Second in my continuing series of opinions, I’ll talk about semiconductors. The chips that are in everything. They are in our cars, or ovens, our dish washing machines, our washers and dryers, our lawn mowers, our electric power drills, our cell phones, our heating and cooling systems, our stereos and televisions. Chips are in everything.
Now, there is a chip shortage. No one anywhere can make enough chips. It’s a shortage that shuts down automobile manufacturing plants, and causes automobile manufacturers to sell cars that don’t have all their features, because the chips that enable those features can’t be obtained.
I can hear the endless screaming about how this wouldn’t be a problem if we made all the chips in the USA. We’d have all the chips we wanted.
Talk about not understanding how life works. Let me explain that “build it here!” problem.
Intel. The king of chips in the United States, is currently sinking 25 or more Billion dollars into chip fabrication facilities. 25 Billion. Here’s what all that money does. It makes Intel competitive with TSMC and Global Foundries. That’s it. That’s all it does. It doesn’t provide any capabilities that TSMC and Global Foundries are not already working on, or even trying to surpass. That 25 billion simply makes Intel competitive.
For the past 10 years, Intel has been stuck. Their chip manufacturing process has not improved. They’ve been stuck at the 14 nm manufacturing process, and just recently managed to get 10 nm process working, after years of effort. Meanwhile, TSMC and Global Foundries already have 4nm manufacturing process in place, and working, and are implementing 2nm processes. Intel’s grand plan for that 25 billion dollars of chip fabrication capability brings them into competition with TSMC and Global Foundries.
How did Intel get so far behind? Money. More accurately, stockholders, and company managers. Literally, Intel stopped investing in improving the chip fabrication process. There are multiple reasons the investment stopped. First, their chips were dominant in the market. They had no serious competition. AMD was viewed as a beginner chip for dirt cheap computers, that couldn’t match the performance of Intel systems in any way. ARM was no threat, at best, it was a chip for slow, battery powered, mobile devices.
Intel ruled. Intel could do what it wanted. Intel made the chips it wanted to make. It improved them in little steps, a tweak here, another tweak there. Just enough to keep the sales going. And, if you wanted the best chips, you bought Intel.
Where have we heard that line before? Something about American automobile manufactures owning the market, right?
Then, ARM expanded, and started showing up in everything, not just slow cell phones. Then, AMD made radical changes in the design of its chips. And ARM and AMD both teamed with TSMC, Global Foundries, and others, to produce their chips.
Suddenly, Intel didn’t have the best chips. And the market responded accordingly, with sales of AMD chips growing dramatically, and with customers abandoning Intel for AMD, or ARM (as Apple did).
This is exactly what happens when you own the market. You don’t have to compete. You don’t have to get better. You stagnate. Just like the US automobile manufacturers did.
Let me continue with this example, by throwing in China and Russia. China made big headlines yesterday, with a new ARM processor that has 128 processor cores, and is made on a 5nm fabrication process. That fabrication is performed by TSMC. And, just like that, China has a processor that is competitive on the global market. But, it’s very clear China does not manufacture the chip itself. That’s because China’s chip fabrication capabilities are currently only at the 14mn fabrication level. In other words, China can’t make the chip itself. It has to have someone else make the chip.
Russian semiconductor manufacturing may be more limited that China’s. It appears that the best Russian CPU is manufactured on a 28 nm process, and is no where near competitive with ARM, Intel, or AMD. For Russia to become competitive will require hundreds of billions of dollars in investment. Additionally, Russia will not likely be able to purchase some advanced manufacturing technologies on the global marketplace, and will have to manufacture and develop their own processes. This will take decades.
If we shut down the use of semiconductor manufactures that are not based in the US, as a means of protecting the US manufacturers, and as a means of guaranteeing our own supply of chips in adequate, we will very likely end up with an obsolete semiconductor manufacturing capacity that makes chips that would have been good one or two decades ago, and improvement to the manufacturing process would stop, like it did at Intel, because there would be no reason, and no need, to improve the process. There would be no competition.
It would be the end of Intel and AMD on the global market. It would limit them to the US market only. Maybe not at first. But over time, they would be surpassed, just like the semiconductor manufacturing sectors of Russia and China, which can not manufacture 5 nm chips.
Go ahead. Scream about how we need to develop our own chip manufacturing capacity here, in the US, so we can solve this chips shortage problem, and not be at the mercy of the international market for semiconductors. Then, watch as the US becomes unable to compete with other companies in the semiconductor marketplace, and ends up with chips that only exist in the US, and that do not have the performance or chips from TSMC, Global Foundries, and others.
Now, let me list some names that have gone into the history books, of what were once grand American companies that are now a fraction of what they used to be, or are gone completely.
International Business Machines.
Digital Equipment Corporation.
Companies that, like Ford, General Motors, and Chrysler, had no real competition. Until technology they didn’t develop left them all obsolete, and noncompetitive. IBM still exists, but it’s now mostly a niche company, specializing in really big computers, and in customized cloud computing. It is a shadow of the company it once was.
Look at the names. Learn what happened to them. Learn the stories of their collapses into history. Learn the ending to their stories. Now tell me again, why we have to have an isolated, US controlled semiconductor manufacturing capacity that will always be there for us, even if it means we can’t compete with anyone outside the captive US marketplace.