The Slow Demise of Intel

[Zack Morris]

DEC, IBM, HP, and now... Intel?

Intel, cutting 12,000 jobs, tries to bend the company without breaking it

A coworker and I were discussing this the other day. Some friends and ex-colleagues of his in Intel's fab research group (working on processing nodes 2 steps out) spoke of particularly deep cuts to that group. Intel is cutting about 11% of its workforce, which is staggering, all the more so given that the company is as profitable as it is (60% gross margin). However, Intel has been suffering from a collapsing PC industry in recent years and they totally missed the mobile boat. Samsung (Korea) and TSMC (Taiwan) are rapidly converging on Intel. The latest Skylake processors are built on a 14nm process and some say that Intel's competition may already be ahead in 10nm node progress. Each process shrink is taking longer and nobody is certain what comes next once single-digit nm transistor channel lengths are reached. Note that the spacing between two crystalline silicon atoms is ~0.2 nm, if memory serves.

Why is this important? In a nutshell: the worst case scenario is that the US loses its edge in a vitally strategic domain.

Semiconductor fabrication is arguably the pinnacle of human technical achievement. New fab facilities cost on the order of $10B (and future nodes will be significantly more costly). If it turns out that cuts to fab R&D are deeper than elsewhere, Intel may be signalling a willingness to abandon this costly effort. Samsung, TSMC, and their governments will be happy to pick up the tab to maintain their technical advantage in the field. Former Intel CEO Andy Grove famously warned of the consequences of nations losing their manufacturing edge in this 2010 article. It's not just about the current generation of technology, but about the know-how to innovate, prototype, and bring new technologies to market. Once the supply chain decays, it can't be snapped back into life at will.

No easy solution to this one. The market doesn't care about America's competitiveness -- it will be happy to hand that over to Asian technocracies. And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

[noddy]

noone saw the rise of smartphone/tablets properly nor the fact the ultra low power SOC's that they are built from would be 'good enough' for the vast magority of humanity... intel was fixated on moar moar power like a muscle car head during the 70's fuel crisis.

almost overnight, intel went from a business model of people building mix n match custom computers with each component being $100-500 to the entire system (sans storage) being $100, plus the requirement it ran all day off a battery.

ARM and its license allows proper competition in the hardware market, which intel never really did.

i wont lose any sleep, the last x86 in my house is my compiling machine and its days are numbered.

[Zack Morris]

Someone has to fab those competitive ARM designs and it won't be Intel. Intel gave the foundry business a go recently and it went nowhere. Those state-of-the-art American factories are too costly.

Intel still has the upper hand in the server space. Cloud computing isn't going away and probably won't be moving to ARM any time soon. But then again, servers are relatively untethered to CPU architecture and change can happen suddenly and swiftly. The likes of Facebook are pioneering the way here.

Ironically, Intel used to produce a line of ARM chips (StrongARM, bought from DEC and rebranded as XScale). Maybe that's where they'll end up. Just another fabless ARM vendor. Chips fabbed in China (after their takeover of Taiwan), and President Zuckerberg signing free trade agreements and law enforcement cooperation pacts with Beijing Oh, America. How stupid your business elites and politicians are.

[noddy]

NVidia is all in on doing ARM gear with killer mips/watts ratios, AMD is also playing in that space.

now that its about running costs and not about raw power the intel servers arent as far ahead as you might think

and yep, facebook, google, amazon, azure etc are all on virtual machines and most of the code is hardware agnostic (php,python,go,dotnet,java blah)

having the most powerful and best server machines didnt help SUN at all in a cost driven environment, this is exactly the path by which x86 kicked those original unix vendors.

[NapLajoieonSteroids]

NVidia is all in on doing ARM gear with killer mips/watts ratios, AMD is also playing in that space.

now that its about running costs and not about raw power the intel servers arent as far ahead as you might think

and yep, facebook, google, amazon, azure etc are all on virtual machines and most of the code is hardware agnostic (php,python,go,dotnet,java blah)

having the most powerful and best server machines didnt help SUN at all in a cost driven environment, this is exactly the path by which x86 kicked those original unix vendors.

beat me to it- noddy running at a faster processing rate than I, confirmed.

[Doc]

DEC, IBM, HP, and now... Intel?

Intel, cutting 12,000 jobs, tries to bend the company without breaking it

A coworker and I were discussing this the other day. Some friends and ex-colleagues of his in Intel's fab research group (working on processing nodes 2 steps out) spoke of particularly deep cuts to that group. Intel is cutting about 11% of its workforce, which is staggering, all the more so given that the company is as profitable as it is (60% gross margin). However, Intel has been suffering from a collapsing PC industry in recent years and they totally missed the mobile boat. Samsung (Korea) and TSMC (Taiwan) are rapidly converging on Intel. The latest Skylake processors are built on a 14nm process and some say that Intel's competition may already be ahead in 10nm node progress. Each process shrink is taking longer and nobody is certain what comes next once single-digit nm transistor channel lengths are reached. Note that the spacing between two crystalline silicon atoms is ~0.2 nm, if memory serves.

Why is this important? In a nutshell: the worst case scenario is that the US loses its edge in a vitally strategic domain.

Semiconductor fabrication is arguably the pinnacle of human technical achievement. New fab facilities cost on the order of $10B (and future nodes will be significantly more costly). If it turns out that cuts to fab R&D are deeper than elsewhere, Intel may be signalling a willingness to abandon this costly effort. Samsung, TSMC, and their governments will be happy to pick up the tab to maintain their technical advantage in the field. Former Intel CEO Andy Grove famously warned of the consequences of nations losing their manufacturing edge in this 2010 article. It's not just about the current generation of technology, but about the know-how to innovate, prototype, and bring new technologies to market. Once the supply chain decays, it can't be snapped back into life at will.

No easy solution to this one. The market doesn't care about America's competitiveness -- it will be happy to hand that over to Asian technocracies. And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

What can replace silicon?

http://www.computerworld.com/article/30 ... licon.html

[monster_gardener]

DEC, IBM, HP, and now... Intel?

Intel, cutting 12,000 jobs, tries to bend the company without breaking it

A coworker and I were discussing this the other day. Some friends and ex-colleagues of his in Intel's fab research group (working on processing nodes 2 steps out) spoke of particularly deep cuts to that group. Intel is cutting about 11% of its workforce, which is staggering, all the more so given that the company is as profitable as it is (60% gross margin). However, Intel has been suffering from a collapsing PC industry in recent years and they totally missed the mobile boat. Samsung (Korea) and TSMC (Taiwan) are rapidly converging on Intel. The latest Skylake processors are built on a 14nm process and some say that Intel's competition may already be ahead in 10nm node progress. Each process shrink is taking longer and nobody is certain what comes next once single-digit nm transistor channel lengths are reached. Note that the spacing between two crystalline silicon atoms is ~0.2 nm, if memory serves.

Why is this important? In a nutshell: the worst case scenario is that the US loses its edge in a vitally strategic domain.

Semiconductor fabrication is arguably the pinnacle of human technical achievement. New fab facilities cost on the order of $10B (and future nodes will be significantly more costly). If it turns out that cuts to fab R&D are deeper than elsewhere, Intel may be signalling a willingness to abandon this costly effort. Samsung, TSMC, and their governments will be happy to pick up the tab to maintain their technical advantage in the field. Former Intel CEO Andy Grove famously warned of the consequences of nations losing their manufacturing edge in this 2010 article. It's not just about the current generation of technology, but about the know-how to innovate, prototype, and bring new technologies to market. Once the supply chain decays, it can't be snapped back into life at will.

No easy solution to this one. The market doesn't care about America's competitiveness -- it will be happy to hand that over to Asian technocracies. And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

Thank You VERY Much for your post, Zack Morris.

Samsung, TSMC, and their governments will be happy to pick up the tab to maintain their technical advantage in the field.

versus

And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

You may be correct but...........

Curious why/if it works for foreign governments...........

Why not for US............

[Zack Morris]

And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

You may be correct but...........

Curious why/if it works for foreign governments...........

Why not for US............

Because Americans are more interested in grifting the government for personal gain than working for the collective benefit of the nation, as is more likely to be the case in East Asia, even among business elites and politicians.

[Zack Morris]

NVidia is all in on doing ARM gear with killer mips/watts ratios, AMD is also playing in that space.

Nvidia's Tegra evolved from an X86 clone. These days, there's not much of an architectural disadvantage to X86.

now that its about running costs and not about raw power the intel servers arent as far ahead as you might think

Raw power does indeed matter in the cloud space, where infrastructure is being pushed to the max. It's all about cost vs. benefit, of course, so performance/watt is important, but where performance is paramount, Xeon-based systems can scale to a degree that ARM hasn't yet demonstrated. Theoretically, one could imagine ARM catching up, but does it really make business sense for companies with expertise developing silicon for mobile phones and embedded devices to try competing in the server space? These are two very different problem domains.

Nvidia's GPU experience makes them especially well positioned going forward. I think we'll start seeing more custom silicon to handle machine learning tasks. If Intel is smart with its investments in configurable computing (Altera), it should be able to hold its own for at least some time. But raising a white flag over its fabs turns them into just another chip company in a brutally competitive landscape where software eats everything.

and yep, facebook, google, amazon, azure etc are all on virtual machines and most of the code is hardware agnostic (php,python,go,dotnet,java blah)

Facebook, Google, and Amazon have invested heavily in building highly customized machines, and are beginning to experiment with novel architectures. The use of VMs for server management is orthogonal to developing custom high-performance computing hardware for data-intensive machine learning.

having the most powerful and best server machines didnt help SUN at all in a cost driven environment, this is exactly the path by which x86 kicked those original unix vendors.

True but I thought the cost was in the actual cost of the hardware itself plus the service agreements and software for a niche architecture aimed at high performance computing. X86 is far more open. You don't buy Intel servers from Intel, after all

[Zack Morris]

What can replace silicon?

http://www.computerworld.com/article/30 ... licon.html

They've been talking about these since before I entered college. Carbon nanotubes are already being used in some applications, just not computing. I don't think they offer any size advantage. Graphene (an unrolled carbon nanotube, essentially) was promising for a while but it's very hard to reliably fabricate and I'm not sure how much of a win it would really be anymore. These are all things that you can bet Intel is researching but if they throw in the towel, it'll be bad news for the US.

[Typhoon]

What can replace silicon?

http://www.computerworld.com/article/30 ... licon.html

They've been talking about these since before I entered college. Carbon nanotubes are already being used in some applications, just not computing. I don't think they offer any size advantage. Graphene (an unrolled carbon nanotube, essentially) was promising for a while but it's very hard to reliably fabricate and I'm not sure how much of a win it would really be anymore. These are all things that you can bet Intel is researching but if they throw in the towel, it'll be bad news for the US.

The problem with graphene is that it has a zero bandgap between the valence band and the conduction band,
thus graphene is not a semiconductor. A great deal of research has gone into how one might create a finite semiconducting band gap in graphene, but I would, as yet, not hold my breath.

[Typhoon]

And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

You may be correct but...........

Curious why/if it works for foreign governments...........

Why not for US............

Because Americans are more interested in grifting the government for personal gain than working for the collective benefit of the nation, as is more likely to be the case in East Asia, even among business elites and politicians.

I don't think that it is that straightforward.
There is no lack of grifting of and by the governments of E Asia.

A case study may be enlightening.
Once upon a time there were three photographic film companies:
Kodak in N America, Agfa in Europe, and Fujifilm in Asia.

Kodak was a very innovative company, but management mostly missed out on the digital camera revolution.
Fujifilm did not miss out and made lots of money.
However, with the commoditization of digital cameras and then the arrival of smartphones,
it was clear that the days of digital cameras were numbered.

All three imaging companies realized this and invested heavily in the healthcare field esp medical imaging,
and all became major players in the industry.
However, Kodak continued to be mismanaged and management sold off the one profitable
part of the company, the medical imaging division, so as to boost quarterly profits, make
Wall Street happy, raise the stock price to exercise their options,
and pull the cords on their golden parachutes while Kodak spiraled into bankruptcy.
The Kodak workers were stiffed as is the custom.

Kodak medical imaging became a Canadian owned successful company called Carestream.

Agfa and Fujifilm continue to be viable companies as far as I know.

[Doc]

And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

You may be correct but...........

Curious why/if it works for foreign governments...........

Why not for US............

Because Americans are more interested in grifting the government for personal gain than working for the collective benefit of the nation, as is more likely to be the case in East Asia, even among business elites and politicians.

I don't think that it is that straightforward.
There is no lack of grifting of and by the governments of E Asia.

A case study may be enlightening.
Once upon a time there were three photographic film companies:
Kodak in N America, Agfa in Europe, and Fujifilm in Asia.

Kodak was a very innovative company, but management mostly missed out on the digital camera revolution.
Fujifilm did not miss out and made lots of money.
However, with the commoditization of digital cameras and then the arrival of smartphones,
it was clear that the days of digital cameras were numbered.

All three imaging companies realized this and invested heavily in the healthcare field esp medical imaging,
and all became major players in the industry.
However, Kodak continued to be mismanaged and management sold off the one profitable
part of the company, the medical imaging division, so as to boost quarterly profits, make
Wall Street happy, raise the stock price to exercise their options,
and pull the cords on their golden parachutes while Kodak spiraled into bankruptcy.
The Kodak workers were stiffed as is the custom.

Kodak medical imaging became a Canadian owned successful company called Carestream.

Agfa and Fujifilm continue to be viable companies as far as I know.

Kodak invented the digital camera but didn't pursue it because it competed with their film business. That is a story that has been repeated many times in the last decades. Companies that dominate a given tech market invent a new disrupt product for their industry and fail to develop it. Motorola once absolutely dominated cell phones when they were analog, but failed to switch over to digital because they thought analog was good enough.

Locally fujifilm to have a large facility. It has been closed for several years now.

[Typhoon]

And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

You may be correct but...........

Curious why/if it works for foreign governments...........

Why not for US............

Because Americans are more interested in grifting the government for personal gain than working for the collective benefit of the nation, as is more likely to be the case in East Asia, even among business elites and politicians.

I don't think that it is that straightforward.
There is no lack of grifting of and by the governments of E Asia.

A case study may be enlightening.
Once upon a time there were three photographic film companies:
Kodak in N America, Agfa in Europe, and Fujifilm in Asia.

Kodak was a very innovative company, but management mostly missed out on the digital camera revolution.
Fujifilm did not miss out and made lots of money.
However, with the commoditization of digital cameras and then the arrival of smartphones,
it was clear that the days of digital cameras were numbered.

All three imaging companies realized this and invested heavily in the healthcare field esp medical imaging,
and all became major players in the industry.
However, Kodak continued to be mismanaged and management sold off the one profitable
part of the company, the medical imaging division, so as to boost quarterly profits, make
Wall Street happy, raise the stock price to exercise their options,
and pull the cords on their golden parachutes while Kodak spiraled into bankruptcy.
The Kodak workers were stiffed as is the custom.

Kodak medical imaging became a Canadian owned successful company called Carestream.

Agfa and Fujifilm continue to be viable companies as far as I know.

Kodak invented the digital camera but didn't pursue it because it competed with their film business. That is a story that has been repeated many times in the last decades. Companies that dominate a given tech market invent a new disrupt product for their industry and fail to develop it. Motorola once absolutely dominated cell phones when they were analog, but failed to switch over to digital because they thought analog was good enough.

The failure of Japan to globally market their then advanced mobile phones is a famous failure.

Locally fujifilm to have a large facility. It has been closed for several years now.

Fujifilm does their software development in the US.

Software has long been the traditional Achilles heel of Japan Inc.

[Doc]

And there isn't much the US government can or should do to save an ailing 100,000+-employee giant.

You may be correct but...........

Curious why/if it works for foreign governments...........

Why not for US............

Because Americans are more interested in grifting the government for personal gain than working for the collective benefit of the nation, as is more likely to be the case in East Asia, even among business elites and politicians.

I don't think that it is that straightforward.
There is no lack of grifting of and by the governments of E Asia.

A case study may be enlightening.
Once upon a time there were three photographic film companies:
Kodak in N America, Agfa in Europe, and Fujifilm in Asia.

Kodak was a very innovative company, but management mostly missed out on the digital camera revolution.
Fujifilm did not miss out and made lots of money.
However, with the commoditization of digital cameras and then the arrival of smartphones,
it was clear that the days of digital cameras were numbered.

All three imaging companies realized this and invested heavily in the healthcare field esp medical imaging,
and all became major players in the industry.
However, Kodak continued to be mismanaged and management sold off the one profitable
part of the company, the medical imaging division, so as to boost quarterly profits, make
Wall Street happy, raise the stock price to exercise their options,
and pull the cords on their golden parachutes while Kodak spiraled into bankruptcy.
The Kodak workers were stiffed as is the custom.

Kodak medical imaging became a Canadian owned successful company called Carestream.

Agfa and Fujifilm continue to be viable companies as far as I know.

Kodak invented the digital camera but didn't pursue it because it competed with their film business. That is a story that has been repeated many times in the last decades. Companies that dominate a given tech market invent a new disrupt product for their industry and fail to develop it. Motorola once absolutely dominated cell phones when they were analog, but failed to switch over to digital because they thought analog was good enough.

The failure of Japan to globally market their then advanced mobile phones is a famous failure.

Locally fujifilm to have a large facility. It has been closed for several years now.

Fujifilm does their software development in the US.

Software has long been the traditional Achilles heel of Japan Inc.

IBM allowed Microsoft to push them out of their dominate position in the computer industry because they saw no value in software

awbPiUNaEUk