Wednesday, July 17th 2019

Intel's CEO Blames 10 nm Delay on being "Too Aggressive"

During Fortune's Brainstorm Tech conference in Aspen, Colorado, Intel's CEO Bob Swan took stage and talked about the company, about where Intel is now and where they are headed in the future and how the company plans to evolve. Particular focus was put on how Intel became "data centric" from "PC centric," and the struggles it encountered.

However, when asked about the demise of Moore's Law, Swan detailed the aggressiveness that they approached the challenge with. Instead of the regular two times improvement in transistor density every two years, Swan said that Intel has always targeted better and greater densities so that it would stay the leader in the business.
With 10 nm, Intel targets improved density by as much as 2.7x compared to the last generation of 14 nm transistors. He addressed the five year delay in delivering the 10 nm node being caused by "too aggressive innovation," adding that "... at a time it gets harder and harder, we set more aggressive goal..." and that's the main reason for the late delivery. Additionally he said that this time, Intel will stay at exactly 2x density improvements over two years with the company's 7 nm node, which is supposed to launch in two years and is already in development.

When talking about the future of Intel, Swan has noted that Intel's current market share is 30% of the "silicon market", saying that Intel is trying to diversify its current offerings from mainly CPUs and FPGAs to everything that requires big compute performance, in order to capture rest of the market. He noted that Artificial Intelligence is currently driving big demand for such performance, with autonomous vehicles expected to be a big source of revenue for Intel in the future. Through acquisitions like Mobileye, Intel plans to serve that market and increase the company's value.

You can listen to the talk here.
Add your own comment

111 Comments on Intel's CEO Blames 10 nm Delay on being "Too Aggressive"

#26
Kn0xxPT
For me ( tin foil/ ), this all Intel Delay, is just bullshit, If I remember correctly, Intel made an agreedment with AMD for GPU tech patents and x86-64 renewal licences.
Is not too hard to think that AMD made Intel ( as an agreedment) to delay the release of 10nm Desktop chips.
What Intel didn't predict, was AMD making a great job in marketing and pushing core count to home users at great prices.

For me, making comparisons with Intel 14nm and AMD 7nm, and claiming superior tech to AMD, sorry, but its not a real comparison. AMD gets the lead on this timeframe of consumer. but Intel will get the edge again. They are focusing on GPU tech now... that is AMD source redone with Intel knowleadge.
The security problems on Intel CPU yes, they need a fix. but even with that Perf/Watt Intel is way better on a 14nm, go figure at 10nm or 7nm where AMD is now, and Intel is 4 years ago.
( I own a R5 2600 )
Posted on Reply
#27
R0H1T
Kn0xxPT, post: 4082540, member: 172760"
What Intel didn't predict, was AMD making a great job in marketing and pushing core count to home users at great prices.
More cores sell, they always have since the first multi core processor from AMD to the most dominant ecosystem (Android) fueled by Octa core revolution, among other things :toast:
Posted on Reply
#28
Konceptz
Fouquin, post: 4082516, member: 157604"
16 months. i7-980X came out March 16, 2010.
Since I need to be specific, the first mainstream 6 physical core CPU, not HEDT or extreme edition CPU.
The first one being the i7-8700 released Oct 2017
The first ever i7 CPU, the i7-920 was released Nov 2008........

As I said, Intel has been #1 since Nov 2008, they had no reason to keep pushing the envelope until the Ryzen 3000s lit a fire under them.
Posted on Reply
#29
bug
Konceptz, post: 4082546, member: 59570"
Since I need to be specific, the first mainstream 6 physical core CPU, not HEDT or extreme edition CPU.
The first one being the i7-8700 released Oct 2017
The first ever i7 CPU, the i7-920 was released Nov 2008........

As I said, Intel has been #1 since Nov 2008, they had no reason to keep pushing the envelope until the Ryzen 3000s lit a fire under them.
Yes, that is how businesses work. They don't compete with themselves if they don't have to. In other news, water is reported to be wet.
Posted on Reply
#30
DeOdView
Mark Little, post: 4082484, member: 168714"
I'm sorry but I can't let Intel try to blame their massive mistakes on being too awesome. The reason for the huge 10 nm delay is the total lack of competition from its rivals. Here is Intel mainstream desktop progression over the last twelve years:

4 cores 65 nm Kentsfield
4 cores 45 nm Clarksfield/Bloomfield/Lynnfield
4 cores 32 nm Westmere/Sandy Bridge
4 cores 22 nm Ivy Bridge/Haswell
4 cores 14 nm Broadwell/Skylake/Kaby Lake

For some reason, they stopped innovating over that entire time on the core architecture but instead focused only on die shrinks in the total absence of any competition. I'm guessing that with each shrink they could make more volume for less money to sell more chips (up to a point of course). Its all about profit and in the absence of competition, companies choose greed over innovation. No Intel, your 10 nm delay is not because you were too awesome. It would have been awesome if you went from 4 cores at 65 nm to over 32 cores at 14 nm regardless of your competitors. That would have been impressive. No, instead you were just greedy. Sadly most companies are like this.
^ Young one, I see. :p You must have forgotten who's pushed for 4+ cores at the time! It's amazed me when people swore to their... that there is no need for "Moar Cores" back in the Pentium days! ... Oh waited... they stilled saying it! NVM!

Whole heart agreed with you. Have seen so many the all MIGHTY companies, down and disappeared in my life time... and will be sure many more to come.

C'est la Vie!

Cheer!
{@}
^(^

PS :
Hopefully now, Intel let the DEVs use more than 4 Cores! : \Evillaughs\
Posted on Reply
#31
Kn0xxPT
Honestly, Android/Twitch/Youtube "scenes" made the push it needed to Multi-core PC ecosystem starting to kick in on mainstream. Multi did never was justified for prices either for consumer needs at mainstream.

Intel, helded back yes, they got greedy yes ... AMD got the chance to finally get a grip on sales. Intel is still best of plug-n-play system. there are no major RAM configurations needed, no over-the-board cooling, super duper PSU.I bought R5 2600 with all the hardware "hand-picked" and know-how needed to make it stable, it was more simple buying a cheap Mobo, basic DDR4 sticks, simple air cooling and would have to worry about one single setup at BIOS level. I wouldnt recommend any AMD CPU for basic casual gamer. it doesn't worth it. i5-9400F for me was enough, because in 3 to 4 years we have PCIe 5, DDR5 tech all arround ..and that is the tech evolution will leap.
Posted on Reply
#32
DeOdView
Kn0xxPT, post: 4082566, member: 172760"
Honestly, Android/Twitch/Youtube "scenes" made the push it needed to Multi-core PC ecosystem starting to kick in on mainstream. Multi did never was justified for prices either for consumer needs at mainstream.

Intel, helded back yes, they got greedy yes ... AMD got the chance to finally get a grip on sales. Intel is still best of plug-n-play system. there are no major RAM configurations needed, no over-the-board cooling, super duper PSU.I bought R5 2600 with all the hardware "hand-picked" and know-how needed to make it stable, it was more simple buying a cheap Mobo, basic DDR4 sticks, simple air cooling and would have to worry about one single setup at BIOS level. I wouldnt recommend any AMD CPU for basic casual gamer. it doesn't worth it. i5-9400F for me was enough, because in 3 to 4 years we have PCIe 5, DDR5 tech all arround ..and that is the tech evolution will leap.
I told my wifie that's I wanted to upgrade the Dell's dual core (63xx series) in my family room hooked up with the 60" TV...
.... And she said... Why?....
I just shakes my head and said nothing....

Just sayinnggg!

"AMD is Too Aggressive"

Fixed that for you
Posted on Reply
#33
Wavetrex
Mark Little, post: 4082484, member: 168714"
4 cores 65 nm Kentsfield
4 cores 45 nm Clarksfield/Bloomfield/Lynnfield
4 cores 32 nm Westmere/Sandy Bridge
4 cores 22 nm Ivy Bridge/Haswell
4 cores 14 nm Broadwell/Skylake/Kaby Lake
THIS is why I'm switching all my computing to AMD side. Not buying anything Intel anymore (well, except for their LAN chips on AMD mobos... lol)

I actually went through some of this... kentsfield, lynnfield, broadwell ... "Oh f***, ANOTHER quad core !??" Well, I'll buy it, no choice.

R7 3700x incoming tomorrow. Bye Intel.
Posted on Reply
#34
Darmok N Jalad
Either AMD got really lucky, or they nailed their product strategy really well with Ryzen. They knew they were lagging in IPC and process, but by making core count a talking point, they squeezed Intel on a place that Intel has to be very good—die sizes. Intel had to counter with more cores, but there has been no die shrink to save them. It has bought AMD time to catch up on IPC. I don’t think AMD expected such production woes at Intel. It really couldn’t have played out any better for them. I think they know this and have doubled down on cores—they will have a 16C/32T chip in September that will really cut into Intel’s highest margin chips. AMD just needs to get mobile Zen2 going.
Posted on Reply
#35
Xx Tek Tip xX
This is what happens when targets are set too high, whatever they bring out will still outperform zen 3000, they just have to pull off the value side of things for people, or further push single thread.
Posted on Reply
#36
HD64G
Typical tactics: Apologising with plenty of cheap excuses just to cover their bases for the incoming lawsuits of investors/shareholders. Total failure that is. 7nm intel is bound to get to the market in full suing -if ever- after 2021. Until then... :slap: :nutkick:
Posted on Reply
#37
Xx Tek Tip xX
HD64G, post: 4082647, member: 95052"
Typical tactics: Apologising with plenty of cheap excuses just to cover their bases for the incoming lawsuits of investors/shareholders. Total failure that is. 7nm intel is bound to get to the market in full suing -if ever- after 2021. Until then... :slap: :nutkick:
Intel's 10nm node is only delayed due to the fact it was badly underestimated, the targets were set too high causing it to be off track, calling it "excuses" is false.
Only the desktop market has been held behind, mobile is already secured and the server market should be their upmost priority.
Posted on Reply
#38
Eliad Buchnik
Mark Little, post: 4082484, member: 168714"
I'm sorry but I can't let Intel try to blame their massive mistakes on being too awesome. The reason for the huge 10 nm delay is the total lack of competition from its rivals. Here is Intel mainstream desktop progression over the last twelve years:

4 cores 65 nm Kentsfield
4 cores 45 nm Clarksfield/Bloomfield/Lynnfield
4 cores 32 nm Westmere/Sandy Bridge
4 cores 22 nm Ivy Bridge/Haswell
4 cores 14 nm Broadwell/Skylake/Kaby Lake

For some reason, they stopped innovating over that entire time on the core architecture but instead focused only on die shrinks in the total absence of any competition. I'm guessing that with each shrink they could make more volume for less money to sell more chips (up to a point of course). Its all about profit and in the absence of competition, companies choose greed over innovation. No Intel, your 10 nm delay is not because you were too awesome. It would have been awesome if you went from 4 cores at 65 nm to over 32 cores at 14 nm regardless of your competitors. That would have been impressive. No, instead you were just greedy. Sadly most companies are like this.
No one said they try to be awesome.
But your comment show lack of knowledge in semiconductors industry. Going for high core might be beneficial in many cases but it has many drawback especially in power consumption price and other complexity, but again we are talking about manufacturing process and not end product and there is distinct separation between the two. Saying 10 nm delay is due lack of competition yet you clearly say they benefit financially when going to smaller node contradicts your claim. Intel spent ten of billions so far on 10nm either by R&D and building new fabs. Intel also suffered a lot from chip shortages due the delays in 10nm meaning they lose money by not selling more chips, so not moving to the new node after they invested so much only damages them financially so how lack of competitiveness can explain any decision to not move to the new node?
And speaking on technical levels you of course would not know how current use of DUV light source which has 193 nm requires them to use quadruple pattering to achieve a features in tens of nm size. And even after that the features are still far from how they meant to be, in addition to other limitations you perceive as nonexistent judging by your rhetoric and claims.
Posted on Reply
#39
Vayra86
Mark Little, post: 4082484, member: 168714"
I'm sorry but I can't let Intel try to blame their massive mistakes on being too awesome. The reason for the huge 10 nm delay is the total lack of competition from its rivals. Here is Intel mainstream desktop progression over the last twelve years:

4 cores 65 nm Kentsfield
4 cores 45 nm Clarksfield/Bloomfield/Lynnfield
4 cores 32 nm Westmere/Sandy Bridge
4 cores 22 nm Ivy Bridge/Haswell
4 cores 14 nm Broadwell/Skylake/Kaby Lake

For some reason, they stopped innovating over that entire time on the core architecture but instead focused only on die shrinks in the total absence of any competition. I'm guessing that with each shrink they could make more volume for less money to sell more chips (up to a point of course). Its all about profit and in the absence of competition, companies choose greed over innovation. No Intel, your 10 nm delay is not because you were too awesome. It would have been awesome if you went from 4 cores at 65 nm to over 32 cores at 14 nm regardless of your competitors. That would have been impressive. No, instead you were just greedy. Sadly most companies are like this.
We can complain about the 4 core rampage all day long but let's face it, nobody had a purpose for anything much more for most things. It just didn't pay off. Intel's focus on IPC was the right one, if you ask me. But they should really have pushed more cores from Broadwell onwards; they just simply had no reason to because there was no competition.

Xx Tek Tip xX, post: 4082651, member: 178884"
Intel's 10nm node is only delayed due to the fact it was badly underestimated, the targets were set too high causing it to be off track, calling it "excuses" is false.
Only the desktop market has been held behind, mobile is already secured and the server market should be their upmost priority.
I agree, this Intel story, for once, seems pretty credible. Its also quite something to admit this publicly, albeit with a positive spin.

But now imagine if they screw up 7nm with their 'realistic target' ;)
Posted on Reply
#40
AsRock
TPU addict
Mark Little, post: 4082484, member: 168714"
I'm sorry but I can't let Intel try to blame their massive mistakes on being too awesome. The reason for the huge 10 nm delay is the total lack of competition from its rivals. Here is Intel mainstream desktop progression over the last twelve years:

4 cores 65 nm Kentsfield
4 cores 45 nm Clarksfield/Bloomfield/Lynnfield
4 cores 32 nm Westmere/Sandy Bridge
4 cores 22 nm Ivy Bridge/Haswell
4 cores 14 nm Broadwell/Skylake/Kaby Lake

For some reason, they stopped innovating over that entire time on the core architecture but instead focused only on die shrinks in the total absence of any competition. I'm guessing that with each shrink they could make more volume for less money to sell more chips (up to a point of course). Its all about profit and in the absence of competition, companies choose greed over innovation. No Intel, your 10 nm delay is not because you were too awesome. It would have been awesome if you went from 4 cores at 65 nm to over 32 cores at 14 nm regardless of your competitors. That would have been impressive. No, instead you were just greedy. Sadly most companies are like this.
Maybe putting to much money in trying to screw AMD over and over. The shit just flipped and it's all their own fault.
Posted on Reply
#41
londiste
Do not confuse CPU architecture design with manufacturing process research.
When it comes to manufacturing process, Intel's 10nm was without a doubt very aggressive. They have eased up the specs by now.

There is a lot of things that went wrong with Intel's 10nm but metal pitch under 40nm (Intel wanted 36nm) seems to be one of the major causes - this necessitated SAQP because EUV was not ready (and it only starts to becomes ready for using in semiconductor mass production about now) and seems to be a major hurdle in couple other ways. 40nm being a soft limit for pre-EUV was theoretically shown a while ago and both TSMC and Samsung opting to stay at 40nm metal pitch until EUV is telling.
Posted on Reply
#42
redzo
Mark Little, post: 4082484, member: 168714"
I'm sorry but I can't let Intel try to blame their massive mistakes on being too awesome. The reason for the huge 10 nm delay is the total lack of competition from its rivals. Here is Intel mainstream desktop progression over the last twelve years:

4 cores 65 nm Kentsfield
4 cores 45 nm Clarksfield/Bloomfield/Lynnfield
4 cores 32 nm Westmere/Sandy Bridge
4 cores 22 nm Ivy Bridge/Haswell
4 cores 14 nm Broadwell/Skylake/Kaby Lake

For some reason, they stopped innovating over that entire time on the core architecture but instead focused only on die shrinks in the total absence of any competition. I'm guessing that with each shrink they could make more volume for less money to sell more chips (up to a point of course). Its all about profit and in the absence of competition, companies choose greed over innovation. No Intel, your 10 nm delay is not because you were too awesome. It would have been awesome if you went from 4 cores at 65 nm to over 32 cores at 14 nm regardless of your competitors. That would have been impressive. No, instead you were just greedy. Sadly most companies are like this.
This.
Maybe Chipzilla will unleash another post Pentium 4 product. Who knows. The duopoly has got intresting once again. Good for everybody.
Posted on Reply
#43
dinmaster
load of crap, being on top for so long has staled the development, no sense of urgency and have turned into a company much run like a government... They did this to themselves along with miss management and now are paying the price. at least they have deep pockets to correct the ship hopefully before its too late.
Posted on Reply
#44
Manu_PT
Wavetrex, post: 4082632, member: 182738"
THIS is why I'm switching all my computing to AMD side. Not buying anything Intel anymore (well, except for their LAN chips on AMD mobos... lol)

I actually went through some of this... kentsfield, lynnfield, broadwell ... "Oh f***, ANOTHER quad core !??" Well, I'll buy it, no choice.

R7 3700x incoming tomorrow. Bye Intel.
So you are so worried about that, yet you go buy exactly one of the less valuable chips from a company line up?



dinmaster, post: 4082711, member: 31212"
load of crap, being on top for so long has staled the development, no sense of urgency and have turned into a company much run like a government... They did this to themselves along with miss management and now are paying the price. at least they have deep pockets to correct the ship hopefully before its too late.
Someone that knows nothing about tech/hardware and reads this post, might think Intel is in big trouble and that AMD totaly dominates Intel chips on everything.

I mean, Ryzen 3000 presents great value, R5 3600 and R9 3900x specifically. But let´s be honest, on a lot of tasks they still competing and going behind chips from 2015, at 14nm, while AMD is at 7nm.
Posted on Reply
#45
Wavetrex
Manu_PT, post: 4082720, member: 168799"
Yet you go buy exactly one of the less valuable chips from a company line up?
-video-
"Odd Man Out or Jack of All Trades ?"

I would say the latter.
8 cores is good today for most tasks and 90% of the reviews out there did point out that while the 12c is "faster", it's marginally, even in professional workloads.
Basically, only two mainline activities are faster on the 12c: 3D rendering and 4K video encoding.

For everything else, the extra 4 cores don't bring much (if anything).
On the other hand, the less 2 cores of 2600 and slightly lower frequencies do affect professional productivity quite a lot !

And a personal reason... I used a 6-core for the last 2.5 years (6800K) and going to another 6 core didn't make any sense at all... but doubling to 12 is overkill for all my activities.
So... I picked the "Odd man out".Or... more precisely, the Jack of all Trades, which is perfect for my day to day operations for the next 2-3 years.

If ever there's need for more, a 16c is incoming in a few months.
So ... in reality the 12c (3900X) is the real "odd man out" !
Posted on Reply
#46
Dave65
Yep 4 core with hyper threading is aggressive:shadedshu::shadedshu:
Posted on Reply
#47
dinmaster
Someone that knows nothing about tech/hardware and reads this post, might think Intel is in big trouble and that AMD totaly dominates Intel chips on everything.
that's right (would think that), if i wasn't talking about 10nm that intel was doing, while everyone else was going to 7nm. remember these companies are building 2-5+ years ahead, think about it like that. they are still a year away from having their next (7nm) processor ready if no delay. so from now until then their hands are somewhat tied to bringing out more cores on current gen/ bandaid (update cause of security issues) processors until the new architecture comes out. AMD had this as well, bulldozer came out and they couldn't do much until the next architecture (zen) was available which almost killed them... that's the approach i was taking. Things depend on how they develop, I use intel cpu's but i think amd is in a good position. the infinity fabric with chiplets allow them to easily add more cores to a processor which can help that period leading to next architecture. Something i want to see the gpu's take, have a look at the new mac pro dual gpu with infinity fabric between the gpu's, maybe next are chiplets on gpu's and could be a big move if they use it. adding these abilities can make refreshing processors huge when you need it and i think they learnt that from bulldozer... intel has perfected its single core performance because the lack of competition and them just sitting on what they had. I haven't seen much improvement from cpu to cpu over a few years now from intel but amd has done big changes from zen to zen+ to zen 2 now over a short time. We as consumers only look at the now, what is out now and amd is looking more favorable each iteration. when intel comes out with their 7nm, things could tip more to them. until that time a'lot can change and stocks will swing, that's why i made the deep pockets comment.
Posted on Reply
#48
lexluthermiester
bug, post: 4082495, member: 157434"
They didn't stop innovating, they shifted focus to the booming mobile market. A 4 core Skylake laptop will blow a 4 core Kentsfield laptop out of the water any day.
You must be talking about battery life because the computational performances are very similar.
Posted on Reply
#49
EarthDog
Wavetrex, post: 4082632, member: 182738"
I actually went through some of this... kentsfield, lynnfield, broadwell ... "Oh f***, ANOTHER quad core !??" Well, I'll buy it, no choice.
Hex cores have been around on the Intel HEDT platform for several years

...or you could choose one of AMD's notably slower hex cores (Bullldozer) which have also been around for several years + some.

You had a choice, you just chose not to do it for whatever reason...be it the higher cost or whatever... but, you had a choice.

Manu_PT, post: 4082720, member: 168799"
So you are so worried about that, yet you go buy exactly one of the less valuable chips from a company line up?
Looks like it. Nothing like spouting off about market stagnanation in core count............and then getting the same core count the other team has. Sure, its cheaper, but, that didn't seem like it was a talking point (until it becomes useful, like now).
Posted on Reply
#50
bug
lexluthermiester, post: 4082752, member: 134537"
You must be talking about battery life because the computational performances are very similar.
Yup, battery life, SpeedStep, better IGP (though not earth shattering). Everything but raw HP has improved.
Posted on Reply
Add your own comment