Perhaps the BEST B1G Forum anywhere, here at College Football Fan Site, CFB51!!!

I worked at AMD for five years and I can confirm this, with great and probably unnecessary detail.

AMD split into two companies about 15 years ago.  The design side of the house which is still AMD, and the wafer fab sites which are now called "Global Foundries."  They did this because the manufacturing side was sitting idle part of the time, at times when AMD didn't require as many wafer starts.  Idle time in a wafer fab is basically just lighting billions of dollars on fire, so they decided to split off that part of the company, so they could act as a foundry for many other semisconductor design firms.  Originally, AMD products were still supposed to get first priority, but I imagine now, the space goes to whomever pays the most for it.

Intel is the only major IC producer that has its own fabs.

Then there's companies like Samsung that produce some low-level semiconductors and memory using their own fabs, but they're not really doing complex ICs and microcontrollers.

I will take exception with some of what the dude is saying, if he's implying that there's no money to be made on manufacturing old chip designs.  You can think of the semiconductor business as being analogous to the pharmaceutical industry.  Huge initial investments in R&D, plus huge initial investments in setting up a new product manufacturing line in a wafer fab, result in profits being delayed heavily into the lifecycle of a product.  A CPU company like AMD might produce a version of an IC for 2-3 years, and only actually make profit on it, for the last 6 months.

Consequently, the idea that producing "old technology" generates little money, is false.  In fact it's quite profitable to harvest more money on old designs, once all the R&D and setup expense are covered.

Another problem with his assertion, although he never states it explicitly, is that he sort of genericizes the idea of "chips in cars" as if they're all the same and sort of homogenous, and they're all really old.  In a modern car, there are 30-40 different ICs.  Some are quite new and modern, and some are of ancient design.  Some are simple, some are super-complex.

But it's still true that manufacturers eventually want to move on.  There's SOME of that at play here, but it's not THE REASON.

We're seeing the same thing in glass for computer Displays.  Smaller screen sizes just aren't all that profitable for manufacturers, and if you look at the market you're seeing everything under 24" slowly disappear.  We're also competing with TVs, which have gone even larger.  In a couple of years, you won't be able to buy a monitor under 24" in the USA.  China/LATAM and emerging countries are another story of course.

Apple could build the infrastructure to manufacture their own chips. But it only makes economic sense if they want to become a major IC manufacturing company. And it would probably take them 4-5 years minimum to be competitive (note that building the actual fab is an 18-24 month process). 94 can probably give better clarity than I on how long, if they actually started from scratch, it would take if they threw limitless resources at becoming a semi manufacturer.

That would be stupid.

Even Apple can't afford to buy TSMC, who currently has a market cap of $555B. Intel has a market cap of $220B, but their own business isn't the sort of chips Apple uses, and Apple has made it perfectly clear that they'd design their own chips RATHER than use anything from Intel, so I don't think there's a real value there. It also doesn't make sense because Intel sold their NAND business to SKHynix, so it wouldn't get them vertically integrated in NAND either. The same is true of Qualcomm, which has a market cap of $146B. Apple famously designs their own chips rather than buy from Qualcomm, and I don't think Qualcomm has any DRAM or NAND exposure, so it doesn't make sense there. That leaves Micron, with a cap of about $75B. I'm sure Apple could make that deal work. Micron currently doesn't do processors, but it would get Apple vertically integrated in DRAM and NAND and might, over 5+ years, give them a roadmap to do their own processors.

But all that is a TON of work for a company that's raking in money hand over fist and can buy their chips from companies who have legitimate expertise in their manufacturing, and STILL make insane margin on their incredibly overpriced hardware.

All correct.  It would make zero sense for Apple to manufacture its own microchips.  That's not their core competency, and they have absolutely no need to develop that as a core competency. 

AMD was actually quite good at manufacturing its own chips, and still they realized it didn't make any sense to do so, and spun off two different companies in Spansion and Global Foundries, as a result. 

Intel probably shouldn't own their wafer fabs either, but there's a lot of inertia there, that's hard to overcome.

I dont care who owns what or makes what

I just wish more of it was in the US

I dont care who owns what or makes what

I just wish more of it was in the US

@utee94 What is the limitation/drawback to chipmaking within the US? 

I know that many chip fabs don't have as much human labor as other businesses due to the necessity of keeping clean rooms clean. And many of the staff in a chip fab are high-tech workers so it's not like you're going to save a ton of money on labor anyway... Given gross margins for semiconductor manufacturing, it doesn't seem like a market that would be a "race to the bottom" cost-wise...

I have heard that there are significant environmental issues with chip fabs (water/pollution), but I'm not sure how accurate that is. 

Your thoughts? 

I dont care who owns what or makes what

I just wish more of it was in the US

...but then Walmart wouldn't exist....

I realize certain things about wafer fabs can be dirty but NFW can it be any dirtier than a petrochemical plant and there are literally hundreds if not thousands all over the state. Wafer fabs probably ship chemicals in by the 55 gallon drum, chemical plants ship em by the barge and tanker ship. 


Just in my county alone we put in over $20 billion dollars in chemical plants in the last 5 years. And still building. 

Well we (the US) better figure out how to make those things here as I suspect that things are going to start going sideways with China in the next few years. I can see a cold war if not some sort of armed conflict on the horizon. China is looking to exert it's influence in more places around the world and look to minimize the United States influence and strength in the world. And if they continue to hold the means of production for some of our most important and critical goods and components, we are in for a world of hurt.