financials first things like that just don't happen.

 

[redgarl]

In the last 4 quarters, Nvidia has generated over $60 billion in profits. One year. That's not monopoly money from the soaring stock value that could disappear tomorrow, that's actual money generated from product sales. Over that same time frame, AMD has generated $1.8 billion in profits. It doesn't matter where AI goes from here. AMD missed the bus.

Basically thriving on Datacenter alone since all other sectors are down.

However, AMD will eventually get their part of the pie, it is just a matter of time and the Accelerator TAM is just growing. With democratization, Nvidia will not have the same hold on the market as they have at the moment.

Not to mention that Intel is dead and their business will go to AMD. I will not be surprised to see AMD getting 15B$ revenue per quarter in the next 24 months.

 

[DavidC1]

Poor Pat. Can we assume that 18A is broken too? I read about his career on Wikipedia - it seemed highly unlikely that he was a fit for the job. If they were looking to hire a technical person - why not at least somebody with PHD?

He's not just an engineer, but a star engineer hired straight from high school. He got his degree in electrical engineering while working at Intel.

He was the Chief Architect of the 486 chip which saved the company as they were doing Itanium-ish project with an alternate chip with a completely different ISA.

Then became the youngest CTO. He was fired by the bean counter Otellini in 2008, the same guy that missed the big Apple deal because he thought the iPhone would be "low volume". Gelsinger should have been the successor, or at the very least stay as CTO. Too many good people get fired for one mistake, even if it's a relatively big one. They should be demoted or reprimanded, but instead they get eliminated.

 

[Makaveli]

Pat Gelsinger: I'm betting the entire company on 18A.

<<Reuter reports 18A has low yields per BroadComm>>

Also Pat Gelsinger: Peace Out.

lmao

 

[palladin9479]

Intel actually invested significantly in ASML to help EUV development as did several other companies. The industry as a whole picked a winner and propped up ASML's EUV development due to how capital intensive it was.

The problem is that Intel management was stuck in a money above all mindset and if 10nm had hit the timeline targets would have saved them short term money. They were always going to need EUV machines and it likely would have been better to be acquiring them early even if 10nm hit targets, but when you're run financials first things like that just don't happen.

Ehh no.

ASML is a Dutch company, Intel had no investment in them during the development of EUV. Back in 1999 the US DOE put together a cooperative research program to study EUV and ASML was one of the members of that group along with Intel, SVG and another manufacturer. Under this sort of agreement the funding comes from the US Tax payer and thus Congress becomes the approval authority for the license. Intel, SVG and ASML all received such approval. Intel decided not to invest heavily into it while SVG and ASML went the other way. Eventually ASML bought SVG and thus become the only manufacturer with EUV capable products.

Over a decade later (July 2012) Intel agreed to provide ASML $4.1 billion in exchange for 15% ownership as a way to accelerate the transition to 450mm wafers and expand on the existing EUV technology. TSMC and Samsung also participated.

https://www.reuters.com/article/us-intel-asml-idUSBRE86819B20120709/

Six years after that ASML released a set of new equipment that enabled TSMC to develop 7nm fabrication technology. Intel was on 14nm and was struggling with yields on it's own 10nm process, the core issue was it's lithography devices were just not up to the task to print 10nm chips economically. Unlike TSMC who had embraced EUV and went all in with ASML (their cooperation goes back to the 70's), Intel had been trying to develop and use their own products. That cost them immensely and ultimately they had to just use ASML and then try to play catch up with Samsung and TSMC on fabrication process's.

Remember ASML is just making the chip printer, you still need to develop a process to use those chip printers to create functional processors out of silicon wafers. Each kind of chip printing technology requires it's own set of fabrication process's that are just as expensive to develop.

 

[thestryker]

Over a decade later (July 2012) Intel agreed to provide ASML $4.1 billion in exchange for 15% ownership as a way to accelerate the transition to 450mm wafers and expand on the existing EUV technology. TSMC and Samsung also participated.

If it wasn't for the large infusion of money it's not guaranteed EUV makes it to market. There was nobody else anywhere near ready for EUV and ASML still wasn't that close. Their first prototype machines were in 2010 and production ready wasn't until 2016. That's a very long time to go sinking a lot of manufacturing capital into something without return.

Eventually ASML bought SVG and thus become the only manufacturer with EUV capable products.

ASML bought them in 2000 they were not a real EUV competitor at this point given the first production machines were still over a decade away.

 

[palladin9479]

If it wasn't for the large infusion of money it's not guaranteed EUV makes it to market. There was nobody else anywhere near ready for EUV and ASML still wasn't that close. Their first prototype machines were in 2010 and production ready wasn't until 2016. That's a very long time to go sinking a lot of manufacturing capital into something without return.

ASML bought them in 2000 they were not a real EUV competitor at this point given the first production machines were still over a decade away.

Umm ... EUV has been around since 1999. It didn't make financial sense at that time due to other process's being cheaper, but it existed. ASML did not need Intel, Samsung and TSMC were already lined up.

I know you really want to make this about Intel as some sort of hidden genius, they are not. They missed the boat and it's pretty well documented with how poorly their 10nm process was going. The history is known and a mater of public record, no amount of pretzel twisting is going to change that.