Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Intel CEO Pat Gelsinger Retires Amid Company's Decline
- Thread starter MacRumors
- Start date
- Sort by reaction score
There is probably more computing power in my iPhone 16 Pro Max than in all the computers used to land on the moon. Think about innovation and technological progress.
Intel bought StrongArm from Digital and sold a line of ARM (still has the license)
XScale - Wikipedia
After Itanium flopped, Intel adopted 64bit x86 instructions from AMD
x86-64 - Wikipedia
Consider also, Intel's iAPX 432 (from ~1981-1986). The iAPX was intended to be programmed with the Ada language and other super-smart compilers, because it had hardware designed for object-oriented programming & other high-level languages.
Unfortunately the compilers weren't very good and the iAPX was notoriously slow. Not all R&D leads to a viable product and not all ideas are good ideas... that said, I do still want to buy an Itanium machine for fun
They’re not even competing effectively against the only other large vendor producing x86 chips, much less Apple Silicon! They’re in Walmart because they exist. If they didn’t, it’d be AMD or someone else would be stocking those shelves with yesterday’s products! Existing will get you a loooooong way, but failing to compete will take down even the largest company. Which is as it should be.
They missed the boat starting ALLLLLLLL the way back at the iPod. In an alternate universe, Intel bent over backwards to learn what they needed to in order to produce efficient performant processors (perhaps smartly not dependent on x86) for Apple at each step when they needed a new level of efficiency and performance. Learning those lessons put them in a good position when different vendors were doing flagship phones, tablets, ultralight laptops and Intel was able to provide them better solutions across the board year after year. AND, with all that experience in small powerful efficient design, by the time they decided to make a 5G chip, they were WELL ready for that challenge and it was no big deal to build that tech into their chips for those vendors that wanted to provide those solutions. And, because Qualcomm in that timeline was challenged, they didn’t charge inordinate amounts of money for their modems, making them even more broadly available.
The more I think about it, all Intel had to was “make what Apple wanted”. Even, like AT&T before the iPhone, they had no idea what Apple had up their sleeves, just meeting the challenge would have made all the difference.
“Intel's next CEO could be Lip-Bu Tan.”
www.tomshardware.com
Intel's next CEO could be Lip-Bu Tan, a former member of the board who reportedly clashed with Gelsinger about strategy
No decision has been made yet.
I would say that is simply the normal divergence of comparison charts zeroing at the start point, and you using the 5 year comparison that started a few months before the date you are comparing. Picking a different period gives a different zero start date, at which point that obvious correlation isn't as obvious.
Max:
1Yr:
I simply wanted to focus on Apple Silicon because I assumed that was the most relevant threat that Intel was facing in 2020. But if mobile in general was the threat to x86, then, you can certainly look at a wider range - I don't disagree.
The divergence in May 2024 is interesting too!
Yeah, since those are just stock price charts, it probably had a lot to do will this result in April 2024: https://www.reuters.com/technology/...d-quarter-revenue-below-estimates-2024-04-25/
But I would have guessed that the most relevant threat Intel was facing in 2020 was the global shutdown of manufacturing due to COVID.
The iAPX432 (formerly the 8800) was the inspiration for David Patterson of Cal to develop their first RISC design circa 1980. From what I've read, Patterson was abhorred by the complexity of the iAPX 432 and decided to go the other direction. Cal's EECS and CS departments were also familiar with Seymour Cray's design philosophy of the CDC6000 and CDC7000 series machines. That is relatively few and simple to decode instructions, make those instructions execute as fast as possible and a large registers set (the CDC 6000/7000 series had (8) each of 18 bit A registers, 18 bit B registers and 60 bit X registers.
Intel also had the 860 which was designed for numerically intensive applications, but they ticked off their customers by EOL'ing it circa 1997. Intel's Itanic processor was pushed primarily to kill off the likes of MIPS, SPARC and POWER processors.
PC shipments actually increased in 2020 as people began working from home so I don't think that was the culprit.
Yeah, but Intel doesn't really manufacture many PCs, they largely manufacture chips, and that chip manufacturing (that was already having issues) took a turn for the worse:
The sun doesn't rise because the rooster crows... Intel isn't having problems because Apple left them; Apple left Intel because Intel were having problems.
This article is an amazing read. There we (not me personally) are complaining that Apple is still ‘stuck’ on 3nm and only four years ago Intel failed at transition from 10nm to 7nm. Also, ‘company had identified a “defect mode”’ is hilarious. Defect mode! Comedy gold.
I think Apple Silicon was a massive blow. Not because Intel lost tons of revenue from Macbooks, but because Apple showed a chip that didn’t need a fan (my husband’s early 2020 Macbook Air turned into a portable helicopter when he was scrolling through Tumblr), shockingly faster, with low power consumption… and thus proved it’s possible to drop Intel and make a huge leap forward rather than compromise. And Rosetta 2 made Intel apps Just Work (TM) on AS as well, hammering the point home further: Intel is not actually necessary. It’s possible to make something completely new that destroys Intel’s pitiful roadmap. That M1 presentation was a BlackBerry meets iPhone moment.
From another article linked on The Verge: “when a $1,000 M1 laptop can outdo a maxed-out, $6,000 MacBook Pro with quadruple the RAM and Intel’s best chip, while also running cooler and quieter in a smaller and lighter form factor and with twice the battery life, where do competitors even go from here?”
I am the opposite of surprised as I watch Intel scrambling and reducing its workforce, while Apple casually plops the absurdly fast M4 into an iPad and extends MB Pro’s battery life to 18-24 hours because it can.
I think Apple Silicon was a massive blow. Not because Intel lost tons of revenue from Macbooks, but because Apple showed a chip that didn’t need a fan (my husband’s early 2020 Macbook Air turned into a portable helicopter when he was scrolling through Tumblr), shockingly faster, with low power consumption… and thus proved it’s possible to drop Intel and make a huge leap forward rather than compromise. And Rosetta 2 made Intel apps Just Work (TM) on AS as well, hammering the point home further: Intel is not actually necessary. It’s possible to make something completely new that destroys Intel’s pitiful roadmap. That M1 presentation was a BlackBerry meets iPhone moment.
From another article linked on The Verge: “when a $1,000 M1 laptop can outdo a maxed-out, $6,000 MacBook Pro with quadruple the RAM and Intel’s best chip, while also running cooler and quieter in a smaller and lighter form factor and with twice the battery life, where do competitors even go from here?”
I am the opposite of surprised as I watch Intel scrambling and reducing its workforce, while Apple casually plops the absurdly fast M4 into an iPad and extends MB Pro’s battery life to 18-24 hours because it can.
No kidding 🤔 And where do these chips end up being utilized?Yeah, but Intel doesn't really manufacture many PCs, they largely manufacture chips, and that chip manufacturing (that was already having issues) took a turn for the worse:
The sun doesn't rise because the rooster crows... Intel isn't having problems because Apple left them; Apple left Intel because Intel were having problems.
Both.
But also, Apple likely would've left regardless. They may have still been interested in Intel for manufacturing, but as far as core design goes, Apple is happy to be independent and in control, even when they aren't best (and they certainly are up there).
True, but since Windows marketshare is around 4 times the marketshare of the Mac, that large 2020 increase in PC sales should have hypothetically increased Intel's sales vs 2019 to offset the loss of Apple quite a bit.
I very specifically said "manufacturing" because that seemed to be Intel's biggest failing in 2020. My thinking that the COVID interuptions didn't help could be off base, but I doubt it helped, and I doubt it helped their chip design at the time, as things got even worse in 2022-2024 for Intel with the Core 13 and 14 voltage failures that caused them to have to replace quite a few chips and extend the warranties on those that didn't yet fail. None of that should have affected or been affected by Apple.
Intel has quite a few internal issues beyond the loss of Apple, including AMD outperforming them and apparently now outselling them in datacentres, which I expect is a huge problem due to the much higher profit margin at that level.
AMD records its highest server market share in decades — Intel fights back in client PCs
Another strong quarter for AMD.
Hey, I agree that Apple switching away from Intel was a huge blow to Intel, but a very large part of that was prestige. Windows on ARM existed for a couple of years before the Apple switch (or 8 years if you want to count Windows RT, and another decade if you count Pocket PC), but Apple showing that Mac OS could work so well on ARM, even running legacy, did impress and make people question why Windows on ARM wasn't better, as well as remove some of the trepidation of the eventual potential of ARM for Windows. But Intel has a lot of excellent outside competition even within X86 and a lot of internal issues that definitely make Apple's switch to AS as only one of the things leading to its current situation.
Last edited:
Apple only started using Intel in 2006 and both companies had previously survived quite well without each other.
If Intel was performing flawlessly, they would likely find the remaining market more than sufficient, but they are nowhere near flawless at the moment and AMD is performing pretty consistently. Add in the AI "buzzword of the moment" effect and Nvidia is suddenly massive. And TSMC is crushing it, while I'm not hearing much about anyone wanting to take Intel up on their third party foundry offering (though I'm not in the loop, so maybe they have lots of business I just don't know about.)
Apple leaving didn't help Intel, but it isn't what is really doing the damage because they don't really compete directly with Intel... sure, Apple now makes CPUs, but they are their own customer. Which is why I find the whole monopoly complaint weird, as Apple just seems to want to control ALL the pieces of their product. It is pretty easy to avoid Apple parts by not buying Apple products.
One difference between the MacOS. X + descendants world and the MS-Windows world is that Apple put a lot of effort in keeping MacOS portable between processors, whereas MS had a brief spell of making Win NT portable but gave up after a very years. Apple also burned some bridges by forcing MacOS apps to use only 64 bit code for Macs on Catalina and later, in order to NOT have to deal with the ugliness of Intel's 32 bit instruction set. That's why Rosetta 2 works significantly better than MS's Intel to ARM translation scheme. (Having written that, I would be happy to see MacOS Mojave running on emulation for one App that was never updated to run on Catalina.)
Intel's other problem was the bane of many tech companies, where the first generation of CEO's were people who knew the tech, second generations was people who knew how to make the tech, and third generation were the Bean Counters. Intel also forgot how to be flexible, witness the Itanic (cute name for Itanium) where Intel was discouraging naysayers in academia from commenting on the problems with the Itanic. AMD saw an out by crafting on a RISC like 64 bit instruction set, by adopting AMD's design as allowed by the 1970's era cross licensing agreement between the two companies.
As others have noted, Intel was having problems with getting new process nodes on-line after years of being at the top or close to the top. I suspect that part of the problem was that Intel forgot how to keep and inspire the kind of people who could tackle problems of getting a new process node on-line. I also suspect that Intel management for got Andy Grove's comment about "only the paranoid survive".
Intel's other problem was the bane of many tech companies, where the first generation of CEO's were people who knew the tech, second generations was people who knew how to make the tech, and third generation were the Bean Counters. Intel also forgot how to be flexible, witness the Itanic (cute name for Itanium) where Intel was discouraging naysayers in academia from commenting on the problems with the Itanic. AMD saw an out by crafting on a RISC like 64 bit instruction set, by adopting AMD's design as allowed by the 1970's era cross licensing agreement between the two companies.
As others have noted, Intel was having problems with getting new process nodes on-line after years of being at the top or close to the top. I suspect that part of the problem was that Intel forgot how to keep and inspire the kind of people who could tackle problems of getting a new process node on-line. I also suspect that Intel management for got Andy Grove's comment about "only the paranoid survive".
I wouldn't say there's a significant difference. At the low level, XNU used to run on 68k, PowerPC and other archs; now it runs on x64 and ARM64. NT used to run on PowerPC, MIPS, and others; now it runs on x64 and ARM64.
Sure, in practice, Mac apps are more likely to run natively on ARM64 than Windows apps, but that's more a function of Apple having designated a clear path forward, whereas Microsoft continues to support both.
Intel's more recent problems seem to be more about hubris, and Gelsinger is hardly a bean counter.
Perhaps, although they do have external threats that are hard to control.
True, as I guess Apple learned from Copland and fortunately didn't squander NeXTSTEP. The weird thing for Windows IIRC, though, was that Windows 10 on ARM only supported Win32, but at least they added Win64 for Windows 11.
Yeah, I miss Woz, too!
As much as I agree with that overall concept, Gelsinger appears to actually be a complete exception to that rule. According to Wikipedia, he was a tech that actually started his career at Intel in QC, earned his engineering degree, and became the lead architect on the 80486. He was named the youngest VP at 32 and was mentored by Andy Grove himself and became CTO in 2001. After leaving Intel in 2009 he was brought back in 2021 as CEO to try to pump up the share price after a re-org under pressure from an activist investor. That reminds me of the Carl Icahn days at Apple, so I guess it makes sense that Gelsinger got a boatload of cash to step into that lousy scenario. Either way, his credentials actually sound like exactly what you would want, but I don't expect 3 years would be enough time to turn a ship of that size, even if he was capable and there were no other outside factors.
Itanium was such an interesting scenario, as it showed that the X86 lock-in power that was always such an advantage for Intel has also become an albatross for them. Every time they have tried to move on to a new instruction set, they keep getting dragged back to it, and now they have to deal with another company that is able to design CPUs with it just as well or better. If Jobs hadn't gotten rid of the Mac clones, assuming Apple even then survived, I cannot imagine how hard it would have been for them to switch to Intel or AS with other companies continuing to manufacture clones using the older architecture. That said, at least Apple could control the features of the OS to push users to move, while Intel is at the whim of Microsoft. It is funny how everything feels like it has sort of looped back to the 1980s Edit: maybe 1970s
Last edited:
Which was my point (and technically exactly what I said in the rest of that paragraph after that sentence.)
Yep.
Even worse, Windows RT (Windows 8 on ARM) IIRC had no emulator at all. And they didn't really want third parties making Win32 apps for it.
Here's a take on Gelsinger that isn't so positive.
If we contrast that with Steve Jobs, a key difference appears to be: when faced with technologies that didn't really sell, Gelsinger thought, "but they should sell!". Jobs, meanwhile, didn't have that nostalgic attachment. He killed them off and moved on.