I just read an article written by Intel board members. "Professor Yoffie, Mr. Hundt, Ambassador Barshefsky, and Professor Plummer all served as longtime directors of the Intel board."
I worked for Intel Corp in the late 1980s until 1997. That was when Intel Corp ruled the CPU world- Intel 286, 386, Pentium. New wafer fabs were being built in NM and AZ. Intel Corp was/is highly dependent on corporate welfare. And there was a consortium of taxpayer funded photolithography research going on. It all ended when SVG corp was sold to ASML. I always wondered how a Dutch company could buy a US funded company?
The photolithography tools were the bottleneck because the geometry of the chips were limited by the frequency of light waves. Nanotechnology. In the 1990s, the limiting factor was the placement/position of the photolithograpy tool, but magnetics were used to resolve that resolution. Talk about a gnats ***.
Today, there is talk about TSMC taking over Intel chip factories. Intel Corp might die.
I worked for Intel Corp in the late 1980s until 1997. That was when Intel Corp ruled the CPU world- Intel 286, 386, Pentium. New wafer fabs were being built in NM and AZ. Intel Corp was/is highly dependent on corporate welfare. And there was a consortium of taxpayer funded photolithography research going on. It all ended when SVG corp was sold to ASML. I always wondered how a Dutch company could buy a US funded company?
The photolithography tools were the bottleneck because the geometry of the chips were limited by the frequency of light waves. Nanotechnology. In the 1990s, the limiting factor was the placement/position of the photolithograpy tool, but magnetics were used to resolve that resolution. Talk about a gnats ***.
Today, there is talk about TSMC taking over Intel chip factories. Intel Corp might die.
There are only two things. Architecture and manufacturing processes. I thought Intel just fell behind the manufacturing process before Arrow Lake. After they migrated Arrow Lake to TSMC, it didn’t perform any better than Ryzen which is one node behind. The intel problem is way bigger than its fab.
I worked for Nortel, in their IC fabs, in the '80s through early 2000, before shifting over to the CAD support, and finally the IC design side, with other semi companies. I worked with Perkin-Elmer (USA), Canon (Japan) and finally ASML (Dutch) photo gear over the years, and watched the photo-lithography market shrink along with the printed feature sizes. It still astounds me to think that the wave length of the UV light they're using is larger than the physical feature they're printing! Crazy...
From my perspective, one of Intel's failings was trying too hard to keep everything internal. For a long time, they didn't use industry standard chip design tools, which hindered their ability to work as a foundry. There was a certain level of arrogance there, We're Intel, makers of the best CPU's, it doesn't consume too much power, use a bigger heat sink! Slowly but surely, Samsung and TSMC, with their foundry buisness model, caught up with, and overtook Intel, learning from each customer what they needed most in a manufacturing and design process. It would be sad to see a company like Intel go under, but the US is hedging it's bets, by investing in TSMC. I'm out of the loop these days, but all of this makes me wonder where the US military is getting their secret stuff made these days. For a long time IBM didn't make any chips for their own use, but maintained a small fab in Vermont.....
From my perspective, one of Intel's failings was trying too hard to keep everything internal. For a long time, they didn't use industry standard chip design tools, which hindered their ability to work as a foundry. There was a certain level of arrogance there, We're Intel, makers of the best CPU's, it doesn't consume too much power, use a bigger heat sink! Slowly but surely, Samsung and TSMC, with their foundry buisness model, caught up with, and overtook Intel, learning from each customer what they needed most in a manufacturing and design process. It would be sad to see a company like Intel go under, but the US is hedging it's bets, by investing in TSMC. I'm out of the loop these days, but all of this makes me wonder where the US military is getting their secret stuff made these days. For a long time IBM didn't make any chips for their own use, but maintained a small fab in Vermont.....
I would challenge the statement that Intel ruled the CPU world during the 90’s. DEC had far more advanced processors and it was only through the theft of DEC IP that Intel managed to a bigger foothold in the CPU market. The court case, which DEC won, was marred by the appalling settlement that the DEC board agreed with Intel.
For the last 10 years of my career at Motorola I worked in a tech center that designed custom RF IC's for Motorola products. The only US foundry that could make the small geometry CMOS stuff we designed was IBM in Armonk NY. I was never an IC guy, so I don't remember the specific process number. Now that those chips are in production, they are fabbed at Global Foundrys and packaged at AMKOR, which comes from AMericaKORea.
I built PCs as far back as the mid 80's. I always used Intel chips even when AMD made equivalent products for less money, I refused to put one penny in Hector Ruis's pocket, since he refused to even listen to me explain a bad call that eventually took down the last of Motorola's paging market. He was a VP in the paging division at Motorola in the late 80's and early 90's. When Hector left AMD, I switched to AMD, and I am typing this today on a DIY Ryzen 9 machine.
I worked at Intel in Portland in 1980, the year Mount Saint Helens blew. Intel's fab "Aloha 3", in the suburbs of Portland, reacted to the volcanic dust cloud by vacuuming the parking lots, vacuuming the walkways, vacuuming the clothes and shoes of everybody who entered (through the double locked airtight entryways), and putting down sticky flypaper on the floor. They did this all day, every day, for several weeks. Surprisingly, the fab's yield went up! All these countermeasures resulted in a cleaner fab during a volcanic ash storm, than before the volcano.
What went wrong? Success, hubris, and (as it turned out) terrible replacements for Robert Noyce, Gordon Moore, and Andy Grove.
definition: yield = (# of good, defect free die) / (total # of die built)
What went wrong? Success, hubris, and (as it turned out) terrible replacements for Robert Noyce, Gordon Moore, and Andy Grove.
definition: yield = (# of good, defect free die) / (total # of die built)
There's probably hundreds of big reasons, but one of them was "not invented here" mentality and also they kept laying off older people that knew how the engineering and manufacturing "machines" ran. So in one way not able to change and in another forced to change for the wrong reasons with fresh, less experienced and cheaper workers. Those less experienced workers had to rebuild their tribal knowledge at the cost of even more delays. And of course that cycle even repeated a few times. Intel got caught for laying off the older workers, with their not so clever scheme of not handing out raises or promotions to certain grade levels and then use that as a general guidelines for the next year's layoffs.
Building fabs that can make chips at the latest process node costs billions. The last time I looked, a bleeding edge fab was >$20 billion, and it lasts only 7 yrs. But a year or two after it starts running, it is already falling behind. And the latest PC and mobile chips demand the very latest process nodes if they are to be competitive. It just makes more sense for Intel to focus on chip design, architecture, and the associated Intel ecosystem, than to build and staff fabs, which are bleeding cash. TSMC focus only on process and manufacturing and nothing else, and they fab for everyone so they have volume and associated scale benefits.
The central role the PC had for 35 yrs has had to give way to the mobile and associated apps and that's another big factor at play here. I do know Intel spent billions over the years to diversify their portfolio and get into other market segments but they were never able to do it successfully.
One of the defining characteristics of the semiconductor industry is the cost and selling price of products goes down as the technology improves. It has made life for consumers very good, but life for semi companies and their investors not that easy. My son works in the pharma industry and they seem to be able to extract more value from their investments over longer time periods.
The central role the PC had for 35 yrs has had to give way to the mobile and associated apps and that's another big factor at play here. I do know Intel spent billions over the years to diversify their portfolio and get into other market segments but they were never able to do it successfully.
One of the defining characteristics of the semiconductor industry is the cost and selling price of products goes down as the technology improves. It has made life for consumers very good, but life for semi companies and their investors not that easy. My son works in the pharma industry and they seem to be able to extract more value from their investments over longer time periods.
Last edited:
Remember the Itanium processor?
In the early 2000s, Intel knew they needed a 64 bit processor in order to stay in the PC and server CPU game.
They decided to build a new processor from the ground up that would not be x86 compatible. At the same time AMD added 64 bit extensions to the x86 processor and came up with a really good 64 bit upgrade for the x86 architecture. Cost low and performance was high. Intel stuck with Itanium. Itanium was never able to compete as far as price or speed. Microsoft also put too many eggs in the Itanium basket and fell behind in the development for the X86-64 processor. At the same time the Linux kernel developers were making great advances with the new AMD processor. By 2005 the AMD Opteron chip was really gaining in the server market. Eventually Intel caught up in the X86-64 processor game though they had lost alot of time and money.
Another sad thing that happened along the way was DEC cancelled the Alpha line and expected Itanium to be their CPU. Alpha was good.
Another side note and a reason for Intel to go in the trash can:
In the early 2000s my father had a suite of vibration alalysis and bearing flaw detection application developed jointly with some others. The company name was Inteletek. The name made more sense for the Russian developers.
Unfortunately the Intel lawyers got wind of the name and forced him to change his company name. My father went to court and had a good case. Unfortunately the judge had to say that you cannot win. They have more money than God. So I really hope Intel dies a slow painful death.
I do not think Intel has more money than God now.
In the early 2000s, Intel knew they needed a 64 bit processor in order to stay in the PC and server CPU game.
They decided to build a new processor from the ground up that would not be x86 compatible. At the same time AMD added 64 bit extensions to the x86 processor and came up with a really good 64 bit upgrade for the x86 architecture. Cost low and performance was high. Intel stuck with Itanium. Itanium was never able to compete as far as price or speed. Microsoft also put too many eggs in the Itanium basket and fell behind in the development for the X86-64 processor. At the same time the Linux kernel developers were making great advances with the new AMD processor. By 2005 the AMD Opteron chip was really gaining in the server market. Eventually Intel caught up in the X86-64 processor game though they had lost alot of time and money.
Another sad thing that happened along the way was DEC cancelled the Alpha line and expected Itanium to be their CPU. Alpha was good.
Another side note and a reason for Intel to go in the trash can:
In the early 2000s my father had a suite of vibration alalysis and bearing flaw detection application developed jointly with some others. The company name was Inteletek. The name made more sense for the Russian developers.
Unfortunately the Intel lawyers got wind of the name and forced him to change his company name. My father went to court and had a good case. Unfortunately the judge had to say that you cannot win. They have more money than God. So I really hope Intel dies a slow painful death.
I do not think Intel has more money than God now.
I think this is the consequence that salesperson replaced engineer running the company, it happened in Apple which ended up bring Jobs back, and recently the collapse of carmaker NISSAN, the list can go on and on.
The first ever meeting of the New York Society of Security Analysts I ever went to was sometime in 1974/5 when Intel presented. That was in the era that Raytheon was making memory chips.
Bloated company not afraid if they will bankrupt similar to GMC in auto industry.
Like the roman empire. Hubris will teach that one day.
Won't miss them. Any company with the attitude the world can not do without them earns dislike.
Won't miss them. Any company with the attitude the world can not do without them earns dislike.
Some mobile chip makers are entering the CPU space, and also making laptops powered by chips not designed by Intel or AMD.
Mediatek is an example, I think.
And some OS are flexible, away from the x86 instruction set that was so dominant all these years.
Part of the credit goes to the US Government, they cut off China from Android and ARM, so those guys said yeah, right...and went and built their own versions, using their own lithography and own chip designs.
Talk about shooting themselves in the foot....
And those phone chip guys are basically fabless, and used to wafer scale integration, the CPU and North / South bridges in same package.
And my cell phone has more cores than my desktop, and more memory.
I know people who run their businesses, and file tax returns, using cell phones and tablet computers. No Windows / Intel.....
People want things done, they are not thinking of chip maker, nor are they OS bound.
Let Intel go, or they will become a money well that can never be filled, like Chrysler and GM.
They are out of date at present, unless they catch up, game over.
The leader is now playing catch up, that is a very important thing.
Mediatek is an example, I think.
And some OS are flexible, away from the x86 instruction set that was so dominant all these years.
Part of the credit goes to the US Government, they cut off China from Android and ARM, so those guys said yeah, right...and went and built their own versions, using their own lithography and own chip designs.
Talk about shooting themselves in the foot....
And those phone chip guys are basically fabless, and used to wafer scale integration, the CPU and North / South bridges in same package.
And my cell phone has more cores than my desktop, and more memory.
I know people who run their businesses, and file tax returns, using cell phones and tablet computers. No Windows / Intel.....
People want things done, they are not thinking of chip maker, nor are they OS bound.
Let Intel go, or they will become a money well that can never be filled, like Chrysler and GM.
They are out of date at present, unless they catch up, game over.
The leader is now playing catch up, that is a very important thing.
Last edited:
Indeed... it's common to design the chip and tape it out to a fab.
Splitting the fab companies from the chip design companies makes a lot of sense nowadays, since most "chips" nowadays are done with software tools. You can do most of the design work with emulators ( like a Cadence Palladium ) and then tape it out and when the silicon comes back do the actual hardware testing....
Intel was strange in that it combined both, almost an anachronism.
Intel was integrated, not dependent on outsiders, and made money till Android came along, and before that Apple also took some market away from Windows.
Intel did some cutting edge stuff, and wanted to keep it secret...that is all.
Now Apple has moved away from Intel chips, and their OS is said to be better.
Buy some popcorn, if you wish, enjoy the action.
I outsource some of my molding work, and it is a headache, everybody wants to do it their way, not mine.
Keeping things in house gets better results, but $20 billion is a big amount...
Intel did some cutting edge stuff, and wanted to keep it secret...that is all.
Now Apple has moved away from Intel chips, and their OS is said to be better.
Buy some popcorn, if you wish, enjoy the action.
I outsource some of my molding work, and it is a headache, everybody wants to do it their way, not mine.
Keeping things in house gets better results, but $20 billion is a big amount...
All good things evening come to an end. The sun is clearly setting on Intel.
I have an iPad Pro and it’s a powerful beast. Add a keyboard and a mouse you have an 8 hour battery life PC level machine.
I prefer my Windows machine because there’s stuff on the PC that just works better, but the point is the PC (or Apple’s equivalent) is no longer preeminent the way it was 10 or 15 years ago. And given the way AI is going, will huge amounts of computing power on one’s desktop even be relevant? My wife spent an hour the other day calculating capital gains tax based on the new rules that will apply after April. This meant getting the details off the government website, running the numbers wrt other income etc. I did it using Gemini and it took about 3 minutes and we were within £20 on a big number. Big changes coming in the next 10 yrs - it will be back to terminals!
YMMV
I have an iPad Pro and it’s a powerful beast. Add a keyboard and a mouse you have an 8 hour battery life PC level machine.
I prefer my Windows machine because there’s stuff on the PC that just works better, but the point is the PC (or Apple’s equivalent) is no longer preeminent the way it was 10 or 15 years ago. And given the way AI is going, will huge amounts of computing power on one’s desktop even be relevant? My wife spent an hour the other day calculating capital gains tax based on the new rules that will apply after April. This meant getting the details off the government website, running the numbers wrt other income etc. I did it using Gemini and it took about 3 minutes and we were within £20 on a big number. Big changes coming in the next 10 yrs - it will be back to terminals!
YMMV
Just imagine what would happen if every ARM chip was turned off? It would be as big a disaster as turning every Intel chip off, maybe worse.
I asked Gemini for a 200 word summary of Intel’s business performance over the last 10 yrs:-
Intel's business performance over the past decade presents a complex picture, marked by both periods of dominance and increasing challenges. Here's a summary:
* Dominance and Challenges:
* For a significant portion of the last 10 years, Intel maintained its position as a leading semiconductor manufacturer. However, the company has faced increasing competition from rivals like AMD, NVIDIA, and others.
* Technological setbacks in chip manufacturing processes have contributed to a loss of market share in certain areas.
* The rise of specialized processors for artificial intelligence and other applications has also presented challenges to Intel's traditional CPU-centric business.
* Financial Performance:
* Financial data reveals fluctuations in Intel's revenue and stock performance. There have been periods of growth, but also significant declines in recent times.
* It is important to note that when looking at the 10 year total return of the stock, that it has been negative. Meaning that an investment made 10 years ago, would currently be worth less.
* Strategic Shifts:
* In response to these challenges, Intel has undergone strategic shifts, including investments in new manufacturing capabilities and a focus on expanding its foundry services.
* Intel is attempting to regain its footing in chip manufacturing, and is making very large investments to do so.
In essence, Intel's journey over the last 10 years reflects the dynamic and competitive nature of the semiconductor industry, with the company navigating technological disruptions and evolving market demands.
Intel's business performance over the past decade presents a complex picture, marked by both periods of dominance and increasing challenges. Here's a summary:
* Dominance and Challenges:
* For a significant portion of the last 10 years, Intel maintained its position as a leading semiconductor manufacturer. However, the company has faced increasing competition from rivals like AMD, NVIDIA, and others.
* Technological setbacks in chip manufacturing processes have contributed to a loss of market share in certain areas.
* The rise of specialized processors for artificial intelligence and other applications has also presented challenges to Intel's traditional CPU-centric business.
* Financial Performance:
* Financial data reveals fluctuations in Intel's revenue and stock performance. There have been periods of growth, but also significant declines in recent times.
* It is important to note that when looking at the 10 year total return of the stock, that it has been negative. Meaning that an investment made 10 years ago, would currently be worth less.
* Strategic Shifts:
* In response to these challenges, Intel has undergone strategic shifts, including investments in new manufacturing capabilities and a focus on expanding its foundry services.
* Intel is attempting to regain its footing in chip manufacturing, and is making very large investments to do so.
In essence, Intel's journey over the last 10 years reflects the dynamic and competitive nature of the semiconductor industry, with the company navigating technological disruptions and evolving market demands.