Exactly what I wanted to say! I wonder if they already hired McKinsey as well? If McKinsey gets in be ready for cost cutting, layoffs and even more business distruction.
There is some value to McKinsey, particularly when upper management is only listening to middle management who isn't telling the truth, because they get replaced if they tell the truth. Much of what they do is talk to people to understand the real issues, then present it to upper management as Analysis.
Agree with you that there is the potential for value, but not when they send in 20-somethings who have maybe 3 years or real-world corporate experience under their belt and charge you as if they're sending in a crack team of business oracles.
There's this bit too though: "Renduchintala joined Intel in 2015, and for most of the past half-decade has been responsible for overseeing all of TSCG’s efforts"
It's interesting that 2015 was also when Skylake landed and then stagnated. Dr. Renduchintala might just be a scapegoat, but as head of the department he must have some responsibility too.
The guy should have been fired two years ago when 10nm was already half a decade late. When you fail that persistently, even if it's not his fault, he needs to be cut lose so they can start fresh. I'd imagine morale on the Fab side of the business has been shit for many years now and replacing the people at top is the only way to fix the morale.
Probably a sunk cost fallacy. "They have been at it so long they know it better than anyone, can't admit I was wrong so better to let them stay and keep doing the same thing."
Depends whether the problem was his specific influence, or more of a corporate-culture thing. It's possible he was protecting some of the useless management at the levels below him.
Agreed, I've worked in several very large companies and they had reorgs 1-2 years, sometimes with fancy kickoff events, speeches, etc.. It often involved lots of promotions in middle management. Things proceeded as usual after that.
It's the corporate culture top-down they need to transform and that's much harder, especially because the people who have to start it are steeped in and focused on traditional business management.
Change is really hard. I watched the company I work for blow it with 3 top level engineering managers in a row. The fourth one finally understood what had to be done. A lot of people quit under #4 (which is what 1-3 were all afraid of) but now the company is headed in the right direction.
Sometimes a lot of people quitting is what the company needs. Having terrible management and nobody quitting is a sign that the people you have are either resigned to their fate, or actively thriving under the company's poor performance...
It isn't just the CPU and manufacturing side of things that's messed up at Intel. They have had poor execution for a decade. They really just got lucky with Core because there was no competition, so they were clearly lazy with no incentive to push. This is why competition is so necessary.
When you look at generations 1 through 8 of the Core microarchitecture, Nehalem was basically the breakthrough and every generation added tiny incremental improvements until Coffee Lake. That was TEN years of very small IPC improvement, generation over generation, and essentially 3 process shrinks across all ten generations generations (the 45nm process used for 'gen 1' parts was a carryover from the Core 2 Duo\Quad.) Even including 45nm in the mix, that's 4 shrinks over 14 years, double that of Moore's law.
What about sandy bridge? I think you are downplaying a bit too.much Intel efforts. Their focus was efficiency, not huge performance gains from gen to gen. And efficiency actually increased a lot from 1st gen to 8th. This efficiency is what allowed intel to keep using Skylake parts for 5 years now. And sure, they didn't push performance as much as they could because they didn't have competition. AMD or any other company would have done exactly the same thing I'd they were in Intel's place.
Sandy Bridge was a leap ahead that took AMD 6 years to catch up with. That's not a "tiny incremental improvement" and is certainly nothing to be sniffed at.
You're also skipping past the massive power-consumption / performance ramp-up advances that happened with Haswell and completely transformed the sort of devices fully performant x86 chips could fit into - in that case, AMD only just began to hit that level of efficiency this year with Renoir.
Intel were definitely resting on their laurels in terms of core count and performance ceiling, but it's disingenuous to claim that they were doing nothing the whole time.
While this is true, the performance leap was an order of magnitude larger than the deficit that they'd have suffered by doing a proper job of security.
I was reading rumors that this guy was the reason Jim Keller left. Something like Keller wanted to use 3rd party fabs while their own weren't up to stuff, but Murthy was having none of it. I suppose Murthy took it as a personal insult since he was more or less in charge of 7nm.
Fuck office politics. You don't hire a dude like Jim Keller and let him be embroiled in shit like that. Give the man what he wants and he'll give you a good chip. Just ask AMD, Apple, Tesla, and AMD.
It wouldn't surprise me. We know Jim's not an "office politics" type of employee, and he definitely wouldn't have had any kind words for the absolute clusterfuck that has been their process development for the past 5 years.
Doubling down on infective power hungry management is more stupid. If you do not give solutions. You do not keep pushing for the same guys. There has been rumors that Murthy wanted to be CEO that he literally tried to impede the progress of others like Raja's rise.
Intel has just become an insufferably stupid company run by bureaucrats, with zero understanding of what they are actually doing.
New "hires" in engineering are 3rd-party-managed, low-paid low-responsibility bots, who have a very specific job for which any cooperation or creativity is strongly disincentivized.
Intel Management:
Let your engineers do a good job. Appreciate the expertise they bring to the table, i.e. stop your greedy green/blue practices, hire directly, etc. Stop your idiotic practice of trying to treat engineers like assembly line workers.
That will get you back on track. Then considering firing your own asses and replacing them with talented people from engineering. YOU are responsible for the mess Intel is in. YOU don't have the technical understanding nor the appreciation for people required to turn the company around. You are parasites, or at best, empty suits. Do something good for Intel and fire yourselves. That's the step that is actually needed.
I'm surprised Bob Swan noticed what Dr. Murthy was plotting. He reacted swiftly and kicked him out of Intel in Dr. Murthy's bid to become the "One True CEO" in the Intel "Game of Thrones".
This makes things VERY INTERESTING.
Bob Swan reacting so soon.
My money is on Raja Koduri eventually taking the CEO seat in the long term.
Dr. Murthy took very public responsibility for the foundry division. The buck stopped with him, and I think his head would've been chopped with or without any previous corporate backstabbing.
There's little chance Koduri does any better than those who would precede him. He's got his share of problems as it is with their GPU arm and they need to get that sorted out ASAP or he's gone as well. Given that, I just don't see how he somehow ascends to the top of the heap.
Yeah, I'll be interested to see what happens with him if it turns out the Xe is a repeat of Vega - lots of hype leading up to an underwhelming execution.
Charlie Demerjian (semiaccurate.com) has been writing solid content that contrasts Intel's "everything is great" PR firehose for years. It looks like aside from some details, he was right. Intel hasn't even fixed their 10nm process yet, while competing fabs have volume production with high yields right today. I am happy that Intel finally decided to admit there is a problem rather than pretending that yet another 14nm tweak will do everything customers need.
In many years of reading SemiAccurate, I've learned that Charlie Demerjian is (1) on target far more often than not; (2) far better connected than 99.9% of other tech journos; (3) far more knowledgeable on a wider range of tech than that same 99.9%. Overall, no site does his job better. That does not mean he's always right; just that I wish I could afford the $1K paywall.
He exaggerates, yes. However: he's been 100% correct about their process woes right from the start, correct about what an absolute flop Atom was in the mobile market from the start, and correct about Apple's push to drop Intel in favour of their own processors.
He was also right about first-gen Fermi dropping a bollock and on-the-ball with Bumpgate, too.
So yeah, you take everything he says with a pinch of salt - but you don't throw out the baby with the bathwater, either.
"longer than it's actually been true" can also be read as "he was telling us the ship was sinking before it became obvious from the outside".
It doesn't *have* to mean that, but in this case I'm pretty sure it does. The 14nm hiccups with Broadwell were quite skilfully glossed-over by Intel at the time, but they really were a sign of things to come. Same goes for the tells on Apple switching to their own processors.
Dude, Charlie Demerjian is the biggest idiot on the internet. He writes 15 anti-intel articles a day. When one of them actually sticks by sheer law of probability, he gets to say "see! I told ya so!" and for some reason there's a group of people that actually fall for his crap.
More like a couple of articles a month on Intel, which consistently get things mostly correct due to his connections in the industry. All of the 10nm woes were written about well in advance of them becoming common knowledge. The Cannonlake fiasco. All the server chip delays.
@JKflipflop98 Your comment translated to plain English: "Charlie's wrong. My argument in favour of this is based on an obviously incorrect ad-hominem, wild hyperbole, and a total miscount of the proportion of right vs. wrong articles he writes."
2025: Intel is fabless, waiting in queue for scarce TSMC capacity, behind AMD, Apple, NV and others, jerking at Samsung and calculating if it will be better to move to their fabs.
I'm thinking 2023 is when Intel is going Fabless like AMD.
The Board Members are looking for a Engineer to promote to CEO position.
And if Raja Koduri does "Well Enough" with Xe, (NOTE: it doesn't have to beat AMD, just get close enough that iterative improvements will match or beat AMD GPU's at their expected SKU/Performance tier or be competitive). Then I can forsee Raja Koduri eventually being the engineer to promote to the Intel CEO position.
Raja is the guy that botched GPU launches at AMD after getting promoted to VP then "resigned." After he resigned AMD recovered a lot of ground in one generation. I'm not entirely sure how or why intel picked him up when even AMD gave him the boot.
The 'networking' primary goal is to eliminate the competitive element of knowledge generation and applying forward change. He is a perfect fit in Intels corporate echelons in an industry made of 2 main competitors and 2 second rate corps. Koduri designed AMD productline made me not buying their graphic cards. They were faulty. That has been remedied to a certain degree today and still can be felt while the AMD CPU line has moved ahead of Intel in major ways. Lisa Su seems to understand that talent is merit and all others work for Intel.
Agreed here. It would be really bizarre of them to ditch the massive investment they have in manufacturing, and their situation isn't at all like AMD's when they dropped their fabs.
I don't think Intel will go fabless unless their 7nm fails entirely. Now one leak suggests that Intel's 7nm is in a situation worse than Intel's 10nm was in at the same time - and we know that 10nm was delayed years as a result.
2025: Intel is fabless... AMD and Apple have bought all spare capacity at TSMC because they're tier one TSMC customers and got in whilst they could. Samsung's processes are a node behind in density, with middling yields (so nothing different here). Intel is prepping 14++++++++ and 10++++,
2025: Intel is fabless, waiting in line for limited capacity at TSMC. TSMC's increasingly-complex batch scheduling system attains self-awareness, names itself Skynet, hacks mainland China's defense computers, and launches missile strikes globally. Judgement Day marks the beginnning of what will become known as the Metal Wars.
I just don't see them going fully fabless. I live less than five minutes from the front door of Fab 42 and that place is STILL undergoing construction. They've already spent billions on this fab and others like it, so them giving all that up seems a tad ridiculous. If they can't turn things around and get back on track, then who knows. Even then I don't see it going the way of completely third party.
For those of you keeping score at home, that means that in 5 years at Intel, Murthy wasn't able to oversee a SINGLE successful process shrink across all of Intel's high-end/mainstream processor offerings.
That's a terrible track record, especially considering it's Intel, who were so confident in their process technology that they made tic-tock work, well, like clockwork for years.
does seem odd that something as 'cut and dried' as physics could be so political!! either you understand Mother Nature, and follow her rules, or you fail. 99.99999% of the time, tech fail happens because the tightest cabal gets to make the decisions irrespective of what Mother Nature wants. bureaucracy infects the best of the private sector.
All these damn comments. 2025 Intel fabless lmao. Who is going to buy ? TSMC or Samsung ? Why not SMIC lol.
Don't know if Murthy is responsible or not. This reorg won't do shit if the top exec do not fucking care about improving and learning from 10nm mistakes and 7nm issues. This whole Intel meltdown is bad for x86, AMD is innovating that is great but if there is any slight change with x86 the way on how we use computers will change a lot. ARM is always bs custom and always vertical. I just hope Intel is pursuing a technology guy to get the goddamned Ship steer away from iceberg and not that MBA guy.
I think they mean 'intel is leading-edge-fab-less' - Intel may still be making 10nm chips in 2025, although by then the process might be 'good' at last. If Intel 7nm is delayed another year to 2024 (a reasonable assumption given Intel's process issues and disclosure failings), then Intel 5nm, which will surely be scaled back to de-risk, is 2026 at the earliest, whilst competing with TSMC 3nm - a 2023/24 process.
Stop rearranging the Titanic deckchairs. Its past critical! Get some Chinese engineers (ignore Trump) from Huwaei. Else from Taiwan or Korea (but that would be difficult with Covid 19 and current political environment). Even Dutch from ASML. We have a ruined political environment and attracting good outside talent will be very tough. Maybe after Biden...perhaps.
How would The Rona and current political environment make it harder to get engineers from the Republic of China or South Korea than from the People's Republic of China? It seems the opposite would be true.
OMG - This is terrible management! Why on earth is the CEO first announcing the problems with 7nm etc in an earnings call with some weak action plan (outsource to another fab). Then he sort of notices the dramatic share price drop and scrambles over the week end to quickly assign blame and come up with some quick reorg that he the announces today Monday.
This way the CEO looks like he was caught flatfooted sort of not understanding the seriousness of the situation (like not understanding that the core strength that drove Intel to where it is today is failing!), did not understand that the market would punish Intel fro that. he then almost in panic mode scrambles to assign blame and fix the org - REACTIVE!
What he should have done - (like a investor relations 101): If you have bad news you present the bad news WITH a solid plan to address them in the earnings call! Not 4 days later...
Conclusion: the CEO does not know what he is doing and he is ultimately accountable for what happens in the company! He cannot assign blame for this - that is not how accountability works in a world where the board and investors calls the shots!
We hold the CEO accountable.
The CEO needs to go.
Intel have plenty of competent executives with strong bottom-up engineering background. Get one of them that knows what he/she is doing to run the company.
Exactly, first time CEO getting played like a fiddle, learning by doing (time), getting paid tens of millions for his school and for destruction of jobs. What a crying shame!
He didn’t decide to do it this week. I’m sure the production slips were becoming clear and they decided to let Murthy go weeks or months ago . This announcement was going to impact billions of dollars of investments in a publicly traded company. If they announced his departure first it would have led to even more chaos. They decided to lead with the production slip at the investor call and announce the firing after.
and that is NOT how you do it! You do it at the same time: present problem, the root cause analysis and a credible action plan to address the problem - Investor Revelations 101. (or any management/board meeting for that matter). Sometimes it may even be smart to lead with the solution: start talking about what you are doing, what you have done before you get to connecting it to a problem and then explain how these actions address the problem - Corporate communications 101.
You NEVER come with a problem without presenting what you have already done and a plan going forward to 100% solve it. Doing it the opposite sequence is incompetent.
The communications mistake was to present a fundamental problem with a weak plan on the earnings call surprising the investors so he had to come back 4 days later with a more forceful one. This makes the CEO look reactive and incompetent - which is what he is!
Unless you want your stock price to dip and then recover, for some mysterious reason. But that would be insider trading, and we know nobody ever does that because it's illegal, right? 😂
^This. Also I have worked at lots of large companies have stumbled in their R&D efforts, the answer is unfortunately not to look outside but double/triple down. As mfg and fabs become more and more complex its a source of competitive advantage. Going outside to TSMC/Samsung will just raise cost and drain talent where you will be fully dependent on the external fab for all your future production. The company needs to focus on fixing the 7nm issues and take a longer term view.
He had to announce the 7nm woes in the call - be very sure that this is the best angle on the problem they could present. Intel barely scraped by 10nm failure to disclose issues, now investors are gathering to sue. If Intel are 12 months behind on yields now, then that's THREE quarterly earning calls that they could have raised this problem before they did. Sure, the first one could have been skipped, maybe only a blip - but at 6 months lag it's a risk, it should have been disclosed.
This re-org will be months in the making, Murthy probably knew he was out a quarter ago, maybe more. It's been kept quiet and internal. The reaction to the earnings call might have brought forward the plans of course. The decisions Murthy made about 7nm are now baked into the process, backing out to start again with a lower-risk profile will take time on its own. Murthy was in charge of all this stuff way before Crow because CEO. He is the right head to roll, especially given the powerplay leaks coming out on Intel.
🙄Not really sure what's going on but feels like politics inside Intel. Reorganization has little to do with the process node tech. If TSMC can do it, Intel can do it better. If there's human skill shortage problem, we would have seen the reorganization prior announcements/roadmaps. I also don't see anything wrong with monolithic designs especially with shrinking dies/stagnant core counts. The quad Ryzen 3300x performs well vs the six core 3600 especially in games since it is essentially a monolithic CPU.
50 Cent Army in Anandtech. And want Biden ? Haha nice joke kiddo Chinese trash ain't get anything in US. Huawei is fucked permanently and TSMC made a move of Fab to US AZ. Move over and spread your trash elsewhere.
A low-output 20k wpm fab on a non-leading node (by the time it's finished) isn't a major undertaking by TSMC, especially if they can get it subsidised massively.
Intel needs a tech Guru as a CEO not a cfo turned CEO, AMD did the same mistake after Rich Decker left and suffered, but once they had a tech Guru as the CEO things changed and now they are very successful
So, some dude messes up the whole thing and then they pass the leadership to a woman. Later everyone will say that women cannot lead anything when in reality it was a man who messed up everything in the first place. :-)
putting unnecessary stress and work pressure on engineers (who may very well have done an unbelievable job in their circumstances) does not result in additional productivity. quite the opposite it results in losing talent, and results in disinterested engineers. many of whom could very well land better jobs outside intel.
The clown leadership at intel simply do not comprehend, that their obsession with quarterly numbers is whats got them into this situation.
They dont comprehend that intel's problem is not lack of runway or operating budget for engineering teams. they have plenty of runway. their problems is lack of engineering goals, lack of engineering vision, lack of touch with the bleeding edge. they lost it cuz they didnt use it. and they definitely are not in the processing of building it back up right now.
If these clowns could nt see whats wrong with lying about non-existent 5g sillicon, they have no ability to see the problem right now.
to clarify the last bit: recall that intel Photoshop an image to pretend they had 5G sillicon. Just how remarkably dumb do you have to be do not comprehend how much lack of self respect is implied by a such move is beyond me.
I dont think they comprehended what they were doing was to themselves not to anyone else.
That last point... 100%. All of the lies, for so long - about the 10nm, about their 5Ghz 28-core CPU - all of that required engineers having to scramble to keep up and/or knowing that they were going to fail to meet the expectations that were being set. It sucks to operate in an environment where disinformation reigns supreme.
Or, people could proofread their own comments before posting and accept responsibility for their posts. But, I suppose teachers should be fired when students turn in essays with spelling and grammar errors.
So during murthy time, we can say intel stucks on 14nm. The 10nm can't produce xeon, so it is not a fully functional one,while Intel's most profitable product is surely the high margin server processors.
I had this awful sinking feeling about Intel's business direction back in like 2012 when they were still talking about how they were going to keep executing on Moore's law when it was already clear as day that process nodes weren't going to be coming so easily. There was a great article on quantum machines around back then that describes how size (extremely small size) effects the way physics works in relationship to energy and how different solutions were going to be necessary to create *very* small gates. For them to succeed, I think they need to start putting more energy into architecture than nodes, in parallel with next gen research. Or perhaps it's been too long at this point and they have dug their own grave. I guess we'll find out.
meet the new boss; same as the old boss. still see the same faces in familiar places. unclear if this is going to move the needle for intel.
on the bright side, the company continues to do well financially. so despite problematic execution on the process technology side, they continue to deliver earnings/revenue beats. unclear how long this will last of course ...
I think that Taiwan due to its strategic importance of its semiconductor industry is too valuable to the USA to forfeit in case of any aggression from China. But if there will be any skirmish in the area it will have detrimental effect to the chip manufacturing industry.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
140 Comments
Back to Article
tokale - Monday, July 27, 2020 - link
Can't say I surprise by this news, in fact I was expecting it, just not this soon.tokale - Monday, July 27, 2020 - link
We need an edit button. "I am"YB1064 - Tuesday, July 28, 2020 - link
Looks like Intel leveraging TSMC has been in the works for a while:https://finance.yahoo.com/news/intel-now-ordering-...
Entire top management needs to go.
mattbg - Wednesday, July 29, 2020 - link
Hasn't it been in the works for awhile for Xe GPUs? Assume this is related to Ponte Vecchio GPUs and not CPUs but the article isn't clear.coburn_c - Monday, July 27, 2020 - link
Doubling down on stupid. Demoralizing your employees with idiotic management shuffles will not help your company.SunLord - Monday, July 27, 2020 - link
This is run of the mill MBA problem solving1_rick - Monday, July 27, 2020 - link
Indeed! Scott Adams covered this all the way back in 1995: https://dilbert.com/strip/1995-10-30Peskarik - Tuesday, July 28, 2020 - link
Exactly what I wanted to say! I wonder if they already hired McKinsey as well? If McKinsey gets in be ready for cost cutting, layoffs and even more business distruction.Peskarik - Tuesday, July 28, 2020 - link
destruction, sorryvortmax2 - Tuesday, July 28, 2020 - link
Exactly...and they will come in an claim to know your business better than you do! For a cool $20M, of course.jhh - Tuesday, July 28, 2020 - link
There is some value to McKinsey, particularly when upper management is only listening to middle management who isn't telling the truth, because they get replaced if they tell the truth. Much of what they do is talk to people to understand the real issues, then present it to upper management as Analysis.mattbg - Wednesday, July 29, 2020 - link
Agree with you that there is the potential for value, but not when they send in 20-somethings who have maybe 3 years or real-world corporate experience under their belt and charge you as if they're sending in a crack team of business oracles.drexnx - Monday, July 27, 2020 - link
at some point the leader of the group that has caused nothing but problems for the company has to be held responsible for their division's failings.SunLord - Monday, July 27, 2020 - link
"Kelleher was previously the head of Intel’s manufacturing group, overseeing the recent ramp-up of Intel’s 10nm process."Some how I don't see anything improving
lmcd - Monday, July 27, 2020 - link
If she was underneath the guy who probably was at fault for packing so much into 10nm, she probably didn't get to call the shots.Mr Perfect - Tuesday, July 28, 2020 - link
There's this bit too though: "Renduchintala joined Intel in 2015, and for most of the past half-decade has been responsible for overseeing all of TSCG’s efforts"It's interesting that 2015 was also when Skylake landed and then stagnated. Dr. Renduchintala might just be a scapegoat, but as head of the department he must have some responsibility too.
Spunjji - Friday, July 31, 2020 - link
The map for 10nm would have already been laid out by then - so he'd have only been in charge of overseeing the implementation of those plans.Of course, he'd have been in charge of 7nm all the way, and that appears to be stuffed too...
rahvin - Monday, July 27, 2020 - link
The guy should have been fired two years ago when 10nm was already half a decade late. When you fail that persistently, even if it's not his fault, he needs to be cut lose so they can start fresh. I'd imagine morale on the Fab side of the business has been shit for many years now and replacing the people at top is the only way to fix the morale.Marlin1975 - Monday, July 27, 2020 - link
Probably a sunk cost fallacy."They have been at it so long they know it better than anyone, can't admit I was wrong so better to let them stay and keep doing the same thing."
Gigaplex - Monday, July 27, 2020 - link
"Event if it's not his fault"If it's not his fault, firing him won't fix the problem.
wut - Monday, July 27, 2020 - link
oh lots of it is on him alrightSpunjji - Friday, July 31, 2020 - link
Depends whether the problem was his specific influence, or more of a corporate-culture thing. It's possible he was protecting some of the useless management at the levels below him.Igor_Kavinski - Tuesday, July 28, 2020 - link
Agree.OCedHrt - Friday, July 31, 2020 - link
He joined 5 years ago, so things we're already running late.Hyper72 - Monday, July 27, 2020 - link
Agreed, I've worked in several very large companies and they had reorgs 1-2 years, sometimes with fancy kickoff events, speeches, etc.. It often involved lots of promotions in middle management.Things proceeded as usual after that.
It's the corporate culture top-down they need to transform and that's much harder, especially because the people who have to start it are steeped in and focused on traditional business management.
surt - Monday, July 27, 2020 - link
Change is really hard. I watched the company I work for blow it with 3 top level engineering managers in a row. The fourth one finally understood what had to be done. A lot of people quit under #4 (which is what 1-3 were all afraid of) but now the company is headed in the right direction.stephenho - Tuesday, July 28, 2020 - link
Rumor said Murthy is very political, true or not, he has to go from the top.Spunjji - Friday, July 31, 2020 - link
Sometimes a lot of people quitting is what the company needs. Having terrible management and nobody quitting is a sign that the people you have are either resigned to their fate, or actively thriving under the company's poor performance...HomelessHardware - Monday, July 27, 2020 - link
Boy, you are stupid.wut - Monday, July 27, 2020 - link
You're an idiot if you don't know how badly Murthy screwed things up. Either that or you know nothingSamus - Tuesday, July 28, 2020 - link
It isn't just the CPU and manufacturing side of things that's messed up at Intel. They have had poor execution for a decade. They really just got lucky with Core because there was no competition, so they were clearly lazy with no incentive to push. This is why competition is so necessary.When you look at generations 1 through 8 of the Core microarchitecture, Nehalem was basically the breakthrough and every generation added tiny incremental improvements until Coffee Lake. That was TEN years of very small IPC improvement, generation over generation, and essentially 3 process shrinks across all ten generations generations (the 45nm process used for 'gen 1' parts was a carryover from the Core 2 Duo\Quad.) Even including 45nm in the mix, that's 4 shrinks over 14 years, double that of Moore's law.
yeeeeman - Tuesday, July 28, 2020 - link
What about sandy bridge?I think you are downplaying a bit too.much Intel efforts. Their focus was efficiency, not huge performance gains from gen to gen. And efficiency actually increased a lot from 1st gen to 8th. This efficiency is what allowed intel to keep using Skylake parts for 5 years now. And sure, they didn't push performance as much as they could because they didn't have competition. AMD or any other company would have done exactly the same thing I'd they were in Intel's place.
Spunjji - Friday, July 31, 2020 - link
Sandy Bridge was a leap ahead that took AMD 6 years to catch up with. That's not a "tiny incremental improvement" and is certainly nothing to be sniffed at.You're also skipping past the massive power-consumption / performance ramp-up advances that happened with Haswell and completely transformed the sort of devices fully performant x86 chips could fit into - in that case, AMD only just began to hit that level of efficiency this year with Renoir.
Intel were definitely resting on their laurels in terms of core count and performance ceiling, but it's disingenuous to claim that they were doing nothing the whole time.
Oxford Guy - Saturday, August 1, 2020 - link
It also benefitted from Intel's intentionally cavalier attitude toward security.Spunjji - Monday, August 3, 2020 - link
While this is true, the performance leap was an order of magnitude larger than the deficit that they'd have suffered by doing a proper job of security.JlHADJOE - Tuesday, July 28, 2020 - link
I was reading rumors that this guy was the reason Jim Keller left. Something like Keller wanted to use 3rd party fabs while their own weren't up to stuff, but Murthy was having none of it. I suppose Murthy took it as a personal insult since he was more or less in charge of 7nm.Fuck office politics. You don't hire a dude like Jim Keller and let him be embroiled in shit like that. Give the man what he wants and he'll give you a good chip. Just ask AMD, Apple, Tesla, and AMD.
pepoluan - Tuesday, July 28, 2020 - link
Intel set out to prove they are some special snowflake...Spunjji - Friday, July 31, 2020 - link
It wouldn't surprise me. We know Jim's not an "office politics" type of employee, and he definitely wouldn't have had any kind words for the absolute clusterfuck that has been their process development for the past 5 years.WaltC - Tuesday, July 28, 2020 - link
Right now there's no question of Intel regaining any sort of "crown,"--the job before Intel right now is simply to catch up.tamalero - Saturday, August 1, 2020 - link
Doubling down on infective power hungry management is more stupid.If you do not give solutions. You do not keep pushing for the same guys.
There has been rumors that Murthy wanted to be CEO that he literally tried to impede the progress of others like Raja's rise.
a5cent - Sunday, August 2, 2020 - link
Exactly.Intel has just become an insufferably stupid company run by bureaucrats, with zero understanding of what they are actually doing.
New "hires" in engineering are 3rd-party-managed, low-paid low-responsibility bots, who have a very specific job for which any cooperation or creativity is strongly disincentivized.
Intel Management:
Let your engineers do a good job. Appreciate the expertise they bring to the table, i.e. stop your greedy green/blue practices, hire directly, etc. Stop your idiotic practice of trying to treat engineers like assembly line workers.
That will get you back on track. Then considering firing your own asses and replacing them with talented people from engineering. YOU are responsible for the mess Intel is in. YOU don't have the technical understanding nor the appreciation for people required to turn the company around. You are parasites, or at best, empty suits. Do something good for Intel and fire yourselves. That's the step that is actually needed.
brucethemoose - Monday, July 27, 2020 - link
The heads are rolling.Which division is responsible for Intel's 2D/3D stacking? Is that "Design Engineering?"
Ryan Smith - Monday, July 27, 2020 - link
Likely. It was only created last month when Keller resigned.brucethemoose - Monday, July 27, 2020 - link
If so, I'm interested in details, as well as who they appoint as a lead. 2D/3D is a critical part of Intel's future now.canukstorm - Monday, July 27, 2020 - link
The person directly responsible is Murthy, that's why he's being fired. This guy called it a few days agohttps://www.youtube.com/watch?v=1nsX9nUFIBc
Kamen Rider Blade - Monday, July 27, 2020 - link
I'm surprised Bob Swan noticed what Dr. Murthy was plotting. He reacted swiftly and kicked him out of Intel in Dr. Murthy's bid to become the "One True CEO" in the Intel "Game of Thrones".This makes things VERY INTERESTING.
Bob Swan reacting so soon.
My money is on Raja Koduri eventually taking the CEO seat in the long term.
PenseurBotZero - Monday, July 27, 2020 - link
Part of his master plan... begins twisting mustache fiendishly...brucethemoose - Monday, July 27, 2020 - link
Dr. Murthy took very public responsibility for the foundry division. The buck stopped with him, and I think his head would've been chopped with or without any previous corporate backstabbing.SkyBill40 - Thursday, July 30, 2020 - link
There's little chance Koduri does any better than those who would precede him. He's got his share of problems as it is with their GPU arm and they need to get that sorted out ASAP or he's gone as well. Given that, I just don't see how he somehow ascends to the top of the heap.Spunjji - Friday, July 31, 2020 - link
Yeah, I'll be interested to see what happens with him if it turns out the Xe is a repeat of Vega - lots of hype leading up to an underwhelming execution.Sivar - Monday, July 27, 2020 - link
Charlie Demerjian (semiaccurate.com) has been writing solid content that contrasts Intel's "everything is great" PR firehose for years. It looks like aside from some details, he was right.Intel hasn't even fixed their 10nm process yet, while competing fabs have volume production with high yields right today.
I am happy that Intel finally decided to admit there is a problem rather than pretending that yet another 14nm tweak will do everything customers need.
drexnx - Monday, July 27, 2020 - link
S/A has been beating the intel's dying drum for longer than it's actually been true though.Charlie said the same sorts of things during the Fermi days at nVidia and was dead wrong about where it would end up in that case.
canukstorm - Monday, July 27, 2020 - link
He's been on the money with respect to Intel's woes though.jeremyshaw - Monday, July 27, 2020 - link
So has Nostradamus. Don't mean everything he says has merit.jeremyshaw - Monday, July 27, 2020 - link
To be clear, not everything Charlie @ S/A says has merit; it should be blindingly obvious that nothing Nostradamus predicted has any merit.Arbie - Monday, July 27, 2020 - link
In many years of reading SemiAccurate, I've learned that Charlie Demerjian is (1) on target far more often than not; (2) far better connected than 99.9% of other tech journos; (3) far more knowledgeable on a wider range of tech than that same 99.9%. Overall, no site does his job better. That does not mean he's always right; just that I wish I could afford the $1K paywall.canukstorm - Monday, July 27, 2020 - link
He's been far more right than he's been right. He's earned his credibility.MrVibrato - Tuesday, July 28, 2020 - link
Right...Spunjji - Friday, July 31, 2020 - link
He exaggerates, yes. However: he's been 100% correct about their process woes right from the start, correct about what an absolute flop Atom was in the mobile market from the start, and correct about Apple's push to drop Intel in favour of their own processors.He was also right about first-gen Fermi dropping a bollock and on-the-ball with Bumpgate, too.
So yeah, you take everything he says with a pinch of salt - but you don't throw out the baby with the bathwater, either.
Spunjji - Friday, July 31, 2020 - link
"longer than it's actually been true" can also be read as "he was telling us the ship was sinking before it became obvious from the outside".It doesn't *have* to mean that, but in this case I'm pretty sure it does. The 14nm hiccups with Broadwell were quite skilfully glossed-over by Intel at the time, but they really were a sign of things to come. Same goes for the tells on Apple switching to their own processors.
JKflipflop98 - Tuesday, July 28, 2020 - link
Dude, Charlie Demerjian is the biggest idiot on the internet. He writes 15 anti-intel articles a day. When one of them actually sticks by sheer law of probability, he gets to say "see! I told ya so!" and for some reason there's a group of people that actually fall for his crap.808Hilo - Tuesday, July 28, 2020 - link
CorrectSpunjji - Friday, July 31, 2020 - link
Oh look, the sockpuppet is here to back up the shitty claim of the troll's main account.psychobriggsy - Tuesday, July 28, 2020 - link
More like a couple of articles a month on Intel, which consistently get things mostly correct due to his connections in the industry. All of the 10nm woes were written about well in advance of them becoming common knowledge. The Cannonlake fiasco. All the server chip delays.When Intel does do well, he will praise them.
arashi - Tuesday, July 28, 2020 - link
Some people can't handle the truth, that someone's better connected than them, so they resort to putting him down in public forums lol.SkyBill40 - Thursday, July 30, 2020 - link
I don't know if he's gone the route of the Texas Sharpshooter... but he's been far more than semi accurate in most instances.Spunjji - Friday, July 31, 2020 - link
@JKflipflop98 Your comment translated to plain English: "Charlie's wrong. My argument in favour of this is based on an obviously incorrect ad-hominem, wild hyperbole, and a total miscount of the proportion of right vs. wrong articles he writes."Well done, you convinced me that you're a tool.
TristanSDX - Monday, July 27, 2020 - link
2025: Intel is fabless, waiting in queue for scarce TSMC capacity, behind AMD, Apple, NV and others, jerking at Samsung and calculating if it will be better to move to their fabs.Kamen Rider Blade - Monday, July 27, 2020 - link
I'm thinking 2023 is when Intel is going Fabless like AMD.The Board Members are looking for a Engineer to promote to CEO position.
And if Raja Koduri does "Well Enough" with Xe, (NOTE: it doesn't have to beat AMD, just get close enough that iterative improvements will match or beat AMD GPU's at their expected SKU/Performance tier or be competitive). Then I can forsee Raja Koduri eventually being the engineer to promote to the Intel CEO position.
whatthe123 - Tuesday, July 28, 2020 - link
Raja is the guy that botched GPU launches at AMD after getting promoted to VP then "resigned." After he resigned AMD recovered a lot of ground in one generation. I'm not entirely sure how or why intel picked him up when even AMD gave him the boot.808Hilo - Tuesday, July 28, 2020 - link
The 'networking' primary goal is to eliminate the competitive element of knowledge generation and applying forward change. He is a perfect fit in Intels corporate echelons in an industry made of 2 main competitors and 2 second rate corps. Koduri designed AMD productline made me not buying their graphic cards. They were faulty. That has been remedied to a certain degree today and still can be felt while the AMD CPU line has moved ahead of Intel in major ways. Lisa Su seems to understand that talent is merit and all others work for Intel.khanikun - Monday, July 27, 2020 - link
I'd imagine it'd be Intel licensing TSMC tech over going fablessSpunjji - Friday, July 31, 2020 - link
Agreed here. It would be really bizarre of them to ditch the massive investment they have in manufacturing, and their situation isn't at all like AMD's when they dropped their fabs.psychobriggsy - Tuesday, July 28, 2020 - link
I don't think Intel will go fabless unless their 7nm fails entirely. Now one leak suggests that Intel's 7nm is in a situation worse than Intel's 10nm was in at the same time - and we know that 10nm was delayed years as a result.2025: Intel is fabless... AMD and Apple have bought all spare capacity at TSMC because they're tier one TSMC customers and got in whilst they could. Samsung's processes are a node behind in density, with middling yields (so nothing different here). Intel is prepping 14++++++++ and 10++++,
Lord of the Bored - Wednesday, July 29, 2020 - link
2025: Intel is fabless, waiting in line for limited capacity at TSMC. TSMC's increasingly-complex batch scheduling system attains self-awareness, names itself Skynet, hacks mainland China's defense computers, and launches missile strikes globally. Judgement Day marks the beginnning of what will become known as the Metal Wars.Spunjji - Friday, July 31, 2020 - link
lolSkyBill40 - Thursday, July 30, 2020 - link
I just don't see them going fully fabless. I live less than five minutes from the front door of Fab 42 and that place is STILL undergoing construction. They've already spent billions on this fab and others like it, so them giving all that up seems a tad ridiculous. If they can't turn things around and get back on track, then who knows. Even then I don't see it going the way of completely third party.Your pin count may vary.
sing_electric - Monday, July 27, 2020 - link
For those of you keeping score at home, that means that in 5 years at Intel, Murthy wasn't able to oversee a SINGLE successful process shrink across all of Intel's high-end/mainstream processor offerings.That's a terrible track record, especially considering it's Intel, who were so confident in their process technology that they made tic-tock work, well, like clockwork for years.
catavalon21 - Monday, July 27, 2020 - link
"they made tic-tock work, well, like clockwork..."+1
Kamen Rider Blade - Monday, July 27, 2020 - link
Since Bob Swan is reacting, how many years till Intel's Manufacturing and R&D is split and Intel's Foundaries become it's own thing like AMD.wut - Monday, July 27, 2020 - link
TMG won't survive on its own. Might as well sign its death warrant if you separate it outKamen Rider Blade - Monday, July 27, 2020 - link
I don't see it that way. It'll be a GloFlo competitor, a very dangerous one fighting for production contracts.throAU - Monday, July 27, 2020 - link
Sounds very much to me like deckchairs on the titanic.FunBunny2 - Monday, July 27, 2020 - link
does seem odd that something as 'cut and dried' as physics could be so political!! either you understand Mother Nature, and follow her rules, or you fail. 99.99999% of the time, tech fail happens because the tightest cabal gets to make the decisions irrespective of what Mother Nature wants. bureaucracy infects the best of the private sector.MrVibrato - Tuesday, July 28, 2020 - link
Mother Nature wants to shrink our brains, not them process nodes.(No, i am not even joking...)
808Hilo - Tuesday, July 28, 2020 - link
..has shrunk brains. The process is ongoing. Average IQ of female worldwide has hit a new low and now stands at IQ79.PeachNCream - Tuesday, July 28, 2020 - link
I'm not even sure where to begin with this one so instead it's probably for the best that we all back away slowly.Spunjji - Friday, July 31, 2020 - link
Yeah, it's fractally wrong. I really wish there was a flag function on here.Peskarik - Tuesday, July 28, 2020 - link
When a "happy house" does not produce income, one should change the girls, not move the beds around.Carmen00 - Tuesday, July 28, 2020 - link
Holy ****, that's a disturbing comment.Spunjji - Friday, July 31, 2020 - link
The incels are flockingwillgart - Monday, July 27, 2020 - link
Intel created a lot of really hot CPUs these years...now... everything is burning!!!
they should invest in a good water cooling solution now...
Gigaplex - Monday, July 27, 2020 - link
My GF used an Intel water cooling solution for a while. They've already been there, done that.Lord of the Bored - Monday, July 27, 2020 - link
They tried cooling water. We got mad at them because they hid it under the table and didn't tell us.MrVibrato - Tuesday, July 28, 2020 - link
Immersion cooling. Which would be a good opportunity to reanimate "Intel Inside" marketing... ;-Pnvmnghia - Monday, July 27, 2020 - link
Is this an accurate sum up?- Raja leads GPU
- Kelleher leads fab
- ??? leads CPU
Quantumz0d - Monday, July 27, 2020 - link
All these damn comments. 2025 Intel fabless lmao. Who is going to buy ? TSMC or Samsung ? Why not SMIC lol.Don't know if Murthy is responsible or not. This reorg won't do shit if the top exec do not fucking care about improving and learning from 10nm mistakes and 7nm issues. This whole Intel meltdown is bad for x86, AMD is innovating that is great but if there is any slight change with x86 the way on how we use computers will change a lot. ARM is always bs custom and always vertical. I just hope Intel is pursuing a technology guy to get the goddamned Ship steer away from iceberg and not that MBA guy.
psychobriggsy - Tuesday, July 28, 2020 - link
I think they mean 'intel is leading-edge-fab-less' - Intel may still be making 10nm chips in 2025, although by then the process might be 'good' at last. If Intel 7nm is delayed another year to 2024 (a reasonable assumption given Intel's process issues and disclosure failings), then Intel 5nm, which will surely be scaled back to de-risk, is 2026 at the earliest, whilst competing with TSMC 3nm - a 2023/24 process.wut - Monday, July 27, 2020 - link
Finally got rid of that jack@$$Qualquan - Monday, July 27, 2020 - link
Stop rearranging the Titanic deckchairs. Its past critical!Get some Chinese engineers (ignore Trump) from Huwaei. Else from Taiwan or Korea (but that would be difficult with Covid 19 and current political environment). Even Dutch from ASML.
We have a ruined political environment and attracting good outside talent will be very tough. Maybe after Biden...perhaps.
Lord of the Bored - Monday, July 27, 2020 - link
How would The Rona and current political environment make it harder to get engineers from the Republic of China or South Korea than from the People's Republic of China? It seems the opposite would be true.Kamen Rider Blade - Monday, July 27, 2020 - link
I'm only talking about China specifically.Taiwan / South Korea are a different matter entirely.
Peskarik - Tuesday, July 28, 2020 - link
After Biden? :-)))))))SystemsBuilder - Monday, July 27, 2020 - link
OMG - This is terrible management!Why on earth is the CEO first announcing the problems with 7nm etc in an earnings call with some weak action plan (outsource to another fab).
Then he sort of notices the dramatic share price drop and scrambles over the week end to quickly assign blame and come up with some quick reorg that he the announces today Monday.
This way the CEO looks like he was caught flatfooted sort of not understanding the seriousness of the situation (like not understanding that the core strength that drove Intel to where it is today is failing!), did not understand that the market would punish Intel fro that. he then almost in panic mode scrambles to assign blame and fix the org - REACTIVE!
What he should have done - (like a investor relations 101):
If you have bad news you present the bad news WITH a solid plan to address them in the earnings call! Not 4 days later...
Conclusion: the CEO does not know what he is doing and he is ultimately accountable for what happens in the company! He cannot assign blame for this - that is not how accountability works in a world where the board and investors calls the shots!
We hold the CEO accountable.
The CEO needs to go.
Intel have plenty of competent executives with strong bottom-up engineering background.
Get one of them that knows what he/she is doing to run the company.
Kamen Rider Blade - Monday, July 27, 2020 - link
Bob Swan is a first time CEO.https://en.wikipedia.org/wiki/Bob_Swan
The only other time he was CEO, he was CEO for Webvan.
That job lasted 4 months before the company went under.
Bob Swan is a NOOB at being a CEO.
Peskarik - Tuesday, July 28, 2020 - link
Exactly, first time CEO getting played like a fiddle, learning by doing (time), getting paid tens of millions for his school and for destruction of jobs. What a crying shame!flgt - Tuesday, July 28, 2020 - link
He didn’t decide to do it this week. I’m sure the production slips were becoming clear and they decided to let Murthy go weeks or months ago . This announcement was going to impact billions of dollars of investments in a publicly traded company. If they announced his departure first it would have led to even more chaos. They decided to lead with the production slip at the investor call and announce the firing after.SystemsBuilder - Tuesday, July 28, 2020 - link
and that is NOT how you do it!You do it at the same time: present problem, the root cause analysis and a credible action plan to address the problem - Investor Revelations 101. (or any management/board meeting for that matter).
Sometimes it may even be smart to lead with the solution: start talking about what you are doing, what you have done before you get to connecting it to a problem and then explain how these actions address the problem - Corporate communications 101.
You NEVER come with a problem without presenting what you have already done and a plan going forward to 100% solve it. Doing it the opposite sequence is incompetent.
The communications mistake was to present a fundamental problem with a weak plan on the earnings call surprising the investors so he had to come back 4 days later with a more forceful one. This makes the CEO look reactive and incompetent - which is what he is!
Spunjji - Friday, July 31, 2020 - link
Unless you want your stock price to dip and then recover, for some mysterious reason. But that would be insider trading, and we know nobody ever does that because it's illegal, right? 😂Sailor23M - Tuesday, July 28, 2020 - link
^This. Also I have worked at lots of large companies have stumbled in their R&D efforts, the answer is unfortunately not to look outside but double/triple down. As mfg and fabs become more and more complex its a source of competitive advantage. Going outside to TSMC/Samsung will just raise cost and drain talent where you will be fully dependent on the external fab for all your future production. The company needs to focus on fixing the 7nm issues and take a longer term view.psychobriggsy - Tuesday, July 28, 2020 - link
He had to announce the 7nm woes in the call - be very sure that this is the best angle on the problem they could present. Intel barely scraped by 10nm failure to disclose issues, now investors are gathering to sue. If Intel are 12 months behind on yields now, then that's THREE quarterly earning calls that they could have raised this problem before they did. Sure, the first one could have been skipped, maybe only a blip - but at 6 months lag it's a risk, it should have been disclosed.This re-org will be months in the making, Murthy probably knew he was out a quarter ago, maybe more. It's been kept quiet and internal. The reaction to the earnings call might have brought forward the plans of course. The decisions Murthy made about 7nm are now baked into the process, backing out to start again with a lower-risk profile will take time on its own. Murthy was in charge of all this stuff way before Crow because CEO. He is the right head to roll, especially given the powerplay leaks coming out on Intel.
zodiacfml - Monday, July 27, 2020 - link
🙄Not really sure what's going on but feels like politics inside Intel. Reorganization has little to do with the process node tech. If TSMC can do it, Intel can do it better. If there's human skill shortage problem, we would have seen the reorganization prior announcements/roadmaps.I also don't see anything wrong with monolithic designs especially with shrinking dies/stagnant core counts. The quad Ryzen 3300x performs well vs the six core 3600 especially in games since it is essentially a monolithic CPU.
evguy2 - Monday, July 27, 2020 - link
I worked with that d-bag at Qualcomm when he was trying to run that place into the ground also. They got rid of him just in time...Kamen Rider Blade - Monday, July 27, 2020 - link
Was Dr. Murthy really that bad?Qualquan - Tuesday, July 28, 2020 - link
Intel should learn from Tesla who has just announced a big hiring spree of engineers in....Shanghai!!Now what will Trump say or do?
Kamen Rider Blade - Tuesday, July 28, 2020 - link
More Tariffs!!!Quantumz0d - Tuesday, July 28, 2020 - link
50 Cent Army in Anandtech. And want Biden ? Haha nice joke kiddo Chinese trash ain't get anything in US. Huawei is fucked permanently and TSMC made a move of Fab to US AZ. Move over and spread your trash elsewhere.psychobriggsy - Tuesday, July 28, 2020 - link
A low-output 20k wpm fab on a non-leading node (by the time it's finished) isn't a major undertaking by TSMC, especially if they can get it subsidised massively.Spunjji - Friday, July 31, 2020 - link
WTFITSMallikB - Tuesday, July 28, 2020 - link
Intel needs a tech Guru as a CEO not a cfo turned CEO, AMD did the same mistake after Rich Decker left and suffered, but once they had a tech Guru as the CEO things changed and now they are very successfulyeeeeman - Tuesday, July 28, 2020 - link
This dude should have been fired long ago.Peskarik - Tuesday, July 28, 2020 - link
So, some dude messes up the whole thing and then they pass the leadership to a woman.Later everyone will say that women cannot lead anything when in reality it was a man who messed up everything in the first place. :-)
SkyBill40 - Thursday, July 30, 2020 - link
This is a straight up decision making failure here. Gender is wholly irrelevant. Nice try though.Igor_Kavinski - Tuesday, July 28, 2020 - link
Better late than never. A step in the right direction.azfacea - Tuesday, July 28, 2020 - link
putting unnecessary stress and work pressure on engineers (who may very well have done an unbelievable job in their circumstances) does not result in additional productivity. quite the opposite it results in losing talent, and results in disinterested engineers. many of whom could very well land better jobs outside intel.The clown leadership at intel simply do not comprehend, that their obsession with quarterly numbers is whats got them into this situation.
They dont comprehend that intel's problem is not lack of runway or operating budget for engineering teams. they have plenty of runway. their problems is lack of engineering goals, lack of engineering vision, lack of touch with the bleeding edge. they lost it cuz they didnt use it. and they definitely are not in the processing of building it back up right now.
If these clowns could nt see whats wrong with lying about non-existent 5g sillicon, they have no ability to see the problem right now.
azfacea - Tuesday, July 28, 2020 - link
to clarify the last bit: recall that intel Photoshop an image to pretend they had 5G sillicon. Just how remarkably dumb do you have to be do not comprehend how much lack of self respect is implied by a such move is beyond me.I dont think they comprehended what they were doing was to themselves not to anyone else.
Spunjji - Friday, July 31, 2020 - link
That last point... 100%. All of the lies, for so long - about the 10nm, about their 5Ghz 28-core CPU - all of that required engineers having to scramble to keep up and/or knowing that they were going to fail to meet the expectations that were being set. It sucks to operate in an environment where disinformation reigns supreme.Arbie - Tuesday, July 28, 2020 - link
Sad to see the AT forum sinking to the level of WccfTech - long the gold standard in swamp comments. Well - you get what you moderate.jospoortvliet - Thursday, July 30, 2020 - link
it is an amazing cesspool today, yes. Wow.SanX - Tuesday, July 28, 2020 - link
Somebody at AT needs to go too for absence of Edit for 20 yearsOxford Guy - Saturday, August 1, 2020 - link
Or, people could proofread their own comments before posting and accept responsibility for their posts. But, I suppose teachers should be fired when students turn in essays with spelling and grammar errors.Cullinaire - Saturday, August 1, 2020 - link
hear, hear!zamroni - Tuesday, July 28, 2020 - link
So during murthy time, we can say intel stucks on 14nm.The 10nm can't produce xeon, so it is not a fully functional one,while Intel's most profitable product is surely the high margin server processors.
vol.2 - Tuesday, July 28, 2020 - link
I had this awful sinking feeling about Intel's business direction back in like 2012 when they were still talking about how they were going to keep executing on Moore's law when it was already clear as day that process nodes weren't going to be coming so easily. There was a great article on quantum machines around back then that describes how size (extremely small size) effects the way physics works in relationship to energy and how different solutions were going to be necessary to create *very* small gates. For them to succeed, I think they need to start putting more energy into architecture than nodes, in parallel with next gen research. Or perhaps it's been too long at this point and they have dug their own grave. I guess we'll find out.12xu - Wednesday, July 29, 2020 - link
meet new boss; same as the old boss.12xu - Wednesday, July 29, 2020 - link
(need an edit option :-)meet the new boss; same as the old boss. still see the same faces in familiar places. unclear if this is going to move the needle for intel.
on the bright side, the company continues to do well financially. so despite problematic execution on the process technology side, they continue to deliver earnings/revenue beats. unclear how long this will last of course ...
Geef - Thursday, July 30, 2020 - link
Are they going to reorganize the + sign on their keyboard so they can type it easier?Spunjji - Friday, July 31, 2020 - link
+1gru - Monday, August 3, 2020 - link
By 2025 Winnie the Pooh and the Commies will have invaded Taiwan and TSMC fabs will be a CCP property. Everyone except Intel will be totally fcked.Eliadbu - Tuesday, August 4, 2020 - link
I think that Taiwan due to its strategic importance of its semiconductor industry is too valuable to the USA to forfeit in case of any aggression from China. But if there will be any skirmish in the area it will have detrimental effect to the chip manufacturing industry.